Field of the invention

The present invention relates to proteins suitable for use in enzyme

replacement therapy of lysosomal storage disease, and to such uses and methods of treatment of lysosomal storage diseases by enzyme replacement therapy.

Background

The lysosomal compartment functions as a catabolic machinery that degrades waste material in cells. Degradation is achieved by a number of hydrolases and transporters compartmentalized specifically to the lysosome. There are today over 40 identified inherited diseases where a link has been established between disease and mutations in genes coding for lysosomal proteins. These diseases are defined as lysosomal storage diseases (LSDs) and are characterized by a buildup of a metabolite (or metabolites) that cannot be degraded due to the insufficient degrading capacity. As a consequence of the excess lysosomal storage of the metabolite, lysosomes increase in size. How the accumulated storage material causes pathology is not fully understood but may involve mechanisms such as inhibition of autophagy and induction of cell apoptosis (Cox & Cachon-Gonzalez, J Pathol 226: 241-254 (2012)).

The missing function caused by a mutated or missing protein may be restored by administration, and thus replacement of, the mutated/missing protein with a protein from a heterologous source. This has been shown for a variety of disease fields. Within the field of hemophilia, administration of both enzymes, such as factor IX and factor VII, and proteins, such as factor VIII, that are part of activation complexes in the coagulation pathway have been successfully employed. These components are of course present in the blood and thus it is easy to administrate a protein to its site of action.

In the field of lysosomal storage diseases, storage can be reduced by administration of a lysosomal enzyme from a heterologous source. It is well established that intravenous administration of a lysosomal enzyme results in its rapid uptake by cells via a mechanism called receptor mediated

endocytosis. This endocytosis is mediated by receptors on the cell surface, and in particular the two mannose-6 phosphate receptors (M6PR) have been shown to be pivotal for uptake of certain lysosomal enzymes (Neufeld; Birth Defects Orig Artie Ser 16: 77-84 (1980)). M6PR recognize phosphorylated oligomannose glycans which are characteristic for lysosomal proteins.

Based on the principle of receptor mediated endocytosis, intravenously administrated enzyme replacement therapies (ERT) are today available for eight LSDs (Gaucher, Fabrys, Pompe and the Mucopolysaccharidosis type I, II, IVA, VI and VII). These therapies are efficacious in reducing lysosomal storage in various peripheral organs and thereby ameliorate some symptoms related to the pathology. Elaprase® and Aldurazyme® are examples of orphan medicinal products indicated for long-term treatment of patients with Hunter syndrome (Mucopolysaccharidosis II, MPSII) and the non-neurological symptoms of patients with Hurler / Scheie syndrome (Mucopolysaccharidosis I, MPS I).

A majority of the LSDs cause build-up of lysosomal storage in the central nervous system (CNS) and consequently presents a repertoire of CNS related signs and symptoms. A major drawback with intravenously

administered ERT is the poor distribution to the CNS. The CNS is protected from exposure to blood borne compounds by the blood-brain barrier (BBB), formed by the CNS endothelium. The endothelial cells of the BBB exhibit tight junctions which prevent paracellular passage, show limited passive

endocytosis and in addition lack some of the receptor mediated transcytotic capacity seen in other tissues.

In addition to the neurological component of LSDs, peripheral pathology is to some extent also sub-optimally addressed in current enzyme replacement treatment. Patients frequently suffer from arthropathy, clinically manifested in joint pain and stiffness resulting in severe restriction of motion. Moreover, progressive changes in the thoracic skeleton may cause respiratory

restriction. Prevailing storage leading to thickening of the heart valves along with the walls of the heart can moreover result in progressive decline in cardiac function. Also pulmonary function can further regress despite enzyme replacement treatment. Distribution of lysosomal protein after intravenous administration is dependent on the pattern of glycosylation of the protein. There are three general types of N-glycans: oligomannose (“high-mannose”), complex and hybrid, all of which are present on lysosomal proteins. In addition, proteins directed to the lysosome carry one or more N-glycans which are phosphorylated. The phosphorylation occurs in the Golgi and generates mannose-6-phospate (M6P) residues that are recognized by mannose-6-phosphate receptors (M6PRs) and initiate the transport of the lysosomal protein to the lysosome. High-mannose glycans steer the protein to mannose receptor rich resident macrophages such as the Kuppfer cells in the liver, whereas the mannose 6- phosphorylated glycans steer the protein to M6PR rich cells such as the hepatocytes in the liver. The uptake of lysosomal proteins into cells is in most cases a rapid process and the half-life of a lysosomal protein in circulation is typically less than 1 hour. Consequently, in many tissues, e.g. tissues that are not well supplied by blood, therapeutically useful levels of lysosomal protein are difficult to achieve.

In view of the above, there is a substantial need for developing novel therapeutic strategies for treatment of lysosomal storage diseases, for instance, lysosomal storage diseases of cells of the central nervous system, and/or cells of bone and cartilage.

Summary of the invention

It is an object of the invention to at least partially overcome or alleviate the problems of the prior art, and to provide a means of improving the distribution of lysosomal proteins to target tissue for the purpose of protein replacement therapy, such as enzyme replacement therapy of a lysosomal storage disease.

This and other objects are achieved by a fusion protein comprising

i) a lysosomal polypeptide; and ii) a polypeptide moiety comprising 2-68 units, each unit being independently selected from the group consisting of all amino acid sequences according to SEQ ID NO: 1 :

X1 -X2-X3-X4-X5-X6-D-X8-X9-X10-X11 (SEQ ID NO: 1 ) in which, independently,

X1 is P or absent;

X2 is V or absent;

X3 is P or T;

X4 is P or T;

X5 is T or V;

X6 is D, G or T;

X8 is A, Q or S;

X9 is E, G or K;

X10 is A, E, P or T;

XI I is A, P or T.

Fusing the lysosomal polypeptide to a polypeptide moiety as defined above to yield a fusion protein according to the invention has been found to reduce the rate of receptor-mediated cellular uptake of the lysosomal polypeptide in vivo, by shielding the lysosomal polypeptide from glycan-recognizing receptors. As a result, the rate of uptake in tissues with abundant glycan-recognizing receptors (in particular the liver) is decreased, and the biological half-life of the lysosomal polypeptide is increased, which allows distribution also to tissues less abundant in glycan-recognizing receptors, and even

transportation across the blood-brain barrier. It has been demonstrated that the fusion protein can retain the biological activity of the lysosomal

polypeptide. Hence, a fusion protein according to the invention can be transported across the blood-brain barrier in mammals and effect an enzymatic activity in the brain.

Also, in addition to allowing improved distribution to the central nervous system, including the brain, the reduced rate of receptor-mediated uptake of the fusion protein allows for better distribution to peripheral tissue involved in a peripheral pathology of a lysosomal storage disease is improved. Thus, the present invention allows development of treatments that could potentially improve clinical outcome in a multitude of lysosomal storage diseases.

From a dosing perspective, reduced clearance of the fusion protein may also advantageously allow for development of long-acting medicaments that can be administered to patients less frequently.

Furthermore, it is believed that the high content of hydrophilic amino acid residues of the shielding polypeptide moiety serves to increase the solubility of the fusion protein relative to the lysosomal polypeptide as such.

It has also been found that the polypeptide moiety may improve the thermodynamic stability of the fusion protein relative to the lysosomal polypeptide alone. For instance, it has been found that the fusion protein may have a reduced tendency to aggregate when present in solution, and/or maintain its native structure (such as folding or dimerization) at elevated temperature. An increased thermodynamic stability may, for instance, enable higher expression levels during production, facilitate protein purification and processing, allow formulation of compositions, including pharmaceutical compositions, of higher concentration of the active polypeptide, and/or provide better storage stability or increased shelf-life of a formulation including the fusion protein.

As used herein, the term“lysosomal polypeptide” means a polypeptide that exerts its biological activity in the lysosomal compartment of cells. Lysosomal polypeptides used in the present invention are typically degrading enzymes, such as sulfatases, glycoside hydrolases or proteases.

The polypeptide moiety that forms part of the fusion protein of the invention may be referred to as a“receptor mediated uptake reducing polypeptide” or as a“shielding polypeptide moiety”. As used herein, by“shielding polypeptide moiety” is meant a polypeptide that, by being fused to a polypeptide of interest, in particular a lysosomal polypeptide, reduces the interactions of the polypeptide of interest with glycan recognition receptors, thereby enabling, upon systemic administration to a subject, an altered in vivo distribution in terms of increased distribution of the polypeptide of interest to one or more diseased tissues or organs, such as a tissue or organ affected by a lysosomal storage disease.

In the context of the present invention, the amino acid sequences of the fusion partners of the fusion protein are referred to using the terms

“polypeptide” and“polypeptide moiety”. Notably, these terms are intended to include amino acid sequences as short as 18 amino acids, which effectively represents the smallest version of the shielding polypeptide moiety (2 units each of 9 amino acids). An amino acid sequence of up to about 50 amino acids may sometimes be referred to as“peptide”; however for the sake of simplicity, in the present specification, the amino acid sequences of the fusion protein will be referred to as“polypeptide” or“polypeptide moiety” throughout. By“glycan recognition receptors” is meant receptors that recognize and bind to proteins mainly via glycan moieties of the proteins. For lysosomal proteins, such receptors can be exemplified by the mannose receptor, which

selectively binds proteins where glycans exhibit exposed terminal mannose residues, and mannose 6-phosphate receptors, which are characteristic of lysosomal proteins. Lectins constitute another large family of glycan

recognition receptors which can be exemplified by the terminal galactose recognizing asialoglycoprotein receptor 1 , recognizing terminal galactose residues on glycans.

By“retains the biological activity” or“retained biological activity” of the lysosomal polypeptide is meant that the biological activity of the fusion protein is retained at least partly from the lysosomal polypeptide as such.

Surprisingly, it was found that a fusion protein according to some

embodiments of the invention had not only a fully retained biological activity, but in fact had an increased or improved biological activity. The provision of a highly active protein offers a possibility of formulating an even more potent therapeutic for treatment of LSDs.

Typically, the lysosomal polypeptide is an enzymatically active lysosomal polypeptide. Thus, the biological activity of the lysosomal polypeptide is typically an enzymatic activity. In embodiments, the lysosomal polypeptide may be a sulfatase, a glycoside hydrolase, or a protease. In particular, the lysosomal polypeptide may be selected from a sulfatase and a glucoside hydrolase. In some embodiments, the lysosomal polypeptide is a sulfatase.

In embodiments, the lysosomal polypeptide may be selected from the group consisting of: deoxyribonuclease-2-alpha; beta-mannosidase; ribonuclease T2; lysosomal alpha-mannosidase; alpha L-iduronidase; tripeptidyl-peptidase 1 ; hyaluronidase-3; cathepsin L2; ceroid-lipofuscinosis neuronal protein 5; glucosylceramidase; tissue alpha-L-fucosidase; myeloperoxidase; alpha- galactosidase A; beta-hexosaminidase subunit alpha; cathepsin D;

prosaposin; beta-hexosaminidase subunit beta; cathepsin L1 ; cathepsin B; beta-glucuronidase; pro-cathepsin H; non-secretory ribonuclease; lysosomal alpha-glucosidase; lysosomal protective protein; gamma-interferon-inducible lysosomal thiol reductase; tartrate-resistant acid phosphatase type 5;

arylsulfatase A; prostatic acid phosphatase; N-acetylglucosamine-6-sulfatase; arylsulfatase B; beta-galactosidase; alpha-N-acetylgalactosaminidase;

sphingomyelin phosphodiesterase; ganglioside GM2 activator; N(4)-(beta-N- acetylglucosaminyl)-L-asparaginase; iduronate 2-sulfatase; cathepsin S; N- acetylgalactosamine-6-sulfatase; lysosomal acid lipase/cholesteryl ester hydrolase; lysosomal Pro-X carboxypeptidase; cathepsin O; cathepsin K; palmitoyl-protein thioesterase 1 ; arylsulfatase D; dipeptidyl peptidase 1 ;

alpha-N-acetylglucosaminidase; galactocerebrosidase; epididymal secretory protein E1 ; di-N-acetylchitobiase; N-acylethanolamine-hydrolyzing acid amidase; hyaluronidase-1 ; chitotriosidase-1 ; acid ceramidase; phospholipase B-like 1 ; proprotein convertase subtilisin/kexin type 9; group XV

phospholipase A2; putative phospholipase B-like 2; deoxyribonuclease-2- beta; gamma-glutamyl hydrolase; arylsulfatase G; L-amino-acid oxidase; sialidase-1 ; legumain; sialate O-acetylesterase; thymus-specific serine protease; cathepsin Z; cathepsin F; prenylcysteine oxidase 1 ; dipeptidyl peptidase 2; lysosomal thioesterase PPT2; heparanase; carboxypeptidase Q; b-glucuronidase, N-sulfoglucosamine sulfohydrolase (sulfamidase), and sulfatase-modifying factor 1.

In some embodiments, the lysosomal polypeptide may be selected from the group consisting of sulfamidase, iduronate 2-sulfatase, arylsulfatase A, arylsulfatase B, N-acetylgalactosamine-6-sulfatase, galactocerebrosidase, alpha-L-iduronidase, and beta-glucuronidase. For instance, the lysosomal polypeptide may comprise an amino acid sequence selected from the group consisting of SEQ ID NOs: 33, 37, 49,232, 65, 78 and 82-149.

Typically, the fusion protein according to embodiments of the invention may comprise a lysosomal polypeptide that is glycosylated. In embodiments, the lysosomal polypeptide may comprise high-mannose glycans, and/or at least one phosphorylated N-glycan, for example an N-glycan comprising at least one mannose-6-phosphate (M6P) moiety. In embodiments, the lysosomal polypeptide of the fusion protein is capable of receptor-mediated uptake via a mannose-6-phosphate receptors (M6PRs), such as a calcium dependent M6PR and calcium independent M6PR.

The polypeptide moiety of the fusion protein according to embodiments of the invention may be a shielding polypeptide moiety capable of reducing the rate of receptor-mediated uptake of the lysosomal polypeptide. The uptake rate reduction may be due to steric hindrance or shielding of glycan moieties from glycan-recognizing receptors on cell surfaces.

