WO2025132764A1 - A sortilin-binding polypeptide - Google Patents

Abstract

The present disclosure generally relates to a sortilin-binding polypeptide, an affibody, and a fusion protein or conjugate comprising the sortilin-binding polypeptide. The present disclosure further relates to the sortilin-binding polypeptide, affibody, fusion protein or conjugate for use as a medicament, and a pharmaceutical composition comprising the same. The sortilin-binding polypeptide may be used in a method of treatment of a variety of disorders, preferably frontotemporal dementia (FTD).

Description

A SORTILIN-B INDING POLYPEPTIDE

TECHNICAL FIELD

The present disclosure generally relates to a sortilin-binding polypeptide, an affibody, and a fusion protein or conjugate comprising the sortilin-binding polypeptide. The present disclosure further relates to the sortilin-binding polypeptide, affibody, fusion protein or conjugate for use as a medicament, and a pharmaceutical composition comprising the same. The sortilin-binding polypeptide may be used in a method of treatment of a variety of disorders, such as frontotemporal dementia (FTD).

BACKGROUND

Frontotemporal dementia (FTD) is a form of non-Alzheimer’ s dementia that is particularly common among cases of early-onset dementia, with the prevalence among dementia patients younger than 65 years ranging from 3 to 27% in different studies [1], The disease is characterized by atrophy of the frontal and temporal lobes of the brain [2], leading to one of three main clinical phenotypes, with symptoms ranging from behavioral changes to language impairments [3],

The disease has a strong genetic component, in the sense that a family history of FTD is present in up to 40% of all cases [4-6], The majority of these cases are due to mutations in the genes encoding either microtubule-associated protein tau (MAPI), progranulin (GRN), or chromosome 9 open reading frame 72 (C9orf72) [7],

Heterozygous mutations in the GRN gene, encoding the protein progranulin (PGRN), are present in about 5-10% of all FTD cases, and up to 26% of familial FTD cases [4,8-11], To date, at least 130 different disease-associated mutations have been identified in the GRN gene [7],

These heterozygous mutations in the GRN gene lead to more than 50% decreased plasma and cerebrospinal fluid levels of progranulin (PGRN), a neurotrophic factor with lysosomal functions.

PGRN is a 593-amino acid glycoprotein consisting of seven and a half cysteine-rich granulin domains [12-14], into which the protein can be cleaved by both intra- and extracellular proteases [15,16], PGRN and the different granulins exert a multitude of, sometimes opposing [16], functions, including roles in lysosomal function [17,18] and as a neurotrophic factor [19-21], The identified pathogenic mutations in the GRN gene are believed to cause FTD through haploinsufficiency [9,22], as they are associated with more than 50% decreased PGRN levels in plasma and CSF of mutation carriers compared to controls [19,23-25],

Accordingly, increasing PGRN levels to the normal range can be envisioned as a potential therapeutic strategy for the treatment of FTD with GRN mutations (FTD-GRA).

An interesting target to this end is the PGRN clearance receptor, sortilin. Sortilin is a type I membrane protein, and the main luminal domain of sortilin is a 10-bladed beta propeller, into the tunnel of which both PGRN and the neuropeptide neurotensin (NT) bind [26-29],

PGRN interacts with sortilin through the PGRN C-terminal tail, leading to endocytosis and lysosomal localization of PGRN [28,29], Thus, sortilin is a negative regulator of extracellular PGRN levels [28,30], Importantly, the neurotrophic effects of PGRN have been shown to be independent of sortilin binding [20,21], making sortilin an attractive target for efforts to increase extracellular PGRN levels.

Blocking the PGRN-sortilin interaction has been demonstrated to increase PGRN levels in vitro and in vivo in several studies [31-33], Most notably, the anti-sortilin IgGl antibody latozinemab [32] is currently in phase 3 clinical trials for FTD-GAA (clinicaltrials.gov ID: NCT04374136), which indicates that the sortilin-targeting strategy might be a viable one.

Anti-sortilin treatment against FTD is envisioned to last for many years. However, antibodies are generally costly and complicated to manufacture. Their large size may also limit tissue penetration. There is consequently a need for a sortilin-targeting entity, which shows high specificity, affinity, and stability and which is easy and cost-efficient to produce.

SUMMARY

Encouraged by the above-mentioned results, and in view of the above- mentioned need, the present inventors have provided an affibody -based sortilin-targeting approach and identified a variety of sortilin-binding polypeptides displaying an enhanced target-affinity for sortilin.

The polypeptides of the present disclosure may inhibit sortilin-mediated progranulin degradation such that the extracellular PGRN levels can be increased.