The polypeptide moiety may form a contiguous sequence of 2-68 units of one or more sequence(s) as defined in SEQ ID NO: 1. In embodiments of the invention, the polypeptide moiety, which may also be referred to as a shielding polypeptide moiety, may comprise from 3 to 51 units as defined above. For example, the polypeptide moiety may comprise 3-34 units, such as from 3-17 units, or from 3 to 9 units, as defined above.

Optionally, the fusion protein may comprise a plurality of polypeptide moieties, such as at least two polypeptide moieties, each of which may be as described herein, e.g. each comprising from 2 to 68 units, such as from 3-34 units, such as from 3-17 units, or from 3 to 9 units. Such multiple polypeptides moieties may be of the same length (having the same number of units), or may be of different lengths.

The polypeptide moiety, or at least one of said plurality of polypeptide moieties, may be positioned N-terminally or C-terminally of said lysosomal polypeptide. Alternatively, the polypeptide moiety, or at least one of a plurality of polypeptide moieties, may constitute an insertion into, or replacement of a part of, the amino acid sequence of the lysosomal polypeptide.

In embodiments of the invention, at least one of the residues X3 and X4 of SEQ ID NO:1 may be P. In some embodiments, at least one of X4 and X5 of SEQ ID NO:1 may be T. In some embodiments, at least one of X10 and X11 of SEQ ID NO:1 may be A or P. In some embodiments, X1 is P and X2 is V.

In embodiments of the fusion protein according to the invention, each unit of the polypeptide moiety may be selected from the group consisting of SEQ ID NOs: 2-11 and 159-230. That is, in embodiments of the invention, the polypeptide moiety may comprise 2-68 units of one or more amino acid sequence(s) independently selected from the group consisting of SEQ ID NOs: 2-11 and 159-230. In some embodiments, each unit of the polypeptide moiety may be selected from the group consisting of SEQ ID NOs: 2-11. SEQ ID NOs: 2-11 sequences represent human variants of SEQ ID NO: 1.

In some embodiments, the polypeptide moiety may correspond to a naturally occurring human amino acid sequence. Such embodiments include, for instance, fusion proteins as defined above wherein the polypeptide moiety comprises, or consists of, at least one amino acid sequence selected from the group consisting of SEQ ID NOs: 12-27. In some embodiments, the polypeptide moiety comprises, or consists of, an amino acid sequence selected from the group consisting of SEQ ID NOs: 73-77.

The use of a sequence of human origin may be advantageous as it is expected to contribute to a low immunogenicity in human subjects. Amino acid sequences based on repeating units selected from SEQ ID NOs: 2-11 evaluated in vitro and in silico were found to have low immunogenic potential. Hence, shielding polypeptide moieties consisting of such units are expected to be well tolerated, in terms of immune response, by human subjects.

In embodiments, the fusion protein comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 28-32, 34-36, 38-48, 50- 60, 62-64, 66-72, 79-81 , and 150-158. The lysosomal polypeptides contemplated for use in the fusion protein of the invention often occur naturally in the form of multimeric protein complexes, such as dimers. Hence, according to another aspect, the invention relates to a multimeric protein complex comprising at least two fusion proteins according to the first aspect of the invention. In particular, the multimeric protein complex may be a dimer, that is, it may comprise a dimer of the fusion protein.

In yet another aspect, the invention provides a composition, typically a pharmaceutical composition, comprising a fusion protein or a multimeric protein complex as defined herein and at least one carrier or ingredient, typically at least one pharmaceutically acceptable carrier, diluent, or ingredient. The composition may be formulated as a pharmaceutical composition intended for administration to a subject in need thereof, in particular for treatment of a lysosomal storage disease. The composition may be formulated e.g. for subcutaneous or intravenous administration.

In further aspects, the invention provides a fusion protein, a multimeric protein complex, or a composition according to the above-mentioned aspects of the invention, for use as a medicament, and particularly for use in treatment of a lysosomal storage disease, by enzyme replacement therapy. Hence, the invention also provides a method of treating a lysosomal storage disease, comprising administering to a subject in need thereof a therapeutically effective amount of the fusion protein, the multimeric protein complex, or the composition according to the above-mentioned aspects.

The lysosomal storage disease to be treated may be manifested in one or more tissues and/or organs selected from: neural tissue, central nervous system, peripheral nervous system, brain, muscular tissue, endothelial tissue, heart, lungs, skeletal muscle, connective tissue, cartilage, bone, skeletal bone, and joints. In some embodiments, the organ may be the brain. In some embodiments, tissue may be cartilage and/or bone. In some embodiments, the lysosomal storage disease may be manifested in peripheral tissue.

As used herein,“peripheral tissue” is intended to mean all tissues outside of the central nervous system. Examples of peripheral tissue include muscular tissue, endothelial tissue, heart tissue, lung tissue, muscular tissue, connective tissue, cartilage, and bone.

When the fusion protein, protein complex, or composition is used as a medicament and/or in a method of treating a lysosomal storage disease, the rate of receptor-mediated cellular uptake of the lysosomal polypeptide may be reduced. The uptake may typically be mediated by mannose receptors and/or mannose-6-phosphate receptors (M6PR). In embodiments, the distribution of the fusion protein or the multimeric protein complex to a target tissue or organ affected by the lysosomal storage disease may be increased, relative to the distribution of the lysosomal polypeptide as such (although possibly in multimeric form) without the polypeptide moiety. The target tissue may in particular be selected from: neural tissue, central nervous system, peripheral nervous system, brain, muscular tissue, endothelial tissue, heart, lungs, skeletal muscle, connective tissue, cartilage, bone, skeletal bone, and joints. Thus, the present invention allows development of treatments that could potentially improve clinical outcome in a multitude of lysosomal storage diseases. It is noted that the invention relates to all possible combinations of the features recited in the claims.

Brief description of the drawings

The above, as well as additional objects, features and advantages of the present inventive concept, will be better understood through the following illustrative and non-limiting detailed description, with reference to the appended drawings, in which:

Figure 1 is a schematic representation of a gene encoding a biologically active polypeptide (white) and one or more gene(s) encoding a shielding polypeptide moiety (shaded) according to embodiments of the invention.

Figure 2 is a photograph of an SDS-PAGE gel showing the results of electrophoretic separation of different fusions proteins of sulfamidase with shielding polypeptide moieties according to embodiments of the invention.

Figure 3 is a graph showing the aggregation propensity measured by static light scattering (counts) as function of increased temperature (°C), for sulfamidase (reference) as well as fusion proteins of sulfamidase with shielding polypeptide moieties, according to embodiments of the invention.

Figure 4 is a graph showing the mean and standard deviation serum concentration of sulfamidase (reference, filled boxes) and sulfamidase fusion proteins, according to embodiments of the invention, versus time following a 10 mg/kg i.v. administration in mice.

Figure 5 is a graph showing the relationship between the apparent molecular weight in solution in kDa (y-axis) of fusion proteins according to embodiments of the invention and the number of repeat units of the shielding polypeptide moiety and number of repeat units (x-axis).

Figure 6 is a graph showing the cellular uptake of iduronate-2-sulfatase and iduronate-2-sulfatase fusion proteins in the presence of increasing

concentration of M6P.

Figure 7 is a photograph of an SDS-PAGE gel showing the results of electrophoretic separation of cell harvest media from production of GalN6S (reference) and GalN6S fusion proteins according to embodiments of the invention.

Figure 8 is a photograph of an SDS-PAGE gel showing the results of electrophoretic separation of ASB (reference) and the cell media harvest from production of ASB fusion protein according to an embodiment of the invention.

Figure 9 is a graph showing the mean and standard deviation serum concentration of iduronate 2-sulfatase (reference, SEQ ID NO: 37, filled boxes) and an iduronate 2-sulfatase fusion protein (SEQ ID NO: 46), according to embodiments of the invention, versus time following a 10 mg/kg i.v. administration in mice.

Figure 10 is a graph showing the normalised scattering signal as function of increased temperature (°C), for arylsulfatase A (PSI0590) represented by a solid black line, and fusion proteins of arylsulfatase A with shielding polypeptide moieties (PSI0681 , SEQ ID NQ:50; PSI0683, SEQ ID NO: 52; PSI0685, SEQ ID NO: 54; PSI0687, SEQ ID NO: 56; PSI0689, SEQ ID NO: 58; PSI0691 , SEQ ID NO: 60), each represented by a dashed line, according to embodiments of the invention.

Figure 11 is a graph showing the normalised scattering signal as function of increased temperature (°C), for iduronidase (PSI0754, SEQ ID NO: 61 ) represented by a solid black line, and fusion proteins of iduronidase with shielding polypeptide moieties (PSI0753, SEQ ID NO: 62; PSI0755, SEQ ID NO: 63; PSI0756, SEQ ID NO: 64), each represented by a dashed line, according to embodiments of the invention.

Figure 12 is a photograph of an SDS-PAGE gel showing the results of electrophoretic separation of the cell media harvest from production of GALC fusion protein according to an embodiment of the invention. Lane A:

Molecular weight marker SeeBlue plus 2 (Thermo Fisher Scientific), lane B: unrelated cell media harvest, lane C: GALC-SPM1 -17 (PSI0817), lane D: SPM1 -17-GALC (PSI0818), lane E: SPM1 -17-GALC-SPM1 -17 (PSI0819).

The star indicates produced fusion protein in harvest media.

Detailed description

The present invention is based on the realization that fusion of a lysosomal polypeptide to a polypeptide moiety as defined hereinbelow provides a means of reducing the rate of receptor-mediated uptake of the lysosomal

polypeptide, in particular reduce the rate of uptake in the liver, and thereby, through an increased biological half-life, enable distribution to tissues that have previously been poorly targeted.

Lysosomal proteins are usually rapidly cleared from circulation when administrated by intravenous injection. As described in more detail below, cellular uptake from the extracellular compartment is facilitated by receptors recognising the characteristic mannose and mannose 6-phosphate rich glycans of lysosomal proteins. Thus, distribution of lysosomal proteins is at least partly controlled by the density of these receptors on different cells. While the mannose recognizing receptors are abundantly present on tissue- resident macrophages and sinusoidal endothelial cells in the liver, the cation independent mannose 6-phosphate receptor is abundant on hepatocytes. Consequently, a major part of the dose of an intravenously administrated therapeutic lysosomal enzyme may distribute to the liver, which is sub-optimal for most therapeutic applications. For example, the two therapeutic a- galactosidase A preparations used as treatment for Fabry disease both show 60-70 % of the dose distributed to liver after a single dose in mice (Lee et al, Glycobiology 13: 305-313 (2003)). In contrast, cells in tissues that are not very well supplied by blood and/or have low abundance of receptors are not sufficiently targeted via these uptake mechanisms. It has been found that by preventing rapid uptake via the glycan-dependent routes, clearance from the circulation is significantly reduced and other, slower processes facilitate uptake into cells, resulting in a different distribution profile. This enables distribution of therapeutic proteins to cells of tissues that are by nature poorly exposed to lysosomal enzymes. In particular embodiments, a fusion protein as disclosed herein may provide a better distribution of a lysosomal polypeptide in joints, connective tissue, cartilage and bone, when

administrated by intravenous infusion. Also the central nervous system, peripheral nerves, skeletal muscle, heart and lung may be better targeted. These are all tissues where a severe pathology is commonly manifested as a consequence of lysosomal storage.

Distribution of lysosomal protein after intravenous administration is highly dependent on the pattern of glycosylation of the protein, in particular N- glycosylation.

In general, N-glycosylations can occur at an Asn-X-Ser/Thr sequence motif. To this motif the initial core structure of the N-glycan is transferred by the glycosyltransferase oligosaccharyltransferase, within the reticular lumen. This common basis for all N-linked glycans is made up of 14 residues: 3 glucose, 9 mannose, and 2 N-acetylglucosamine residues. This precursor is then converted into three general types of N-glycans; oligomannose, complex and hybrid, by the actions of a multitude of enzymes that both trims down the initial core and adds new sugar moieties. Each mature N-glycan contains the common core Man(Man)2-GlcNAc-GlcNAc-Asn, where Asn represents the attachment point to the protein.

In addition, proteins directed to the lysosome carry one or more N-glycans which are phosphorylated. The phosphorylation occurs in the Golgi and is initiated by the addition of N-acetylglucosamine-1 -phosphate to C-6 of mannose residues of oligomannose type N-glycans. The N- acetylglucosamine is cleaved off to generate mannose-6-phospate (M6P) residues, that are recognized by mannose-6-phosphate receptors (M6PRs) and will initiate the transport of the lysosomal protein to the lysosome. The resulting N-glycan is then trimmed to the point where the M6P is the terminal group of the N-glycan chain. The binding site of the M6PR requires a terminal M6P group that is complete, as both the sugar moiety and the phosphate group is involved in the binding to the receptor (Kim et al, Curr Opin Struct Biol 19(5):534-42 (2009)).

Generally, a lysosomal polypeptide is a polypeptide that exerts its biological activity in the lysosome. Lysosomal polypeptides used in the present invention are typically degrading enzymes, such as sulfatases, glycoside hydrolases or proteases.

In embodiments, the lysosomal polypeptide is a lysosomal protein lacking transmembrane helices and having at least one N-glycosylation site.

Examples of such lysosomal proteins are listed in the table I below: In Table I a number of N-glycosylated lysosomal proteins are listed. Some of the proteins might be known under other names. It should be understood that the protein listing above also encompasses any and all alternative names. Table I. Examples of lysosomal proteins

In one embodiment, the lysosomal protein is selected from the group consisting of deoxyribonuclease-2-alpha; beta-mannosidase; ribonuclease T2; lysosomal alpha-mannosidase (Laman); tripeptidyl-peptidase 1 (TPP-1 ); hyaluronidase-3 (Hyal-3); cathepsin L2; ceroid-lipofuscinosis neuronal protein 5; glucosylceramidase; tissue alpha-L-fucosidase; myeloperoxidase (MPO); alpha-galactosidase A ;beta-hexosaminidase subunit alpha; cathepsin D; prosaposin; beta-hexosaminidase subunit beta; cathepsin L1 ; cathepsin B; beta-glucuronidase; pro-cathepsin H; cathepsin H; non-secretory

ribonuclease; lysosomal alpha-glucosidase; lysosomal protective protein; gamma-interferon-inducible lysosomal thiol reductase; tartrate-resistant acid phosphatase type 5 (TR-AP); arylsulfatase A (ASA); prostatic acid

phosphatase (PAP); N-acetylglucosamine-6-sulfatase; arylsulfatase B (ASB); beta-galactosidase; alpha-N-acetylgalactosaminidase; sphingomyelin phosphodiesterase; ganglioside GM2 activator; N(4)-(beta-N- acetylglucosaminyl)-L-asparaginase; iduronate 2-sulfatase; cathepsin S; N- acetylgalactosamine-6-sulfatase; alpha-L-iduronidase; lysosomal acid lipase/cholesteryl ester hydrolase (Acid cholesteryl ester hydrolase) (LAL); lysosomal Pro-X carboxypeptidase; cathepsin O; cathepsin K; palmitoyl- protein thioesterase 1 (PPT-1 ); N-sulfoglucosamine sulfohydrolase

(sulfamidase); arylsulfatase D (ASD); dipeptidyl peptidase 1 ; alpha-N- acetylglucosaminidase; galactocerebrosidase (GALCERase); epididymal secretory protein E1 ; di-N-acetylchitobiase; N-acylethanolamine-hydrolyzing acid amidase; hyaluronidase-1 (Hyal-1 ); chitotriosidase-1 ; acid ceramidase (AC); phospholipase B-like 1 ; proprotein convertase subtilisin/kexin type 9; group XV phospholipase A2; putative phospholipase B-like 2;

deoxyribonuclease-2-beta; gamma-glutamyl hydrolase; arylsulfatase G (ASG); L-amino-acid oxidase (LAAO) (LAO); sialidase-1 ; legumain; sialate O- acetylesterase; thymus-specific serine protease; cathepsin Z; cathepsin F (CATSF); prenylcysteine oxidase 1 ; dipeptidyl peptidase 2; lysosomal thioesterase PPT2 (PPT-2); heparanase; carboxypeptidase Q; b- glucuronidase, and sulfatase-modifying factor 1.

In certain embodiments of aspects disclosed herein, the lysosomal protein is a sulfatase. Said sulfatase preferably has a FGIy residue at its active site. In some embodiments, said sulfatase is thus selected from arylsulfatase A; N- acetylglucosamine-6-sulfatase, arylsulfatase B; iduronate 2-sulfatase; N- acetylgalactosamine-6-sulfatase; N-sulfoglucosamine sulfohydrolase

(sulfamidase); arylsulfatase D, and arylsulfatase G. In particular, said sulfatase is arylsulfatase A; N-acetylglucosamine-6-sulfatase; arylsulfatase B; iduronate 2-sulfatase; N-acetylgalactosamine-6-sulfatase or sulfamidase. In one embodiment, said sulfatase is arylsulfatase A. In one embodiment, said sulfatase is sulfamidase. In one embodiment, said sulfatase is iduronate-2- sulfatase. In one embodiment, said sulfatase is arylsulfatase B. In one embodiment, said sulfatase is N-acetylgalactosamine-6-sulfatase.

In embodiments of aspects disclosed herein, the lysosomal protein is a glycoside hydrolase. In some embodiments, said glycoside hydrolase is selected from alpha-galactosidase A; tissue alpha-L-fucosidase;

glucosylceramidase; lysosomal alpha-glucosidase; beta-galactosidase; beta- hexosaminidase subunit alpha; beta-hexosaminidase subunit beta;

galactocerebrosidase; lysosomal alpha-mannosidase; beta-mannosidase; alpha-L-iduronidase; alpha-N-acetylglucosaminidase; beta-glucuronidase; hyaluronidase-1 ; alpha-N-acetylgalactosaminidase; sialidase-1 ; di-N- acetylchitobiase; chitotriosidase-1 ; hyaluronidase-3, and heparanase.

Preferably, said glycoside hydrolase is alpha-L-iduronidase, beta- glucuronidase or galactocerebrosidase. In one embodiment, said glycoside hydrolase is alpha-L-iduronidase. In one embodiment, said glycoside hydrolase is beta-glucuronidase. In one embodiment, said glycoside hydrolase is galactocerebrosidase.

In embodiments of aspects disclosed herein, the lysosomal protein is a protease. In some embodiments, said protease is selected from cathepsin D; cathepsin L2; cathepsin L1 ; cathepsin B; pro-cathepsin H; cathepsin S;

cathepsin O; cathepsin K; dipeptidyl peptidase 1 ; cathepsin Z; cathepsin F; legumain; gamma-glutamyl hydrolase; tripeptidyl-peptidase 1 ;

carboxypeptidase Q; lysosomal protective protein; lysosomal pro-X

carboxypeptidase; thymus-specific serine protease; dipeptidyl peptidase 2, and proprotein convertase subtilisin/kexin type 9. In one embodiment, said protease is tripeptidyl-peptidase 1.

In embodiments, the lysosomal protein comprises polypeptide consisting of an amino acid sequence selected from any one of SEQ ID NOs: 33, 37, 49, 232, 65, 78 and 82-149, or a polypeptide having at least 90 % sequence identity with an amino acid sequence selected from SEQ ID NOs: 33, 37, 49, 61 , 65, 78 and 82-149. In a non-limiting example, said polypeptide has at least 95 % sequence identity with an amino acid sequence selected from SEQ ID NOs: 33, 37, 49, 61 , 65, 78 and 82-149, such as at least 98 % sequence identity or at least 99 % sequence identity with an amino acid sequence selected from SEQ ID NOs: 33, 37, 49, 61 , 65, 78 and 82-149.

In other embodiments, the lysosomal polypeptide is a sulfatase and

comprises an amino acid sequence selected from any one of SEQ ID NO: 49; SEQ ID NO 108; SEQ ID NO 109; SEQ ID NO 37; SEQ ID NO 65; SEQ ID NO: 33, SEQ ID NO: 121 ; and SEQ ID NO: 136. In a preferred embodiment, said polypeptide has an amino acid sequence is selected from SEQ ID NO: 49, SEQ ID NO: 109; and SEQ ID NO: 37.

In another embodiment, the lysosomal protein is a glycoside hydrolase and comprises a polypeptide consisting of an amino acid sequence selected from any one of SEQ ID NO: 93; SEQ ID NO: 91 ; SEQ ID NO: 90; SEQ ID NO: 103; SEQ ID NO: 110; SEQ ID NO: 94; SEQ ID NO: 97; SEQ ID NO: 78; SEQ ID NO: 85; SEQ ID NO: 83; SEQ ID NO: 61 ; SEQ ID NO: 123; SEQ ID NO: 100; SEQ ID NO: 127; SEQ ID NO: 111 ; SEQ ID NO: 138; SEQ ID NO: 125; SEQ ID NO: 128; SEQ ID NO: 87; and SEQ ID NO: 147. In a preferred embodiment, said polypeptide has an amino acid sequence as set out in SEQ ID NO: 61 or SEQ ID NO:78.

In another embodiment, the lysosomal protein is a protease and comprises a polypeptide consisting of an amino acid sequence selected from any one of SEQ ID NO: 95; SEQ ID NO: 143; SEQ ID NO: 86; SEQ ID NO:104; SEQ ID NO:131 ; SEQ ID NO:122; SEQ ID NO:119; SEQ ID NO:88; SEQ ID NO:98; SEQ ID NO:99; SEQ ID NO:101 ; SEQ ID NO:115; SEQ ID NO:118; SEQ ID NO:142; SEQ ID NO:139; SEQ ID NO:135; SEQ ID NO:117; SEQ ID NO:141 ; and SEQ ID NO:145. In a preferred embodiment, said polypeptide has an amino acid sequence as set out in SEQ ID NO: 86.

The present inventors surprisingly found that the shielding polypeptide as described herein has the potential, when fused to a lysosomal protein, to decrease the rate of receptor-mediated uptake of the lysosomal protein, possibly by steric hindrance. The shielding polypeptide is designed based on the C-terminal domain of human bile salt-stimulated lipase (BSSL).

Bile salt-stimulated lipase (BSSL), also referred to as bile salt-activated lipase (BAL) or carboxylic ester lipase (CEL) is a lipolytic enzyme produced by the human lactating mammary gland and pancreas. The protein is arranged in two domains, a large globular amino-terminal domain and a smaller but extended carboxy-terminal (C-terminal) domain (for a review, see e.g. Wang & Hartsuck (1993) Biochim. Biophys Acta 1166: 1 -19). The present inventors have found that repetitive sequences based on or derived from the C-terminal domain of human BSSL can be successfully fused to various lysosomal polypeptides and confer reduced receptor-mediated uptake of the fusion partner, thereby extending its biological half-life and allowing for a different distribution pattern in vivo.

The C-terminal domain of human BSSL consists of repeating units of, or similar to, the formula“PVPPTGDSGAP”. Table 2 in Example 1 below lists the repeating units from human BSSL variants. The most common form of the C-terminal domain contains 18 repeating units (UniProt entry P19835).

However, there are variations in the human population, both with regards to the number of repeating units, and the amino acid sequence of the individual repeating units. Furthermore, each repeating unit has one site that may be O- glycosylated, increasing the hydrophilicity and size of the region (Stromqvist et al. Arch. Biochem. Biophys. 1997). The C-terminal end of the domain is however hydrophobic, and has been shown to bind into the active site of BSSL and cause auto-inhibition of the enzyme. The most frequent human sequence of this hydrophobic portion is“QMPAVIRF” (SEQ ID NO: 231 ) (Chen et al. Biochemistry 1998).

It has previously been speculated that the C-terminal domain may be responsible for the stability of BSSL in vivo, for example its resistance to denaturation by acid and aggregation under physiological conditions (Loomes et al., Eur. J. Biochem. 1999, 266, 105-111 ). In contrast, another study of the cholesterol esterase structure showed that the C-terminal domain, which is enriched with Pro, Asp, Glu, Ser and Thr residues, is reminiscent of the PEST-rich sequences in short-lived proteins, suggesting that the protein may have a short half-life in vivo due to the repetitive sequences in the C-terminal domain (Kissel et al., Biochimica et Biophysica Acta 1989, 1006).

In the present invention, the extended biological half-life of a fusion protein comprising a shielding polypeptide moiety as defined herein, based on or derived from the C-terminal domain of human BSSL, is believed to be due mainly to the decreased receptor mediated uptake of the protein. However, it is also envisaged that other mechanisms may contribute to the increased biological half-life.

As used herein, the expressions“fused” and“fusion” refer to the joining of two or more portions of chemical entities of the same kind, such as peptides, polypeptides, proteins, or nucleic acid sequences. A fusion protein as referred to herein typically comprises at least two polypeptide portions, which may be of different origin; for instance, a biologically active polypeptide, which is not BSSL, and a shielding polypeptide moiety, which may be derived from BSSL. Generally, a fusion may contain the fused portions in any order and at any position; however, a fusion of genes is typically made in-frame (in-line), such that the open reading frames (ORFs) of the fused genes are maintained, as appreciated by persons of skill in the art. In embodiments, both fusion partners of the fusion protein may be of human origin, but the fusion protein does not correspond to a naturally occurring human protein.

Figure 1 schematically illustrates nucleic acid constructs encoding a fusion protein according to embodiments of the present invention, comprising at least one gene encoding a biologically active polypeptide (Fig. 1 (a)-(g), white bar), such as a lysosomal polypeptide, and at least one gene encoding a shielding polypeptide moiety ((b)-(g), dashed bar). For simplicity other elements such as promoter or enhancer sequences and the like are not marked, although a person of skill in the art will appreciate that such elements may be included as necessary. For instance, the gene encoding the biologically active polypeptide may be preceded by a signaling peptide for expression in mammalian cells. As shown in Fig. 1 , the gene encoding the shielding polypeptide moiety may be located C-terminally, see Fig. 1 (b); N- terminally, see Fig. 1 (c); or both N- and C-terminally, Fig. 1 (d), to the gene(s) encoding at least one biologically active polypeptide. Alternatively, a nucleotide sequence encoding a shielding polypeptide moiety may be positioned within the boundaries of the gene encoding the biologically active polypeptide (in-line positioning). In such embodiments, sequences encoding shielding polypeptide moieties may optionally be present at multiple sites, e.g. at three sites as shown in Fig. 1 (f), or more sites as desired, as long as the insertion does not disrupt the tertiary or folding structure of the biologically active polypeptide. In-line positioning of one or more shielding moieties may be combined with N- and/or C-terminal fusion(s). However, as explained above with reference to Figure 1 , the shielding polypeptide moiety is not necessarily located at the C-terminal of the biologically active polypeptide. In embodiments of the invention, the at least one shielding polypeptide moiety may be located at the N-terminal of the biologically active polypeptide (Fig. 1 (c)), or shielding moieties may be located each at the N-terminal and C- terminal, respectively (Fig. 1 (d)). In other embodiments, one or more shielding polypeptides may be inserted at a position within the biologically active polypeptide (Fig. 1 (e)), for example in a position located in a surface-exposed loop of the biologically active polypeptide.

In some embodiments, the shielding polypeptide moiety may replace a specific sequence segment of the biologically active polypeptide. For instance, when positioned as an insert, the shielding polypeptide moiety may replace a part of a surface-exposed loop on the biologically active

polypeptide. Alternatively, a shielding polypeptide may replace an entire domain, such as a N-terminal or a C-terminal domain, or an internal domain, of the biologically active polypeptide.

In yet other embodiments, an in-line inserted shielding polypeptide moiety may be combined with either an N-terminal moiety, a C-terminal moiety, or both N-terminal and C-terminal shielding polypeptide moieties, see Fig. 1 (f). Notably, in embodiments of the invention comprising multiple shielding moieties, located at different positions, each such shielding moiety may be independently defined as described herein. Otherwise stated, each such shielding moiety may comprise from 2 to 68 units of an amino acid sequence according to SEQ ID NO: 1.