The inventors have performed phage display selections against human and murine sortilin and identified the best leads through various affinity selection and surface plasmon resonance (SPR) screening experiments. Four affibodies, referred to as A3, G11, F6, and Cl, respectively (described in more detail hereinbelow), as well as variants of these affibodies, were shown to display an enhanced binding to human and/or murine sortilin.

Furthermore, mutational studies have been performed to evaluate the binding contribution of amino acids and how substitution with other amino acids affects the binding to sortilin.

Affibody molecules, i.e. affibodies, are small (58-amino acid, ~6.5 kDa) affinity proteins based on the three-helix scaffold of the Z domain derived from Staphylococcus aureus protein A [34,35], The Z domain is a mutated form of the B domain of Protein A. Typically, 13-14 surface-exposed positions in helix 1 and 2 are randomized to generate a library from which affibodies with affinity for new targets can be selected [36] (Fig. la).

Domain Z has a large tolerance for amino acid replacements despite its small size of only 58 amino acids. Domain Z provides a stable scaffold that can be adapted for various target interactions by replacement of its natural binding affinity for the Fc region of antibodies, by e.g. sortilin-binding. The inherent stability and refolding ability of domain Z also constitutes a potential for the development of polypeptides remaining intact or retaining their structural integrity after various biological and chemical conditions, e.g. high temperatures, and extreme pH.

In comparison to antibodies, the affibody scaffold lacks inherent effector functions, as well as disulfide bonds, and generally benefits from high stability. Its small size also makes it amenable to production in bacterial hosts and through chemical synthesis. The small non-binding surface area may also minimize non-specific binding, which may be a problem with larger proteins.

Affibodies have been investigated in various medical applications and have demonstrated good safety and tolerability in humans [45], Compared to antibodies, the small size of the affibody molecules entails a higher binding site density, which enables smaller injected volumes and faster administration, with similar brain uptake [46],

In the general case of dementia, and specifically the case of the genetic FTD- GRN, ease of administration and production costs are important factors since long treatments are likely required.

In the affibody molecule, positions 9-35 are of particular relevance for sortilin- binding. Position 35 represents the most C-terminal position of the positions varied in the selection from the original library of Z domain variants. The sortilin-binding polypeptides of the present disclosure are based on these positions of the affibody molecules.

According to a first aspect of the present disclosure, there is provided a sortilin- binding polypeptide comprising the amino acid sequence:

X9X10X11X12X13X14X15X16X17X18X19X20X21X22X23X24X25X26X27X28X29X30X31X32X33X34X35

(SEQ ID NO: 1), wherein

X₉ is I, N, R, H, T, K, or Q, preferably N, R, Q, K, or I;

X10 is E, A, V, K, T, H, Y, D, Q, R, S, or I, preferably E, H, Q, R, Y, A, or V;

Xu is E, G, W, H, M, R, K, Y, A, I, Q, T, V, F, L, N, or D, preferably A, E, I, Q, V, W, R, G, or T;

X12 is A, V or T, preferably A;

X13 is G, R, S, K, Y, W, F, or Q, preferably G, F, K, Q, R, or Y;

X_His A, W, N, H, K, R, Y, F, Q, or G, preferably A, F, G, or R;

X15 is E, D or V, preferably E;

Xi6 is I, N or T, preferably I;

X17 is I, T, R, M, V, W, Y, L, or K, preferably I, R, V, W, or T;

Xi8 is Q, F, H, V, or Y, preferably F, Y or Q;

X19 is L, P or Q, preferably L;

X20 is P, S, Q or T, preferably P;

X21 is N, K, S or Y, preferably N;

X22 is L, M or Q, preferably L;

X23 is N or T;

X₂₄ is R, K, S, Q, L, E, or N, preferably E, K, N, S, or R;

X₂₅ is W, N, K, R, H, F, L, or Q, preferably F, H, K, N, R, or W;;

X26 is Q, P or H, preferably Q;

X27 is K, G, H, Q, S, T, A, W, or absent, preferably K, H, Q, or G;

X₂8 is G, W, Q, H, Y, M, or absent,;

X29 is A or T, preferably A;

X30 is F, I or Y, preferably F;

X31 is I, K, H, Y, W, F, S, T, or D, preferably I, W, D, H, K, S, or W;

X32 is V, H, M, F, R, or K, preferably H, M, K, or V;

X33 is S, G, T or I, preferably S;

X34 is L, S, F or V, preferably L, and

X35 is K, W, A, E, F, H, T, M, D, Y, G, Q, or R, preferably D, K, W, R, Q, M, or F. The sortilin-binding polypeptide as defined hereinabove forms part of an affibody. This part represents the part of the affibody molecule associated with an enhanced sortilin-binding.

More specifically, in SEQ ID NO: 1, the amino acid residues in position X9, X10, Xu, X13, X14, X17, Xis, X24, X25, X27, X28, X31, X32, and X35 are believed to contribute to the enhanced binding to sortilin.

The amino acid residues in position X12, X15, Xi6, X19, X20, X21, X22, X23, X26, X29, X30, X33, X34 of SEQ ID NO: 1 may vary as defined hereinbefore without significantly affecting the binding of the polypeptide to sortilin.

It is also conceivable that the amino acid residues in position X12, X15, Xi6, X19, X20, X21, X22, X23, X26, X29, X30, X33, X34 of SEQ ID NO: 1 correspond to the amino acids of the wild type Z domain.

Hence, the sortilin-binding polypeptide may comprise the amino acid sequence:

X9X10X11AX13X14EIX17X18LPNLNX24X25QX27X28AFX31X32SLX35 (SEQ ID NO:2).

In exemplary embodiments, the sortilin-binding polypeptide may be derived from or based on the affibody referred to as A3.

In this case, the amino acids of SEQ ID NO: 1 or SEQ ID NO:2 may vary in the following manner:

X9 is I, N, K, Q, or R, preferably I, N, or K;

X10 is E, H, Q, or R;

Xu is A, Y, I, W, L, N, E, or Q, preferably A, E, I, Q, V, or W;

X13 is G, K, R, W, F, Y, or Q, preferably G, F, K, Q, R, or Y;

Xuis W, Y, A, F, G, H, R, or Q, preferably A, F, G, or R ;

X17 is I, L, W, K, V, or R, preferably I, R, V, or W;

Xi8 is H, Y, F, V, or Q, preferably F, Q, or Y;

X24 is K, R, L, E, or N, preferably E, K, or R;

X₂₅ is W, F, L, N, H, K, or Q, preferably F, H, K, N, or W;

X27 is A, K, Q, H, or absent, preferably H, or K;

X28 is G, H, Y, or absent;

X31 is I, or W, preferably W;

X32 is V; and

X35 is D, K, T, or W, preferably D, K, or W. Alternatively, the sortilin-binding polypeptide may be derived from or based on the affibodies referred to as G11, F6, or Cl.

In these cases, the amino acids of SEQ ID NO: 1 or SEQ ID NO:2 may vary in the following manner: X<; is N, R, or Q;

Xio is Y, A, V, or H;

Xu is A, R, G, W, T, or V;

X13 is K, R, or F ;

Xu is G;

X17 is I, T, or V;

Xis is F, or Y;

X₂₄ is N, K, or S;

X₂₅ is R, K, or N;

X27 is H, Q, or G;

X₂₈ is M, Y, W, or Q;

X31 is D, H, K, S, or W;

X32 is M, K, or H; and

X35 is R, Q, M, D, or F.

In exemplary embodiments, the sortilin-binding polypeptide comprises the amino acid sequence as defined in any one of SEQ ID NO:4-7, SEQ ID NO:29-41, or a corresponding sequence having at least 70%, preferably at least 80 %, more preferably at least 90% sequence homology with SEQ ID NO:4-7 or SEQ ID NO:29-41.

The sequences are listed in table 1 hereinbelow, and SEQ ID NO:3 represents the original Z domain (positions X9-X35). In the context of the present disclosure, the original Z domain may also be referred to as the “wildtype” Z domain.

Table 1 : Wild-type affibody polypeptide and sortilin-binding polypeptide sequences

The sortilin-binding polypeptide defined by SEQ ID NO:4 represents the amino acid sequence in positions X9-X35 of the affibody referred to as G11.

The sortilin-binding polypeptide defined by SEQ ID NO: 5 represents the amino acid sequence in positions X9-X35 of the affibody referred to as F6.

The sortilin-binding polypeptide defined by SEQ ID NO: 6 represents the amino acid sequence in positions X9-X35 of the affibody referred to as Cl.

The sortilin-binding polypeptide defined by SEQ ID NO: 7 represents the amino acid sequence in positions X9-X35 of the affibody referred to as A3.

The sortilin-binding polypeptides defined by SEQ ID NO:29-38 represent the amino acid sequences in positions X9-X35 of the A3 affibody variants used in the Example section.

The sortilin-binding polypeptides defined by SEQ ID NO: 39-41 represent the amino acid sequences in positions X9-X35 of the G11 affibody variants used in the Example section.