Finally, the present invention is not limited to the use of a single biologically active polypeptide as fusion partner; rather, as illustrated in Fig. 1 (g), it is envisaged that in some embodiments the fusion protein may comprise multiple biologically active polypeptides, at least one being a lysosomal polypeptide, for instance two or more lysosomal polypeptides, or a

combination of at least one lysosomal polypeptide(s) with one or more other biologically active polypeptides, separated by linkers, and/or, as in the example of Fig. 1 (g), by a shielding polypeptide. Alternatively, or additionally, one or more shielding polypeptide moiety or moieties may also be located at the N- or C-terminal of the fusion protein.

In the case of multiple biologically active polypeptides, these may be the same or different. Hence, the fusion protein may comprise two different biologically active polypeptides, at least one being a lysosomal polypeptide, optionally separated by a linker or spacer sequence and/or a shielding polypeptide moiety. Alternatively, the fusion protein may comprise three different biologically active polypeptides, at least one being a lysosomal polypeptide. In embodiments where the fusion protein comprises more than one, e.g. two, lysosomal polypeptides, these may be the same or different.

In embodiments of the fusion protein including biologically active polypeptides other than lysosomal polypeptides, such biologically active polypeptides may be selected from the group consisting of growth factors, cytokines, enzymes and ligands, and that the remaining biologically active polypeptide(s) may be selected from antibodies or antibody fragments. As an example, the shielding polypeptide moiety may be positioned as a linker between different antigen- binding regions.

According to the invention, the shielding polypeptide moiety used for fusion with a lysosomal polypeptide comprises an amino acid sequence comprising 2-68 repeating units, each unit being independently selected from the group of amino acid sequences defined by SEQ ID NO: 1 :

X1 -X2-X3-X4-X5-X6-D-X8-X9-X10-X11 (SEQ ID NO: 1 ) in which, independently,

XI is P or absent;

X2 is V or absent;

X3 is P or T;

X4 is P or T;

X5 is T or V;

X6 is D, G or T;

X8 is A, Q or S;

X9 is E, G or K;

X10 is A, E P or T;

XI I is A, P or T.

As used herein, a“unit” refers to an occurrence of an amino acid sequence of the general formula according to SEQ ID NO: 1 as defined above, including for instance any of the sequences according to SEQ ID NOs: 2-11 and 159- 230, such as any of the sequences according to SEQ ID NO: 2-11. The shielding polypeptide comprises from 2 to 68 such units, which may be the same or different, within the definition set out above. The units of the shielding polypeptide may also be referred to as“repeating units” (or“repeat units”) although there is some variation in the amino acid sequence between individual units, and hence“repeating units” is not to be understood exclusively as the repetition of one and the same sequence. Stated

differently, the shielding polypeptide moiety comprises from 2 to 68 units, wherein each unit is an amino acid sequence independently selected from the group consisting of the individual sequences falling within the definition of SEQ ID NO:1.

In embodiments of the invention, the shielding polypeptide moiety may comprise at least 3 at least 4, at least 6, at least 8, at least 10, or at least 17 units of one or more amino acid sequence(s) according to SEQ ID NO: 1. Furthermore, in embodiments of the invention, the shielding polypeptide moiety may comprise up to 3, up to 5, up to 9, up to 10, up to 17, up to 18, up to 34, or up to 51 units of one or more amino acid sequence(s) according to SEQ ID NO: 1. Thus for example, the shielding polypeptide moiety may comprise from 3 to 51 units of one or more amino acid sequence(s) according to SEQ ID NO: 1 , such as 3 to 34 units, such as 3 to 17 units, such as 5 to 17 units independently selected from the group consisting of SEQ ID NOs: 2-11 and 159-230, such as from the group consisting of SEQ ID NOs: 2-11.

The number of repeating units of the shielding polypeptide may be made with regards to the specific lysosomal polypeptide to which it is to be fused.

Generally, the number of N-glycosylation sites of the lysosomal polypeptide and their location in three-dimensional space, may be considered when selecting the desired number and length of the shielding polypeptide(s). For instance, many N-glycosylation sites, which are positioned far from each other, may require use of a longer (higher number of units) shielding polypeptide, and/or the use of more than one shielding polypeptide, to achieve a certain shielding effect (reduced rate of receptor-mediated cellular uptake). It is also envisaged that monomeric protein complexes may benefit from more and/or longer shielding polypeptide(s) to achieve a certain shielding effect, compared to dimeric protein complexes where the shielding polypeptides of the second fusion protein may to some extent shield the N- glycans of the lysosomal polypeptide of the first fusion protein. When the lysosomal polypeptide is Iduronate 2-sulfatase, it may be desirable that the shielding polypeptide moiety comprises at least 17 contiguous repeats, and preferably a fusion protein of iduronate-2-sulfatase comprises two shielding polypeptide moieties, each of which comprises at least 17 contiguous units of SEQ ID NO: 1.

The shielding polypeptide moiety may comprise a contiguous sequence of at least 18 amino acids (corresponding to two units that are both 9-meric versions of SEQ ID NO:1 ), and typically up to 748 amino acids

(corresponding to 68 units which are all 11 -mer versions of SEQ ID NO:1 ). The repeating units may be contiguous with one another, although it is also possible that the repeating units are separated by short spacing sequences. For instance, two repeating units may be separated by up to 10 amino acid residues that do not correspond to SEQ ID NO: 1 ; for instance, the short spacing sequence may be a peptide linker of the formula (G ₄ S) ₂ . In some embodiments, a spacing sequence may be up to 5 amino acid residues. In some embodiments one or more amino acid residue(s) may be positioned between two repeating units, e.g. to impart a desired functionality such as an N-glycosylation site, or to provide a site for another type of modification, for instance employing a single Cys residue. In some embodiments, a linker, such as one or more G ₄ S linkers, may be used as spacing sequences between adjacent repeating units. Hence, in view of this possibility, the contiguous sequence comprising up to 68 repeating units may be longer than 748 amino acids, for instance up to 800 amino acids.

The repeating units of the shielding polypeptide moiety are defined by SEQ ID NO: 1 , which is based on the repeating units of human variants of the BSSL C-terminal domain, and which allows some variation of amino acid residues in amino acid positions 3, 4, 5, 6, 8, 9, 10 and 11. In contrast, the residues at positions 1 (X1 ), 2 (X2) and 7 are fixed, although X1 and X2 may be absent. Typically, both X1 and X2 are absent, and in such embodiments, a repeating unit consists of 9 amino acids only.

A shielding polypeptide moiety comprising 2 to 68 units (repeating units) typically comprises several variants of the amino acid sequence motif generally defined by SEQ ID NO:1 , such as at least two different variants according to SEQ ID NO:1. For instance, in embodiments of the invention where the shielding polypeptide moiety comprises at least 4 units, it may comprise at least one unit of each of SEQ ID NO:3, SEQ ID NO:4 and SEQ ID NO:5. In embodiments of the invention where the shielding polypeptide moiety comprises at least 2 units, these may be independently selected from the group consisting of SEQ ID NO:3, SEQ ID NO:4 and SEQ ID NO:5.

Advantageously, the shielding polypeptide moiety may comprise SEQ ID NOs: 3-5 in this order, optionally preceded by SEQ ID NO: 2. A unit according to SEQ ID NO: 2 may especially be located at the N-terminal end of the shielding polypeptide moiety, representing the first unit of the shielding polypeptide moiety. While other specific variations of the repeating units (e.g. the units according to SEQ ID NOs: 3-11 ) may appear repeatedly, SEQ ID NO: 2, if present, typically only appears once, as the first repeating unit of the shielding polypeptide moiety.

The conformation of the shielding polypeptide moiety is generally

unstructured. For instance, in embodiments of the invention, the shielding polypeptide does not contribute to the a-helix and/or b-sheet content of the fusion protein as determined by circular dichroism or FTIR (Fourier Transform Infrared Spectroscopy).

In embodiments of the invention, a repeating unit defined by SEQ ID NO:1 is of human origin, and preferably all of the repeating units of the shielding polypeptide moiety correspond(s) to naturally occurring repeating units of a variant of the C-terminal domain of human BSSL. Such repeating units are represented by SEQ ID NOs: 2-11 (See also Example 1 , Table 1.2). In embodiments of the invention, all repeating units of the shielding polypeptide moiety are selected from the group consisting of SEQ ID NOs: 2-11 , e.g. SEQ ID NOs: 3-11. That is, the shielding polypeptide moiety may comprise 2-68 units, each independently selected from the group consisting of SEQ ID NO: 2-11 , or from the group consisting of SEQ ID NOs: 3-11. The use of a sequence of human origin may be advantageous as it is expected to contribute to a low immunogenicity in human subjects.

Furthermore, in embodiments of the invention, the shielding polypeptide moiety comprises, or consists of, a sequence of repeating units that corresponds to a naturally occurring human sequence of repeating units. Examples of such natural human sequences of repeating units are presented in SEQ ID NO: 12-19 and 20-27. Such sequences may comprise, as the first five repeating units, in this order: [SEQ ID NO: 2] - [SEQ ID NO: 3] - [SEQ ID NO: 4] - [SEQ ID NO: 5] - [SEQ ID NO: 5], or, alternatively, as the first four repeating units, in this order: [SEQ ID NO: 3] - [SEQ ID NO: 4] - [SEQ ID NO: 5] - [SEQ ID NO: 5]. In other embodiments, the first repeating unit may be SEQ ID NO: 5. In embodiments where the shielding polypeptide comprises only three repeating units, these may in particular be, in this order [SEQ ID NO: 2] - [SEQ ID NO: 3] - [SEQ ID NO: 4] or [SEQ ID NO: 3] - [SEQ ID NO: 4] - [SEQ ID NO: 5]. In embodiments where the shielding polypeptide moiety consists of five repeating units, these may in particular be: [SEQ ID NO: 3] - [SEQ ID NO: 4] - [SEQ ID NO: 5] - [SEQ ID NO: 5] - [SEQ ID NO: 5] In yet other embodiments, the shielding polypeptide moiety may comprise, or consist of, 9 identical repeats SEQ ID NO: 5, which may be contiguous.

In embodiments of the invention, the shielding polypeptide moiety comprises an amino acid sequence according to any one of SEQ ID NOs: 12-27. In some embodiments the shielding polypeptide moiety consists of a multiple of any one of the sequences of the group comprised of SEQ ID NOs: 12-27. For instance, the shielding polypeptide moiety may consist of two or more, such as three or more, contiguous multiples, or copies, of an amino acid sequence according to any one of SEQ ID NOs: 12-27; for instance SEQ ID NO: 20. SEQ ID NO: 20 comprises 17 units of an amino acid sequence according to SEQ ID NO: 1 , and thus a three-copy multiple of SEQ ID NO: 20 would comprise at least 51 units.

In some embodiments, the shielding polypeptide moiety comprises an amino acid sequence that corresponds in part or in full to any one of SEQ ID NOs: 12-27, such as SEQ ID NO: 20. In such embodiments, the shielding polypeptide moiety may contain from 3 to 17 contiguous repeating units. For instance, the shielding polypeptide moiety may have an amino acid sequence selected from among any one of SEQ ID NOs: 73-77.

It should be noted that the repeating units of the shielding polypeptide moiety can be independently selected from all units according to SEQ ID NO:1 and the invention is thus not limited to certain sequences of units being repeated. Accordingly, for instance a 51 -unit shielding polypeptide moiety is not necessarily formed of three copies of a 17-unit sequence, but may be formed of any combination of units according to SEQ ID NO:1 , and in particular of any combination of repeating units selected from SEQ ID NOs: 2-11 and 159- 230.

In some embodiments, in particular where the fusion protein is produced in mammalian cells, said one or more shielding polypeptide moieties may completely lack glycosylation sites. In such embodiments, one or more shielding polypeptide moieties of the fusion protein may each comprise 2 to 68 units of one or more amino acid sequence(s) independently selected from the group consisting of SEQ ID NOs: 159-230, such as from the group consisting of SEQ ID NOs: 159-165.

As mentioned above, the fusion protein may comprise a linker, typically a peptide linker, linking the lysosomal polypeptide to one or more shielding polypeptide moieties as described herein. Hence, in embodiments of the invention the fusion protein further comprises a peptide linker positioned between an amino acid sequence of the lysosomal polypeptide and an amino acid sequence of the shielding polypeptide moiety. For example, the peptide linker may be selected from -GS- and -(G4S) _n -, wherein n is an integer from 1 to 5, typically from 1 to 3, or from 2 to 3. The use of a linker may be

advantageous in that it may reduce the occurrence of, or, in the case of n being at least 2, prevent the formation of neo epitopes and subsequent binding of such neo epitopes by antigen-presenting cells of the immune system.

The fusion protein of the invention comprises a lysosomal polypeptide e.g. as described above, fused to a shielding polypeptide moiety as defined herein. In embodiments, the fusion protein comprises an amino acid sequence has at least 95 % sequence identity with an amino acid sequence selected from SEQ ID NO: 28-32, 34-36, 38-48, 50-60, 62-64, 66-72, 79-81 , and 150-158, such as at least 98 % sequence identity, or at least 99 % sequence identity, with an amino acid sequence selected from SEQ ID NO: 28-32, 34-36, 38-48, 50-60, 62-64, 66-72, 79-81 , and 150-158.

In a further embodiment, said lysosomal polypeptide may be extended by one or more C- and/or N-terminal amino acid(s), making the actual lysosomal polypeptide sequence longer than in the sequence referred to above

Similarly, in other instances the lysosomal protein may have an amino acid sequence which is shorter than the corresponding part of the relevant amino acid sequence referred to above, the difference in length e.g. being due to deletion(s) of amino acid residue(s) in certain position(s) of the sequence.

In an embodiment, the fusion protein is isolated.