In SEQ ID NO:29, 31, and 33-36, the amino acid in position 28 (X28) is absent.

The affibodies G11, F6, Cl, and A3, and the variants thereof display an improved binding to murine and/or human sortilin (see SPR sensorgrams of figures 1C and ID, as well as figures 13-15).

In exemplary embodiments, the sortilin-binding polypeptide may further comprise the amino acid sequence XiX2X3X4X5X6X?X8X8b (SEQ ID NO:8) at the N terminus of any one of SEQ ID NO: 1, SEQ ID NO:2, SEQ ID NO:4-7, or SEQ ID NO:29-41 wherein Xi is V;

X₂ is D;

X3 is N or T, preferably N;

X4 is K, R or E, preferably K;

X5 is F, I, L or S, preferably F or S, more preferably F;

Xe is N or S, preferably N;

X7 is K or R, preferably K;

Xs is E, V or D, preferably E or V, more preferably E; and

Xsb is V, or absent, preferably absent.

In exemplary embodiments, the sortilin-binding polypeptide may further comprise the amino acid sequence X36X37X38X39X40 (SEQ ID NO: 9) at the C terminus of any one of SEQ ID NO: 1-2, SEQ ID NO:4-7, or SEQ ID NO:29-41 wherein

X36 is D, E or V, preferably D or V,

X37 is D or N, preferably D,

X38 is P or S, preferably P,

X39 is S, T, I or G, preferably S or T, more preferably S, and

X40 is Q, L, E or H, preferably Q, E or H, more preferably Q or H.

In exemplary embodiments, the sortilin-binding polypeptide may further comprise the amino acid sequence X41X42X43X44X45X46X47X48X49X50X51X52X53X54 (SEQ ID NO: 10) at the C terminus of SEQ ID NO: 9, wherein

X41 is S, N, R or G, preferably S or N,

X42 is A, G, V or T, preferably A,

X43 is N or D, preferably N,

X44 is L, M or F, preferably L or M,

X45 is L, P or Q, preferably L or P,

X46 is A, P, G, T or S, preferably A, P or T,

X47 is E, D or K, preferably E or K,

X48 is A, V or T, preferably A,

X49 is K, R, N or I, preferably K or R,

X50 is K, N, E or M, preferably K or M,

X51 is L, I, Q or P, preferably L,

X52 is N, I, D or K, preferably N,

X53 is D, N or E, preferably D or N, and

X54 is A, T, P or V, preferably A. In exemplary embodiments, the sortilin-binding polypeptide may further comprise the amino acid sequence QAX57K (SEQ ID NO: 11) at the C terminus of SEQ ID NO: 10, wherein X57 is P or S, preferably P.

According to a second aspect, there is provided an affibody comprising the sortilin-binding polypeptide as described hereinabove.

As used herein, the term “affibody” or “affibody molecule” means a protein based on the three-helix scaffold of the Z domain of Staphylococcal protein A (SpA). The affibody molecule consists of 58 amino acids and has a molar mass of about 6 kDa. As mentioned hereinbefore, affibody molecules may be engineered to bind a large number of target proteins or peptides with high affinity. Specific affibody molecules which bind a desired target can be extracted from pools containing billions of different variants using phage display or E. coli display. The wildtype affibody sequence is defined by SEQ ID NO: 12.

Accordingly, the affibody may have a three-helix-bundle structure.

As used herein, the term “three-helix bundle structure” refers to a polypeptide scaffold comprising three alpha helices that are arranged in a compact, folded configuration. The helices are typically packed closely together in parallel or antiparallel orientations, stabilized by hydrophobic core interactions and other non-covalent bonds. The three-helix bundle anti-parallel conformation provides a framework that supports the display of the sortilin-binding amino acids in a structurally favorable orientation. This conformation helps maintaining the structural integrity of the polypeptide and allows for high-affinity interactions with sortilin.

In the context of the present disclosure, the binding surface comprising the side chains of the 13-14 randomized amino acids may be transplanted into variants of the Z scaffold, having modifications in non-binding parts of the protein domain. Many such variants of the Z domain having retained binding functionality to the selected target may be envisioned, having replacements before (N-terminal of) the binding residues in helix 1, after these and up to the binding residues in helix 2, and after those residues in helix 2. Exchanges may be done in many positions in the supporting helix 3.

Furthermore, the loops between helices 1 and 2, and between helices 2 and 3 may be replaced, not only in certain positions, but in their entirety by longer or shorter loop segments, while retaining the three-helix bundle core supporting the amino acid residues composing the binding surface. Accordingly, the numbering of e.g. the binding residues of helix 2 may be shifted if the first loop is replaced. Therefore, the residues corresponding to the first loop (i.e, approximately no. 19-22) in the affibodies exemplifying the invention thus only illustrates one exemplary embodiment of the invention. The same applies to any numbering beyond residue no. 37, which initiates the second loop region.