In an embodiment, the lysosomal polypeptide is glycosylated. The fusion proteins described herein can be produced by recombinant techniques using eukaryotic, such as mammalian (including human), expression systems using conventional methods known to persons of skill in the art. The examples below describe cloning and production of fusion proteins in which shielding polypeptide moieties are fused to biologically active polypeptides, represented by lysosomal proteins. It should be noted that the invention is by no means limited to use of those strains and cell types of the examples; in contrast, suitable cell lines for production of fusion proteins are known to persons of skill in the art, and examples include Pichia pastoris, Saccharomyces cerevisiae, algae, moss cells, plant cells such as carrot cells, and mammalian cells such as CHO, HEK-293, and HT1080.

In one embodiment, said fusion protein is expressed in mammalian, Chinese hamster ovary, plant or yeast cells. The resulting protein is thus glycosylated by one or more oligomannose N-glycans.

In embodiments, the fusion protein may be recombinantly produced in a continuous human cell line.

Regarding the design of DNA constructions encoding the shielding

polypeptide moiety, it may be advantageous to use synthetic genes which utilize the redundancy of the genetic code by including different, or all, codon variants for each amino acid that is to be encoded. The use of more variable DNA sequences may facilitate characterization of the nucleic acid

components, as characterization of highly repetitive sequences may be problematic.

In addition to reducing the rate of receptor mediated uptake, the shielding polypeptide moiety may provide increased solubility to the fusion protein. In particular, the hydrophilic nature of the shielding polypeptide moiety, may be beneficial in that it may increase the bioavailability of a fusion protein that is administered subcutaneously, relative to the bioavailability of the lysosomal polypeptide alone. In such cases, the increased solubility of the fusion protein may promote transfer to the blood stream rather than remaining in the tissue extracellular matrix after injection. This could mean that for some lysosomal polypeptides that otherwise require intravenous administration due to limited bioavailability, subcutaneous administration may be a possibility if the lysosomal polypeptide is fused to a shielding polypeptide moiety as described herein.

As described above, by preventing rapid uptake via the glycan-dependent routes, clearance from the circulation is significantly reduced slower cellular uptake processes are favored, which results in a different distribution profile of the lysosomal polypeptide. This may enable distribution of therapeutic fusion proteins to cells of tissues that are normally poorly exposed to lysosomal enzymes.

In an embodiment, the fusion protein distributes to peripheral tissue when administered to a mammal. Examples of peripheral tissue are given above. Moreover, the lysosomal protein may display biologic activity, such as retained enzymatic or catalytic activity, in said peripheral tissue.

In some embodiments, the fusion protein according to aspects described herein may distribute to the brain when administered to a mammal, and may also display (retained) biological activity, such as retained enzymatic or catalytic activity, in the brain of said mammal. In one embodiment, the fusion protein has enzymatic activity in the brain.

By“retains the biological activity” or“retained biological activity” of the lysosomal polypeptide is meant that the biological activity of the fusion protein is retained fully or partly from the lysosomal polypeptide as such. For instance, the fusion protein may retain at least 25 %, preferably at least 50 %, and more preferably at least 75 %, of a biological activity or effect from the lysosomal polypeptide as such. To avoid complete loss of activity of a lysosomal protein upon fusion with the shielding polypeptide moiety, the fusion should be carried out carefully. The position of fusion must not alter the functional epitope or the active site of the lysosomal polypeptide such that the resulting fusion protein becomes inactive. Thus, the fusion protein as disclosed herein may affect lysosomal storage in the brain, visceral organs or peripheral tissue of mammals, such as to decrease lysosomal storage, for example lysosomal storage of lipids, GAGs, glycolipids, glycoprotein, amino acids or glycogen. Surprisingly, it was found that a fusion protein according to some

embodiments of the invention had not only a fully retained biological activity, but in fact had an increased or improved biological activity, compared to the lysosomal polypeptide as such. In particular such an increased biological activity was observed for fusion proteins of iduronidase and iduronate-2- sulfatase, respectively, with shielding polypeptide moieties. Not wishing to be bound by any particular theory, it is hypothesized that an increased biological activity may be due to increased solubility of the fusion protein and/or improved post-translational modification of the fusion protein, resulting in the production of a higher quality and/or more“activated” protein. Whatever the cause, the provision of a highly active protein offers a possibility of

formulating a more potent therapeutic.

In one aspect, there is provided a composition, comprising a fusion protein where a lysosomal polypeptide is fused to one or several shielding

polypeptides, and optionally one or more carriers or ingredients. Examples of suitable carriers and ingredients include conventional pharmaceutically acceptable substances such as water, a buffer solution, a salt, a pH regulator, an oil, a preservative, an osmotically active agent, and any combination thereof. The term“composition” as used herein should be understood as encompassing solid and liquid forms. A composition may preferably be a pharmaceutical composition, suitable for administration to a patient (e.g. a mammal) for example by injection or orally. The pharmaceutical composition may be formulated for any route of administration, including intravenous, subcutaneous, nasal, oral, and topical administration. In particular the pharmaceutical composition may be formulated, or intended to be formulated, for intravenous or subcutaneous administration.

The advantages disclosed for other aspects also apply to the composition aspect. Similarly, the embodiments disclosed for other aspects also apply to the composition aspect. In particular, the embodiments related to content of glycan moieties, protein activity, and particular examples of lysosomal proteins (see Table I and lists above) are applicable also to this aspect.

The fusion protein of the invention may be used in a method of treatment of a lysosomal storage disease, comprising the step of administering to a patient suffering from a lysosomal storage disease a fusion protein comprising a lysosomal polypeptide fused to a shielding polypeptide moiety as described herein. The patient is typically a mammal, such as a human. In this method, the administering may be effected by intravenous infusion of a duration of at least 5 minutes, such as at least 15 minutes, and up to 6 hours, such as 4 hours or up to 2 hours. As an example, administration may occur at a frequency of once daily to once monthly, for instance once weekly.

The invention will be further described in the following non-limiting examples.

Examples

Example 1A: Identification of repeating units of human origin

A blast search was performed with the catalytic domain of Bile salt-stimulated lipase (BSSL) versus the non-redundant protein sequence database at the National Institute of Health (NIH), USA and identified 10 reported protein sequences for the protein of human origin that contained the whole or part of the C-terminal repetitive unstructured domain.

Material and methods

Blast at NIH was used to search for proteins of human origin that match the catalytic domain of Bile salt stimulated lipase with UniProt ID P19835

(Accession number CEL_HUMAN).

Results

The BLAST search resulted in finding 10 entries that contained both a significant portion of the catalytic domain and the C-terminal repetitive unstructured domain. The number of the repeating units in the domains differed and some variability among the sequence of the repeating units was noted, see Table 1.1 for the different hits. Each repeating domain is initiated by a truncated sequence of 9 residues, while the most prevalent repeating units are 11 residues long. In the table below, the repeating units are separated by a "~" sign for clarity. In the enclosed sequence listing, the repetitive portions are represented by SEQ ID NOs: 12-19. Table 1.1. Variants of human BSSL-CTD

Table 1.2 below lists the unique sequences of repeating units of human origin, with reference to the sequence identity number in the enclosed sequence listing. Absent residues of the first sequence are marked by a dash. Table 1 .2. Units corresponding to repeating units found in human BSSL-CTD.

Hence, there exists a variety of lengths of the C-terminal domain in the human population. Furthermore, the order of the repeating units can vary in the human population. This could imply that variations in the order of the repeating units and the length of the entire domain motifs are allowed.

The most prevalent human form is made up of the combination of the following sequence of repeating units:[SEQ ID NO: 2] - [SEQ ID NO: 3] - [SEQ ID NO: 4] - [SEQ ID NO: 5] - [SEQ ID NO: 5] - [SEQ ID NO: 5] - [SEQ ID NO: 5] - [SEQ ID NO: 5] - [SEQ ID NO: 5] - [SEQ ID NO: 5] - [SEQ ID NO: 5] - [SEQ ID NO: 6] - [SEQ ID NO: 5] - [SEQ ID NO: 5] - [SEQ ID NO:

7] - [SEQ ID NO: 5] - [SEQ ID NO: 8] - [SEQ ID NO: 9], or expressed differently: [SEQ ID NO: 2] - [SEQ ID NO: 3] - [SEQ ID NO: 4] - [SEQ ID NO: 5]x8 - [SEQ ID NO: 6] - [SEQ ID NO: 5]x2 - [SEQ ID NO: 7] - [SEQ ID NO: 5] - [SEQ ID NO: 8] - [SEQ ID NO: 9]

Example 1B: Design of repeating units lacking O-glycosylation sites

This example describes sequences of repeating units without O-glycan sites in the sequence. By utilizing the variable positions in SEQ ID NO: 1 the following sequences were designed that lacked serine or threonine that could be O-glycosylated during cultivation in eukaryotic expression systems such as CHO, HEK or yeast. The sequences are listed in Table 1.3 below. Table 1.3. Units corresponding to the general formula in SEQ ID NO: 1 without serine or threonine present.

Other possible units that lack O-glycan site are listed in Table 1.4 below.

Table 1.4. Units corresponding to the general formula in SEQ ID NO: 1 without serine or threonine present.

Example 2: Expression and purification of N-sulfoglucosamine sulfohydrolase (sulfamidase) fusion proteins with the shielding polypeptide

Materials and methods

DNA constructions: DNA sequences encoding a set of fusion proteins including shielding polypeptide moieties were codon optimized for expression in Chinese hamster ovary (CHO) cells and synthesized by the Invitrogen GeneArt Gene Synthesis service at Thermo Fisher Scientific. The genes were cloned into expression vectors suitable for expression in mammalian cells. The encoded proteins are presented in Table 2.

Cultivation and purification: Recombinant proteins were expressed using the ExpiCHO expression system (Thermo Fisher Scientific), essentially according to the manufacturer’s protocol. ExpiCHO cells were transfected with expression vectors with synthetic genes for the encoded proteins listed in Table 2. Supernatants were harvested by centrifugation 6-9 days after transfection of expression vectors and stored at -70°C.

Supernatants were thawed and filtered before purification. Each recombinant fusion protein was purified from the supernatant using conventional chromatography methods. Recombinant fusion proteins for use in animal studies were also subjected to an endotoxin removal step. Purified fusion proteins were formulated in 50 mM L-Arginine, 75 mM NaCI, 2% Sucrose, pH 7.8. The purity of the fusion proteins was analyzed by SDS-PAGE

(NuPAGE™ 4-12% Bis-Tris protein gels (Thermo Fisher Scientific)) stained with InstantBlue (Thermo Fisher Scientific). The molecular mass of each protein was analyzed by mass spectrometry (LC-MS or MALDI-TOF/MS) and multi-angle light scattering (MALS), and apparent size by size exclusion chromatography (SEC), as described in Example 3 below. Identity of sulfamidase in the fusion proteins was determined by peptide map LC-MS for a subset of the proteins.

In the following,“shielding polypeptide moiety” is abbreviated“SPM”.“SPM1 - 3” denotes a shielding polypeptide moiety formed of the first three units (units

1 -3) of a repetitive human BSSL-CTD sequence. By analogy,“SPM1 -5” denotes a shielding polypeptide moiety formed of the first five units (units 1 -5) of a repetitive human BSSL-CTD sequence,“SPM1 -17” denotes a shielding polypeptide moiety formed of the first 17 units (units 1 -17) of a repetitive human BSSL-CTD sequence, and“SPM4-12” denotes a shielding

polypeptide moiety formed of the fourth to the twelfth unit (units 4-12) of a repetitive human BSSL-CTD sequence.

Table 2. Sulfamidase and Sulfamidase fusion protein sequences

Results

The fusion proteins of sulfamidase with shielding polypeptides were expressed well in ExpiCHO and purifications resulted in protein preparations of high purity, assessed by SDS-PAGE as seen in Figure 2, in which: Lane A is molecular weight marker SeeBlue plus 2 (Thermo Fisher Scientific), lane B is sulfamidase-SPM1 -17 (PSI0438), lane C is SPM1 -5-Sulfamidase-SPM1 -17 (PSI0618), lane D is SPM1 -5-Sulfamidase-SPM4-12 (PSI0619), lane E is SPM4-12-Sulfamidase-SPM4-12 (PSI0620), lane F is sulfamidase-SPM4-12 (PSI0518) and lane G is sulfamidase (PSI0222).

The correct identity of sulfamidase in the fusion proteins was confirmed by high amino acid coverage peptide map mass spectrometry analysis for PSI0438, PSI0618, PSI0619, PSI0620, PSI0518 and PSI0222. Conclusions

Fusion proteins of sulfamidase with shielding polypeptide moieties can be produced by construction of synthetic genes, followed by expression in mammalian cell systems and purification to high purity using conventional techniques.

Example 3: Altered biophysical properties of N-sulfoglucosamine

sulfohydrolase (sulfamidase) fusions with shielding polypeptides compared to un fused sulfamidase

This example describes the biophysical characterization of fusion proteins of sulfamidases with shielding polypeptide moieties, using unfused sulfamidase as reference. Apparent size, molecular mass and Stokes radius in solution were determined by SEC/MALS, and aggregation propensity as a function of increased temperature by static light scattering (SLS).

Materials and methods

The apparent size of the fusion proteins and unfused proteins in solution was assessed by analytical SEC on an AKTA Micro (GE Healthcare Life Sciences) using a calibrated Superdex 200 Increase 3.2/300 column (GE Healthcare Life Sciences). The column was calibrated with Gel Filtration Calibration Kit LMW (code no. 28-4038-41 , GE Healthcare Life Sciences) and Calibration Kit HMW (code no. 28-4038-42, GE Healthcare Life Sciences), containing 8 globular proteins in the size range of 6 to 669 kDa and Blue Dextran 2000, using a running buffer of 25 mM NaP and 125 mM NaCI, pH 7.0 and a flow rate of 75 mI/min at a temperature of 25 °C. The corresponding size and hydrodynamic radius in solution can be calculated from the elution volume of a protein on a calibrated column by the methods described in appendix 10 of Handbook of Size Exclusion Chromatography Principles and Methods (order no 18-1022-18, GE Healthcare Life Sciences).

The molecular mass of the proteins was determined by MALS using a static light scattering detector miniDAWN TriStar and the Astra 5 software (Wyatt Technology Europe, Germany) connected to an Akta Micro (GE Healthcare Life Sciences) with a Superdex 200 Increase 3.2/300 column (GE Healthcare Life Sciences) using a column temperature of 25 °C, a running buffer of PBS, pH 7.4 and a flow rate of 75 mI/min. Aggregation propensity as function of increased temperature was determined by static light scattering (SLS) using UNcle (Unchained Labs). Thermal aggregation was measured through 90 degree (SLS) at two different wavelengths (266 and 473 nm). The samples were diluted to a concentration of 1 mg/ml with 50 mM Arg, 75 mM NaCI, 2% sucrose pH 7.8 and

temperature was ramped from 20 to 95 °C at 0.5 °C/min.

Results

Table 3 lists the values obtained for apparent size, molecular mass and number of repeat units in the fusion protein. Figure 3 shows the aggregation propensity (SLS counts) as function of increased temperature (°C). In this figure, sulfamidase (PSI0222) is represented by solid black line, whereas sulfamidase fusion proteins (PSI0438, PSI0618, PSI0619, PSI0620 and PSI0518) are represented each by a dashed line.

Table 3. Biophysical characterization of sulfamidase and sulfamidase fusion proteins

Conclusions

For the fusion proteins, which included shielding polypeptides, the apparent size in solution was larger than expected for globular proteins of similar masses. A correlation between number of repeat units in the fusion proteins and their size in solution was observed. The increase in apparent size in solution was independent of the position of shielding polypeptide - the apparent size in solution to the same extent. From the aggregation propensity analysis it can be concluded that unfused sulfamidase (i.e., without shielding polypeptide) aggregates at a lower temperature compared to the fusion proteins.

Example 4: Retained formyl glycine content in active site of sulfamidase fusion proteins with shielding polypeptides

Sulfamidase activity is dependent on post-translational conversion of cysteine in the active site of the protein (Cys 50) into a formyl glycine.

Material and methods

Determination of relative amount of formyl glycine in active site of the protein was performed by LC-MS analysis. Sulfamidase and sulfamidase fusion proteins with shielding polypeptides (20 mg) were reduced, alkylated and digested with trypsin. The protein was dissolved in a denaturing buffer (6 M Gua-HCI, 0.2 M Tris, 3mM EDTA, pH 8.3) during the reduction and alkylation steps. Reduction of the protein was done by incubation with DTT (10 mM) at room temperature (RT) for 1 h. Subsequent alkylation with iodoacetamide (55 mM) was performed at RT and in darkness for 45 min. Prior digestion the buffer was exchanged to 50 mM Tris 5mN CaCI2 pH 7.5 by dialysis. Lastly, the tryptic digestion was performed by addition trypsin to a protease: protein ratio of 1 :20 (w/w)). Digestion was allowed to take place over night at 37 °C. The resulting tryptic peptides containing cysteine 50 variants (cysteine50 (alkylated), oxidized cysteine 50, FGIy50 and Ser50) were all semi-quantified using peak area calculations from reconstructed ion chromatograms.

Results

The relative formyl glycine content of sulfamidase fusion proteins with shielding polypeptides is in a range of 0.8-2.2 of sulfamidase. The relative formyl glycine content of sulfamidase fusion proteins to sulfamidase is tabulated in Table 4.

Table 4. Relative formyl glycine (FGIy) content to sulfamidase

Conclusion

The sulfamidase fusion proteins with shielding polypeptides display a similar formyl glycine content as sulfamidase. The formyl glycine conversion in the active site is not compromised by the fusion of shielding polypeptides to sulfamidase in the fusion proteins.

Example 5: Reduced In vitro glycan receptor mediated cellular uptake of sulfamidase fusion proteins with shielding polypeptides

This example demonstrates cellular uptake of sulfamidase and sulfamidase fusion proteins and effect on uptake by inhibition of the M6PR by addition of excess M6P.

Materials and methods

Mouse embryonic fibroblasts (MEF-1 ) were seeded the day before treatment with proteins. The proteins were diluted in cell media in a concentration range between 2 and 15 nM for sulfamidase, and between 2 and 75 nM for the sulfamidase fusion proteins. The cells were incubated approximately 24 hours, washed in PBS and harvested in lysis buffer with protease and phosphatase inhibitors. The inhibition experiments with M6P were performed according to the same procedure as described above. The concentrations of sulfamidase and of the sulfamidase fusion proteins tested were 2 and 7.5 nM. The concentration range of M6P was 0.3 - 2.5 mM. Cell lysates were analyzed for sulfamidase and sulfamidase fusion protein content by an immunoassay using the Meso Scale Discovery (MSD) platform. Streptavidin coated MSD plates were blocked with 5% Blocker A in PBS. The plates were washed in PBS-T and different dilutions of standard and samples were distributed on the plate. A mixture of a biotinylated anti-Sulfamidase mouse monoclonal antibody and Sulfo-Ru-tagged rabbit anti-Sulfamidase antibodies were added and the plates were incubated at room temperature. Complexes of sulfamidase and labelled antibodies bind to the streptavidin coated plate via the biotinylated monoclonal antibody. After washing in PBS-T, the amount of bound complexes was determined by adding a read buffer to the wells and the plates were read in a MSD SI2400 instrument. The recorded

electrochemiluminescence counts were proportional to the amount of sulfamidase in the samples and evaluated against a relevant sulfamidase standard.

Results

The cellular uptake of the sulfamidase fusion proteins is tabulated in Table 5.1 as % decreased uptake compared to sulfamidase without shielding polypeptide. The cellular uptake in the presence of increasing concentration of M6P is listed in Table 5.2.

Table 5.1. Decrease in cellular uptake of sulfamidase fusion proteins compared to sulfamidase in MEF-1 cells

Table 5.2. Cellular uptake of sulfamidase and sulfamidase fusion proteins in MEF-1 cells in the presence of M6P.

Conclusions

The sulfamidase fusion proteins showed lower cellular uptake compared to sulfamidase alone. For fusion proteins with C-terminal shielding polypeptides the number of repeat units correlated to reduced uptake. For fusion proteins with both N- and C-terminal shielding polypeptides the cellular uptake was further reduced compared to fusion proteins with C-terminal shielding polypeptides only.

As demonstrated by M6P inhibition of sulfamidase uptake, the M6PR is indicated as the major endocytotic receptor for sulfamidase uptake in MEF-1 cells. The low uptake of fusion proteins indicates that the shielding

polypeptides reduce the M6PR-mediated uptake of sulfamidase fusion proteins. By adding M6P the uptake of unfused protein can be lowered to a level close to that of the fusion proteins without any M6P addition. The remaining cellular uptake for sulfamidase fusion proteins can be further reduced by inhibiting the M6PR.

Example 6: Shielding polypeptides interfere with sulfamidase binding to mannose-6-phosphate receptor

This example demonstrates the binding of sulfamidase fusion proteins to M6PR in vitro. Binding properties are compared to properties of unfused sulfamidase (i.e., without shielding polypeptide).

Materials and methods

Interactions between sulfamidase fusion proteins and M6PR were analyzed by SPR technology using a Biacore T200. The extracellular part of M6PR (human ciM6PR, R&D systems) was used as ligand and was immobilized by amine coupling on a CM5 Biacore chip. Analyte concentrations were kept constant at 500 nM with an association phase of 3 min and a dissociation phase of 2.5 min.

Results

M6PR was coupled to the chip with a total coupling response of 7967 RU. All sulfamidase fusion proteins showed interaction with immobilized M6PR although with a significantly reduced response magnitude as compared to sulfamidase. By the end of the 3 minute association phase binding had reached equilibrium and responses at this time point is found in Table 6. Table 6. Interactions between M6PR and sulfamidase proteins with shielding polypeptides using SPR technology

Conclusions

Fusion proteins of sulfamidase with shielding polypeptides show significantly reduced interaction with the M6PR receptor as compared to sulfamidase alone. Fusion of shielding polypeptides to both N- and C-terminus of sulfamidase is particularly beneficial for effective blocking of the M6PR interaction driven by the phosphorylated high mannose glycans of

sulfamidase.

Example 7; Fusion proteins of sulfamidase with shielding polypeptides show increased serum exposure and improved distribution to the brain in mice Material and methods

In this example, serum exposure and distribution to CNS of sulfamidase and fusion proteins of sulfamidase with shielding polypeptides were investigated in mice (C57BL/6J). The proteins were produced as described in Example 2. The mice were given an intravenous single dose administration in the tail vein of 10 mg/kg. Sulfamidase and fusion proteins of sulfamidase with shielding polypeptides, respectively, were formulated at 2 mg/mL in 50 mM arginine, 75 mM NaCI and 20 mg/mL sucrose, pH 7.8 and administered at 5 mL/kg. Two non-terminal blood samples were taken from vena saphena at different time points with 3 mice per time point. At termination, the mice were anaesthetized by isoflurane and blood was collected from the orbital plexus. Blood samples were allowed to clot at room temperature, centrifuged at 1200 xg at 4 °C for 10 min. Serum samples were stored -70 °C until bioanalysis. Time-points collected were 5 min, 30 min, 1 , 2 (terminal), 4, 8, 24 (terminal), 32 and 48 (terminal) hours after the dose with a total of nine mice per fusion protein. At termination and after the last blood sample was taken, animals were subjected to cardiac perfusion with cold 0.9% saline. Brain was dissected, weighed and frozen rapidly in dry ice-chilled isopentane. Brain homogenates were prepared in buffer (29 mM diethylbarbituric acid, 29 mM sodium acetate, 0.68 % (w/v) NaCI, pH 6.5) using a Lysing Matrix D device (MP Biomedicals, LLC, Ohio, US). Homogenization was performed for 25 s in a Savant

FastPrep FP120/Bio101 homogenizer 30 (LabWrench, ON, Canada) and the homogenate was subsequently centrifuged in an Eppendorf centrifuge 5417R at 10,000 ref at 10 °C for 10 min.

The serum and brain homogenate levels of sulfamidase and sulfamidase fusion proteins were analyzed by an immunoassay using the Meso Scale Discovery (MSD) platform. Streptavidin coated MSD plates were blocked with 5% Blocker A in PBS. The plates were washed and different dilutions of standard and samples were distributed on the plate. A mixture of a

biotinylated anti-Sulfamidase mouse monoclonal antibody (mAB) and Sulfo- Ru-tagged rabbit anti-Sulfamidase antibodies was added and the plates were incubated at RT. Complexes of sulfamidase and labelled antibodies bind to the streptavidin coated plate via the biotinylated mAb. After washing, the amount of bound complexes was determined by adding a read buffer to the wells and the plates were read in a MSD SI2400 instrument. The recorded electrochemiluminescence counts were proportional to the amount of sulfamidase in the samples and evaluated against a relevant sulfamidase standard. The area under the serum concentration versus time curve extrapolated to infinity (AUC~) corrected for dose was calculated using Phoenix WinNonlin software version 8 (Certara, U.S.A.) by non- compartmental analysis.

Results

Fusion proteins of sulfamidase with shielding polypeptides show increased AUC~ as compared to sulfamidase, see Figure 4 and Tables 7.1 and 7.2. Figure 4 plots the mean and standard deviation serum concentration versus time following a 10 mg/kg i.v. administration in male C57BL/6 mice of sulfamidase (PSI0222 as filled boxes) and sulfamidase fusion proteins (PSI0438: open boxes; PSI0618: open circles; PSI0619: filled triangles).

Further, whereas concentration of sulfamidase in brain homogenate was below limit of quantification at 2h, 24h and 48h after dose (0.26 nM), the fusion proteins showed concentrations in brain homogenates higher than the limit of quantification obtained for sulfamidase (PSI0222) at least at 2 hours post dose, with SPM1 -5-Sulfamidase-SPM1 -17 (PSI0618) up to 24 h post dose and SPM1 -5-Sulfamidase-SPM4-12 (PSI0619) up to 48 h post dose. Table 7.1. Serum exposure (AUC ) and mean concentration in brain homogenate following a 10 mg/kg i.v. administration of sulfamidase and sulfamidase fusion proteins at 2, 24 and 48 hours after the dose

BLOQ = be ow LOQ 0.26 nM for sulfamidase based on average brain weight in this group. If two or more values were below LOQ, no mean is reported.

Table 7.2. Dose normalized serum exposure (AUC¥/Dose) and mean dose normalized concentration in brain homogenate following a 10 mg/kg i.v. administration of sulfamidase and sulfamidase fusion proteins at 2, 24 and 48 hours after the dose

BLOQ = below LOQ, 1.4 g/L for sulfamidase based on average brain weight in this group. If two or more values were below LOQ, no mean is reported.

Conclusion

Fusion proteins of sulfamidase with shielding polypeptides show increased AUC~ as compared to sulfamidase alone, at least partly due to the inhibition of receptor mediated uptake in peripheral tissue. Fusion proteins containing sulfamidase and shielding peptides showed significant improved distribution to brain.

Example 8: Expression and purification of fusion proteins of iduronate 2- sulfatase (idursulfase) with shielding polypeptides

Materials and methods

DNA constructions: DNA constructions were done as in Example 2. The encoded proteins are presented in Table 8. Table 8. Iduronate 2-sulfatase (idursulfase) and idursulfase fusion proteins

Cultivation and purification: ExpiCHO cells were transfected with expression vectors and cultured as described in Example 2. Purifications were performed as in Example 2. Purified fusion proteins were formulated in PBS pH 7.4. Identity of idursulfase in the fusion proteins was determined by peptide map LC-MS analysis for a subset of proteins.

Results

The fusion proteins expressed well in ExpiCHO and purifications resulted in protein preparations with high purity as assessed by SDS-PAGE. The correct identity of idursulfase in the fusion proteins was confirmed by high sequence coverage peptide map mass spectrometry analysis for the following proteins: PSI0593, PSI0692, PSI0693, PSI0694, PSI0695, PSI0696 and PSI0697.

Conclusions

Fusion proteins of idursulfase with shielding polypeptide moieties can be produced by construction of synthetic genes, followed by expression in mammalian cell systems and purification to high purity using conventional techniques.

Example 9: Altered biophysical properties of idursulfase fusion proteins with shielding polypeptides compared to un fused idursulfase

In this example, the biophysical characterization of fusion proteins of idursulfase are described, using unfused idursulfase (i.e., without shielding polypeptide) as reference. Apparent size, molecular mass and Stokes radius in solution were determined by SEC/MALS.

Materials and methods

Performed as in Example 3. For determination of molecular size and mass for idursulfase-SPM1 -5 (PSI0692), idursulfase-SPM1 -17 (PSI0693), SPM1 -5- idursulfase-SPM1 -5 (PSI0694), SPM1 -17-idursulfase-SPM1 -17 (PSI0695), SPM1 -5-idursulfase (PSI0696) and SPM1 -17-idursulfase (PSI0697), a Superdex 200 Increase 10/300 GL column (GE Healthcare Life Sciences), calibrated as in Example 3, was used.

Results

Table 9 lists the values obtained for apparent size and molecular mass and the number of repeat units in the fusion proteins. Figure 5 shows the relationship between the apparent molecular weight in solution (y-axis; in kDa) of the fusion proteins and number of repeat units of the shielding polypeptide moiety (x-axis). Table 9. Biophysical characterization of idursulfase and idursulfase fusion proteins

Conclusions

For the fusion proteins the apparent size in solution was larger than expected for globular proteins of similar masses. A correlation between number of repeat units in the fusion proteins and their size in solution was observed. The increase in apparent size in solution was independent of the position of the shielding polypeptide(s) (N-terminal or C-terminal or both) - the apparent size in solution increased to the same extent.

Example 10: Retained catalytic activity of fusion proteins of id ursu If ase with shielding polypeptides

Idursulfase activity is dependent on post-translational conversion of a cysteine (cys 84) in the active site of the protein into a formyl glycine. The specific activity of the protein can be measured by the method described in this example.

Material and methods

Idursulfase enzymatic activity was assessed using a specific substrate 4- methylumbelliferyl-a-L-iduronate 2-sulphate at 0.1 mM in 0.1 M sodium acetate, pH 4.5 at 37°C (Voznyi et al 2001 , J Inherit Metab Dis 24:675-680). Aldurazyme (iduronidase) was added in an excess at 5 mU/mL as a coupling enzyme to release 4-methylumbelliferone (4-MU). The reaction was typically stopped after 35 minutes using an excess of 0.5 M sodium carbonate at pH 10.7 and the end product 4-MU formed was measured utilizing fluorescence intensity at X355/E460 nm.

Results

The activity of fusion proteins of idursulfase with shielding polypeptides was found to be in a range of 0.5- 1.7 relative to idursulfase and is tabulated in Table 10 below.

Conclusion

The idursulfase fusion proteins display a similar in vitro activity as idursulfase. The in vitro activity is not compromised by the fusion of shielding polypeptides to idursulfase. Table 10. In vitro activity of idursulfase fusion proteins relative to idursulfase activity

Example 11: Reduced in vitro glycan receptor-mediated cellular uptake of fusion proteins of iduronate-2-sulfatase (idursulfase) with shielding

polypeptides

This example demonstrates cellular uptake of idursulfase and idursulfase fusion proteins and the effect on uptake by inhibition of the M6PR by addition of excess M6P.

Materials and methods

Experiments were performed as in Example 5 apart from the concentration of added protein to the cells and the immunoassay used for determining amount of protein taken up by the cells. Concentration of added protein was between 2 and 7.5 nM and the concentration of M6P was 0.2-1.3 mM. Cell lysates were analyzed for idursulfase and idursulfase fusion protein content by an immunoassay using the Meso Scale Discovery (MSD) platform. Streptavidin coated MSD plates were blocked. The plates were washed in PBS-T and a biotinylated goat anti-idursulfase polyclonal antibodies (pAb) was added to the wells and the plates were incubated at room temperature followed by wash in PBS-T. Different dilutions of standard and samples were distributed on the plates and the plates were incubated at room temperature followed by wash in PBS-T. Sulfo-Ru-tagged goat anti-idursulfase polyclonal antibodies was added and the plates were incubated at room temperature. Complexes of idursulfase and labelled antibodies bind to the streptavidin coated plate via the biotinylated pAb. After washing, the amount of bound complexes was determined by adding a read buffer to the wells and the plates were read in a MSD SI2400 instrument. The recorded electrochemiluminescence counts were proportional to the amount of idursulfase in the samples and evaluated against a relevant idursulfase standard.

Results

The decrease of cellular uptake of idursulfase fusion proteins compared to unfused idursulfase is tabulated in Table 11. The cellular uptake was reduced for fusion proteins compared to unfused idursulfase. Figure 6 shows the cellular uptake of idursulfase and idursulfase fusion proteins, respectively, in the presence of increasing concentration of M6P. In this figure, circles represent PSI0593 (idursulfase), squares represent PSI0693 and triangles represent PSI0659. Concentration of M6P in mM is shown on the x-axis; intracellular concentrations of idursulfase proteins in pM is shown on the y- axis. By adding increasing concentrations of M6P the uptake of unfused idursulfase can be reduced to a low level, similar to that of the fusion protein without any M6P addition. Table 11. Decrease in cellular uptake of idursulfase fusion proteins compared to idursulfase in MEF-1 cells.

Conclusion

The idursulfase fusion proteins showed decreased cellular uptake compared to unfused idursulfase, as shown in Table 11. A correlation of decreased cellular uptake with increasing number of repeat units was found. As demonstrated by M6P inhibition of idursulfase uptake, the M6PR is indicated as the major endocytotic receptor for idursulfase uptake in MEF-1 cells. The low uptake of fusion proteins indicates that the shielding polypeptides reduce the M6PR-mediated uptake of idursulfase fusion proteins. The remaining cellular uptake for idursulfase fusion proteins can be further reduced by inhibiting the M6PR.

Example 12: Shielding polypeptides interfere with idursulfase binding to mannose-6-phosphate receptor

This example demonstrates the binding of idursulfase fusion protein to M6PR in vitro. Binding properties are compared to properties of unfused idursulfase (i.e., without shielding polypeptide).

Materials and methods

Interactions between idursulfase fusion protein and idursulfase, respectively, with M6PR were analyzed by SPR technology essentially as in Example 6. Analyte concentrations were kept constant at 1000 nM with a association phase of 3 min and a dissociation phase of 3 min.

Results

M6PR was coupled to the chip with a total coupling response of 7967 RU. The idursulfase fusion proteins showed interaction with immobilized M6PR although with a significantly reduced response magnitude as compared to idursulfase. By the end of the 3 minutes association phase, binding had reached equilibrium and responses at this time point are found in Table 12. Table 12. Interactions between M6PR and idursulfase protein with shielding polypeptides using SPR technology

Conclusions

The fusion protein of idursulfase with shielding polypeptides shows significantly reduced interaction with the M6PR receptor as compared to idursulfase alone.

Example 13: Fusion proteins of idursulfase with shielding polypeptides show increased serum exposure and improved distribution to the brain in mice In this example serum exposure and distribution to CNS of idursulfase (SEQ ID NO: 37) and a fusion protein of idursulfase with shielding polypeptides (SEQ ID NO:46) were investigated in mice (C57BL/6J). Idursulfase and fusion protein were produced as described in Example 2. The mice were given a single intravenous dose of 10 mg/kg in the tail vein. Idursulfase or a fusion protein of idursulfase with shielding polypeptides were formulated at 2 mg/mL in 50 mM arginine, 75 mM NaCI and 20 mg/mL sucrose, pH 7.8 and administered at 5 mL/kg. Two non-terminal blood samples were taken from the sublingual plexus at different time points with 3 mice per time point. At termination, the mice were anaesthetized by pentobarbital, blood was collected from the sublingual plexus and CSF was obtained by needle puncture of the cisterna magna.

Blood samples were allowed to clot at room temperature, centrifuged at 1200 x g at 4 °C for 10 min. Serum samples were stored -70 °C until bioanalysis. Time-points of blood sampling were 5 min, 30 min, 1 , 2 (terminal), 4, 8, 24 (terminal), 32 and 48 (terminal) hours after the dose with a total of nine mice per protein. At termination and after the last blood sample and the CSF sample were taken, animals were subjected to cardiac perfusion with cold 0.9% saline. Brain was dissected, weighed and snap frozen in an air-tight cryo tube in liquid nitrogen. Brain homogenates were prepared in buffer (29 mM diethylbarbituric acid, 29 mM sodium acetate, 0.68 % (w/v) NaCI, pH 6.5) using tubes with Lysing Matrix D and a Homogenizer FastPrep-24™ 5G (MP Biomedicals, LLC, Ohio, US). The homogenates were subsequently centrifuged in an Eppendorf centrifuge 5417R at 10,000 ref at 10 °C for 10 min. and the supernatant was collected for concentration determinations.

The idursulfase or the idursulfase fusion protein concentrations in serum,

CSF and brain homogenate were analyzed by an immunoassay using the Meso Scale Discovery (MSD) platform. MSD Streptavidin Gold plates were blocked, and samples or standards were incubated with polyclonal anti- idursulfase capture (BAF2449, R&D Systems) and detection (AF2449, R&D Systems, conjugated with MSD SULFO-TAG containing reutenium) antibodies. The following day the plates were washed with PBS-Tween, and read in a MSD QuickPlex SQ120 using electrochemiluminescence for detection. For the serum samples, 1 % mouse serum was added to the diluent (1 % fish gelatin in PBS-Tween) to compensate for matrix effects. The recorded electrochemiluminescence counts were proportional to the amount of idursulfase in the samples and evaluated against the relevant idursulfase standard. The area under the serum concentration versus time curve extrapolated to infinity (AUC ) corrected for dose was calculated using Phoenix WinNonlin software version 8 (Certara, U.S.A.) by non- compartmental analysis.

Results

The fusion protein of idursulfase with shielding polypeptides showed an increased AUC~ in serum as compared to idursulfase, see Figure 9 and Tables 13.1 and 13.2. Figure 9 plots the mean and standard deviation serum concentration versus time following a 10 mg/kg i.v. administration in male C57BL/6 mice of (PSI0593, SEQ ID NO: 37, as filled boxes) and the idursulfase fusion protein (PSI0695, SEQ ID NO: 46, open boxes). Further, the idursulfase fusion protein (PSI0695) showed concentrations in CSF and brain homogenates higher than those obtained for idursulfase (PSI0593).

Table 13.1. Serum exposure (AUC ) and mean concentration in CSF and in brain homogenate following a 10 mg/kg i.v. administration of idursulfase and an idursulfase fusion protein at 2, 24 and 48 hours after the dose

Table 13.2. Dose normalized serum exposure (AUC¥/Dose) and mean dose normalized concentration in brain homogenate following a 10 mg/kg i.v. administration of idursulfase and an idursulfase fusion protein at 2, 24 and 48 hours after the dose

Conclusion

A fusion protein of idursulfase with shielding polypeptides showed an increased AUC~ in serum as compared to idursulfase alone, at least partly due to the inhibition of receptor mediated uptake in peripheral tissue. The fusion protein containing idursulfase and shielding peptides showed significant improved distribution to CSF and brain.

Example 14: Expression and purification of fusion proteins of arylsulfatase A (ASA) with shielding polypeptides

Materials and methods

DNA constructions: DNA constructions were done as in Example 2. The encoded proteins are presented in Table 12.

Table 12. Protein sequences of ASA and ASA fusion proteins

Cultivation and purification: ExpiCHO cells were transfected with expression vectors and cultured as described in Example 2. Purifications were performed as in Example 2. Purified fusion proteins were formulated in 25 mM NaP, 125 mM NaCI, pH 7.0.

Results

The fusion proteins expressed well in ExpiCHO and purifications resulted in protein preparations with high purity as assessed by SDS-PAGE. Conclusions

Fusion proteins of ASA with shielding polypeptide moieties can be produced by construction of synthetic genes, followed by expression in mammalian cell systems and purification to high purity using conventional techniques.

Example 15: Altered biophysical properties of ASA fusion proteins with shielding polypeptides compared to ASA

This example describes the biophysical characterization of fusion proteins of ASA with shielding polypeptide moieties, using unfused ASA as reference. Apparent size, molecular mass and Stokes radius in solution were determined by SEC/MALS. Aggregation propensity was determined by backlight scattering.

Materials and methods

Performed as in Example 3 with the exception that a Superdex 200 Increase 10/300 GL column (GE Healthcare Life Sciences), calibrated as in example 3, was used. Aggregation propensity as function of increased temperature was determined by backlight scattering using Prometheus NT.48 (NanoTemper). The samples were diluted to a concentration of 1 mg/ml in 15 ml using 25 mM NaP, pH 7.0; 125 mM NaCI and temperature was ramped from 20 to 90 °C at 2 °C/min.

Results

Table 13 below lists the values obtained for apparent size and the number of repeating units of the shielding polypeptide of the fusion protein. Figure 10 shows the normalised scattering signal as function of increased temperature (°C). In this figure, arylsulfatase A (PSI0590, SEQ ID NO: 49) is represented by solid black line, whereas arylsulfatase A fusion proteins (PSI0681 ,

PSI0683, PSI0685, PSI0687, PSI0689, PSI0691 ) are represented each by a dashed line.

Conclusions

For the fusion proteins the apparent size in solution was larger than expected for globular proteins of similar masses. A correlation between number of repeat units in the fusion proteins and their size in solution was observed. The increase in apparent size in solution was independent of the position of the shielding polypeptides (the N-terminus, C-terminus or both) - the apparent size in solution increased to the same extent. From the aggregation propensity analysis it can be concluded that unfused arylsulfatase A (i.e., without shielding polypeptide) aggregates at a lower temperature compared to the fusion proteins.

Table 13. Biophysical characterization of ASA and ASA fusion proteins

Example 16: Retained catalytic activity of fusion proteins of ASA with shielding polypeptides

ASA activity is dependent on post-translational conversion of a cysteine (Cys 69) in the active site of the protein into a formyl glycine. The specific activity of the protein can be measured by the method described in this example.

Material and methods

ASA enzymatic activity was assessed using para-nitrocatechol sulfate as substrate at 25 mM in 50 mM sodium acetate, pH 4.5 at 0°C. The reaction was typically stopped after 90 minutes using an excess of 0.5 M sodium carbonate at pH 10.7 and an end product nitrocatechol formed was measured utilizing absorbance at 515 nm, which is essentially based on Lee-Vaupel and Conzelmann 1987, Clin Chim Acta 164:171-180.

Results

The activity of ASA fusion proteins is in a range of 0.5-1.2 of ASA and is tabulated in Table 14.

Table 14. In vitro activity of ASA fusion proteins relative to ASA activity

Conclusion

The ASA fusion proteins display a similar in vitro activity as ASA. The in vitro activity is not compromised by the fusion of shielding polypeptides to ASA.

Example 17: Reduced in vitro glycan receptor mediated cellular uptake of fusion proteins of arylsulfatase A (ASA) with shielding polypeptides

In this example we show cellular uptake of ASA and ASA fusion proteins. Material and methods

Experiments were performed as in Example 5 apart from the concentration of added protein to the cells and the immunoassay used for determining amount of protein taken up by the cells. Concentrations of added protein was 3 - 50 nM. Cell lysates were analyzed for ASA and ASA fusion protein content by an immunoassay using the Meso Scale Discovery (MSD) platform. Streptavidin coated MSD plates were blocked with 5% Blocker A in PBS. The plates were washed in PBS-T and a biotinylated goat anti-ASA polyclonal antibodies was added to the wells and the plates were incubated at room temperature followed by wash in PBS-T. Different dilutions of standard and samples were distributed on the plates and the plates were incubated at room temperature followed by wash in PBS-T. Sulfo-Ru-tagged rabbit anti-ASA polyclonal antibodies was added and the plates were incubated at room temperature. Complexes of ASA and labelled antibodies bind to the streptavidin coated plate via the biotinylated pAb. After washing, the amount of bound complexes was determined by adding a read buffer to the wells and the plates were read in a MSD SI2400 instrument. The recorded electrochemiluminescence counts were proportional to the amount of ASA in the samples and evaluated against a relevant ASA standard.

Results

The ASA fusion proteins showed decreased cellular uptake compared to unfused ASA as shown in Table 15. The decrease in cellular uptake of the ASA fusion proteins was similar for all fusion proteins tested.

Conclusion

The cellular uptake was greatly reduced for fusion proteins compared to ASA alone. The fusion of shielding polypeptide(s) to ASA resulted in decreased cellular uptake of the proteins compared to unfused ASA. All fusion proteins showed similarly low cellular uptake, irrespective of the number of repeat units or their position (N- or C-terminal, or both).

Table 15. Cellular uptake of ASA fusion proteins compared to ASA in MEF-1 cells

Example 18: Expression and purification of fusion proteins of alpha-L-

Iduronidase (iduronidase) with shielding polypeptides

Materials and methods

DNA constructions: DNA constructions were done as in Example 2. The encoded proteins are presented in Table 16. Table 16. Iduronidase and iduronidase fusion protein sequences.

Cultivation and purification: ExpiCHO cells were transfected with expression vectors and cultured as described in Example 2. Purifications were performed as in Example 2. Purified fusion proteins were formulated in PBS, pH 7.4.

Results

The fusion proteins expressed well in ExpiCHO and purifications resulted in protein preparations of high purity as assessed by SDS-PAGE. Conclusions

Fusion proteins of iduronidase with shielding polypeptide moieties can be produced by construction of synthetic genes, followed by expression in mammalian cell systems and purification to high purity using conventional techniques.

Example 19: Altered biophysical properties of fusion proteins of iduronidase with shielding polypeptides compared to iduronidase

This example describes the biophysical characterization of fusion proteins of iduronidase with shielding polypeptide moieties using unfused iduronidase (i.e., without shielding polypeptide) as reference. Apparent size, molecular mass and Stokes radius in solution were determined by SEC/MALS.

Aggregation propensity was determined by backlight scattering. Materials and methods

Performed as in Example 3 with the exception that the unfused iduronidase was run at 0.58 mg/ml and a Superdex 200 Increase 10/300 GL column (GE Healthcare Life Sciences), calibrated as in example 3, was used. Aggregation propensity as function of increased temperature was determined by backlight scattering using Prometheus NT.48 (NanoTemper). The samples were diluted to a concentration of 1 mg/ml in 15 ul using 25 mM NaP, pH 7.0; 125 mM NaCI and temperature was ramped from 20 to 90 °C at 2 °C/min. Results

Table 17 below lists the values obtained for molecular weight, apparent size in solution and Stokes radius and the number of repeat units of the shielding polypeptides of the fusion protein. Figure 11 shows the normalised scattering signal as function of increased temperature (°C). In this figure, iduronidase (PSI0754, SEQ ID NO: 61 ) is represented by solid black line, whereas iduronidase fusion proteins (PSI0753, SEQ ID NO: 62; PSI0755, SEQ ID NO: 63; PSI0756, SEQ ID NO: 64) are represented each by a dashed line.

Conclusions

For the fusion proteins with shielding polypeptides the apparent size in solution was larger than expected for globular proteins of similar masses. A correlation between number of repeat units of the shielding polypeptide moieties of the and their size in solution was observed. The increase in apparent size in solution was independent of the position of the shielding polypeptide (N-terminus, C-terminus, or both) - the apparent size in solution increased to the same extent. From the aggregation propensity analysis it can be concluded that unfused iduronidase (i.e., without shielding polypeptide) aggregates at a lower temperature compared to the fusion proteins.

Table 17. Biophysical characterization of iduronidase and iduronidase fusion proteins

Example 20: Retained catalytic activity of fusion proteins of iduronidase with shielding polypeptides

The specific activity of iduronidase was measured by the method described in this example.

Material and methods

Iduronidase enzymatic activity was assessed using 4-methylumbelliferyl a-L- iduronide as substrate (Isemura et al 1978, J Biochem 84:627-632) at 0.1 mM in 0.2 M sodium formate, pH 3.5 at 37°C. The reaction was typically stopped after 20 minutes using an excess of 0.5 M sodium carbonate at pH 10.7 and an end product 4-MU formed was measured utilizing fluorescence intensity at X355/E460 nm. Results

The activity of iduronidase fusion proteins was in the range of 1.5-1.9 of iduronidase and are tabulated in Table 18.

Table 20. In vitro activity of iduronidase fusion proteins to relative to iduronidase

Conclusion

The iduronidase fusion proteins display a similar in vitro activity as

iduronidase. The in vitro activity was not compromised by the fusion of shielding polypeptides to iduronidase.

Example 21: Shielding polypeptides interfere with iduronidase binding to mannose-6-phosphate receptor

This example demonstrates the binding of iduronidase fusion proteins to M6PR in vitro. Binding properties are compared to properties of unfused iduronidase (i.e., without shielding polypeptide).

Materials and methods

Interactions between iduronidase fusion protein and iduronidase, respectively, with M6PR were analyzed by SPR technology essentially as in Example 6. Analyte concentrations were kept constant at 1000 nM with a association phase of 3 min and a dissociation phase of 3 min.

Results

M6PR was coupled to the chip with a total coupling response of 7967 RU. All iduronidase fusion proteins showed interaction with immobilized M6PR although with a significantly reduced response magnitude as compared to iduronidase. By the end of the 3 minutes association phase binding had reached equilibrium and responses at this time point are found in Table 21.

Table 21. Interactions between M6PR and iduronidase protein with shielding polypeptides using SPR technology

Conclusions

Fusion proteins of iduronidase with shielding polypeptides show significantly reduced interaction with the M6PR receptor as compared to iduronidase alone. Fusion of shielding polypeptides to both N- and C-terminus of iduronidase blocks the M6PR interaction more efficiently than fusion of shielding polypeptides to either N- or C-terminus.

Example 22: Reduced In vitro glycan receptor mediated cellular uptake of iduronidase fusion proteins with shielding polypeptides

This example demonstrates effect on cellular uptake by iduronidase fusion proteins compared to iduronidase.

Materials and methods

Mouse embryonic fibroblasts (MEF-1 ) were seeded the day before treatment with proteins. The proteins were diluted in cell media in a concentration range between 78 pM and 10 nM. The cells were incubated approximately 24 hours, washed in PBS and harvested in lysis buffer with protease and phosphatase inhibitors. Cell lysates were analyzed for iduronidase and iduronidase fusion protein content by an immunoassay using the Meso Scale Discovery (MSD) platform. Streptavidin coated MSD plates were blocked with 5% Blocker A in PBS. The plates were washed in PBS-T and different dilutions of standard and samples were distributed on the plate. A mixture of a biotinylated anti- iduronidase rabbit polyclonal antibody and Sulfo-Ru-tagged rabbit anti- idurondase antibodies were added and the plates were incubated at room temperature. Complexes of iduronidase and labelled antibodies bind to the streptavidin coated plate via the biotinylated polyclonal antibody. After washing in PBS-T, the amount of bound complexes was determined by adding a read buffer to the wells and the plates were read in a MSD SI2400 instrument. The recorded electrochemiluminescence counts were proportional to the amount of iduronidase in the samples and evaluated against a relevant iduronidase standard.

Results

The cellular uptake of the iduronidase fusion proteins is tabulated in Table 23.1 as % decreased uptake compared to iduronidase without shielding polypeptide.

Table 23.1. Decrease in cellular uptake of iduronidase fusion proteins compared to iduronidase in MEF-1 cells

Conclusions

The iduronidase fusion proteins SPM1 -17-lduronidase and SPM1 -17- lduronidase-SPM1 -17 showed lower cellular uptake compared to iduronidase alone. For fusion proteins with both N- and C-terminal shielding polypeptides the cellular uptake was further reduced compared to fusion proteins with N- terminal shielding polypeptides only. However, for fusion protein Iduronidase- SPM1 -17, this particular positioning of the shielding polypeptides allows for a cellular uptake of the fusion protein similar to the unfused protein. Example 23: Expression of fusion proteins of N-acetylgalactosamine-6- sulfatase (GalN6S) with shielding polypeptides

Materials and methods

DNA constructions: DNA constructions were done as in Example 2. The encoded proteins are presented in Table 19.

Table 19. GalN6S and GalN6S fusion protein sequences

Cultivation and purification: ExpiCHO cells were transfected with expression vectors and cultured as described in Example 2. Purifications were performed as in Example 2. Purified fusion protein was formulated in PBS pH 7.4

Results

The fusion proteins expressed in ExpiCHO and purification resulted in protein preparation with high purity as assessed by SDS-PAGE. Figure 7 shows the resulting cell media harvest of GalN6S and GalN6S fusion proteins, analyzed by SDS-PAGE. Lane A: Molecular weight marker SeeBlue plus 2 (Thermo Fisher Scientific), lane B: GalN6S (PSI0592), lane C: GalN6S-SPM1 -5 (PSI0777), lane D: GalN6S-SPM1 -17 (PSI0778), lane E: SPM1 -5-GalN6S-

SPM1 -5 (PSI0779), lane F: SPM1 -17-GalN6S-SPM1 -17 (PSI0780), lane G: SPM1 -5-GalN6S (PSI0784), lane H: SPM1 -17-GalN6S (PSI0786). Stars indicate produced protein of interest in harvest media. Conclusions

Fusion proteins of GalN6S with shielding polypeptide moieties can be expressed in mammalian cell systems. Example 24: Expression and purification of fusion protein of arylsulfatase B (ASB) with shielding polypeptide

Materials and methods

DNA constructions: DNA construction was carried out as set out in Example 2. The encoded protein sequence is presented in Table 20.

Table 20. ASB and ASB fusion protein Cultivation and purification: ExpiCHO cells were transfected with expression vector and cultured as described in Example 2. Purifications were performed as in Example 2. Purified fusion protein was formulated in PBS, pH 7.4.

Results

The fusion protein expressed in ExpiCHO and purifications resulted in protein preparations with high purity as assessed by SDS-PAGE. Figure 8 shows the resulting cell media harvest of ASB fusion protein, analyzed by SDS-PAGE and compared to ASB (Galsulfase; purchased). Lane A: Molecular weight marker SeeBlue plus 2 (Thermo Fisher Scientific), lane B: Galsulfase, lane C: SPM1 -17-ASB (PSI0746). Star indicates produced protein of interest in harvest media.

Conclusions

ASB fusion proteins with shielding polypeptide moieties can be expressed in mammalian cell systems.

Example 25: Expression of fusion protein of galactocerebrosidase (GALC) with shielding polypeptide

Materials and methods

DNA constructions: DNA construction was done as in Example 2. The encoded protein sequences are presented in Table 20. Table 21 . GALC fusion protein

Cultivation: ExpiCHO cells were transfected with expression vectors and cultured as described in Example 2.

Results

The fusion proteins were expressed in ExpiCHO. Figure 12 shows the resulting cell media harvest of GALC fusion protein.

LIST OF REFERENCES 1. Cox & Cachon-Gonzalez, J Pathol 226: 241 -254 (2012)

2. Neufeld; Birth Defects Orig Artie Ser 16: 77-84 (1980)

3. Kim et al, Curr Opin Struct Biol 19(5):534-42 (2009)

4. Lee et al, Glycobiology 13: 305-313 (2003)

5. Stromqvist et al. Arch. Biochem. Biophys. 1997

6. Loomes et al., Eur. J. Biochem. 1999, 266, 105-111 ).

7. Kissel et al., Biochimica et Biophysica Acta 1989, 1006).

8. Wang & Hartsuck (1993) Biochim. Biophys Acta 1166: 1 -19

9. Chen et al. Biochemistry 1998

10. Rosenberg, AAPS J, 8:E501 -7 (2006)

11. Wang, Int J Pharm, 185:129-88 (1999)

12.Voznyi et al 2001 , J Inherit Metab Dis 24:675-680

13. Lee-Vaupel and Conzelmann 1987, Clin Chim Acta 164:171-180 14. lsemura et al 1978, J Biochem 84:627-632