The affibody of the present disclosure may be defined by the sequence: X1X2X3X4X5X6X7X8X9X10X11X12X13X14X15X16X17X18X19X20X21X22X23X24X25X26X27X28X29

X30X31X32X33X34X35X36X37X38X39X40 X41X42X43X44X45X46X47X48X49X50X51X52X53X54 QAX57K (SEQ ID NO:28), wherein the various X positions contain the amino acid residues as defined hereinabove.

Preferably, the affibody of the present disclosure is defined by any one of SEQ ID NO: 13-16, SEQ ID NO:42-54, or a corresponding sequence having at least 70%, preferably at least 80 %, more preferably at least 90% sequence homology with SEQ ID NO: 13-16 or SEQ ID NO:42-54.

In table 2, hereinbelow, SEQ ID NO: 12 represents the wild type affibody sequence (Z wt).

Table 2: Affibody wild-type and sortilin-binding affibodies SEQ ID NO: 13 represents the amino acid sequence of the affibody referred to as Gi l.

SEQ ID NO: 14 represents the amino acid sequence of the affibody referred to as F6.

SEQ ID NO: 15 represents the amino acid sequence of the affibody referred to as Cl.

SEQ ID NO: 16 represents the amino acid sequence of the affibody referred to as A3.

SEQ ID NO:42-51 represent the amino acid sequences of the A3 affibody variants used in the Example section.

SEQ ID NO: 52-54 represent the amino acid sequences of the G11 affibody variants used in the Example section.

In SEQ ID NO:42, 44, and 46-49, the amino acid in position 28 (X28) is absent.

As mentioned hereinbefore, each one of these affibodies displays an improved binding to human and/or murine sortilin.

The inventors have also evaluated a variety of fusion constructs based on the affibodies described hereinbefore and identified promising candidates displaying a significantly improved target affinity for sortilin.

Hence, according to a third aspect of the present disclosure, there is provided a fusion protein or conjugate comprising:

- a first part comprising the sortilin-binding polypeptide as described hereinbefore or the affibody as described hereinbefore; and

- a second part comprising at least one peptide adapted to enhance the binding to sortilin.

In exemplary embodiments, the at least one peptide of the second part comprises at least four amino acids, and wherein the C-terminal end amino acid of the second part is leucine (L) or isoleucine (I).

The present inventors have performed several fusion trials with peptides derived from the progranulin C-terminus, as well as the neuropeptide neurotensin (NT). Both of these peptides are known to bind to sortilin.

As mentioned hereinbefore, progranulin (PGRN) interacts with sortilin through the PGRN C-terminal tail, leading to lysosomal degradation of PGRN. The inventors thus tested a variety of PGRN peptides to evaluate their compatibility with the sortilin- binding affibodies, and whether an enhanced sortilin-binding could be observed. As described in more detail in the Example sections, the affibodies of the present disclosure were successfully fused with various lengths of PGRN peptides derived from the C-terminus of progranulin, as well as peptides derived from neurotensin.

In these tests, the C-terminal end amino acid leucine (L), or isoleucine (I) were identified as important to accomplish an improved binding to sortilin. Furthermore, the inventors found that the peptide of the second part should comprise at least four amino acids to secure an improved binding to sortilin, as well as for providing a stable construct.

In embodiments where the peptide of the second part is derived from progranulin or neurotensin, the second part is arranged C-terminally of the first part of the fusion protein or conjugate.

Hence, the second part forms the C-terminal end of the fusion protein or conjugate.

The peptide of the second part may comprise the sequence: X79X80X81X82 (SEQ ID NO: 17), wherein

X79 is selected from any amino acid residue;

Xso is G, N, P, Q, or Y, preferably Q, or Y;

Xsi is L, V, or I, preferably V, or I; and

X82 is L or I, preferably L.

For example, the peptide of the second part may comprise the sequence: QLL (SEQ ID NO: 18) or YIL (SEQ ID NO: 19).

SEQ ID NO: 18 represents the sequence of the C-terminal end amino acids of human progranulin.

SEQ ID NO: 19 represents the sequence of the C-terminal end amino acids of neurotensin.

Alternatively, the peptide of the second part may comprise the sequence QLI (SEQ ID NO: 55).

In exemplary embodiments, the peptide of the second part may comprise from 4 to 24, preferably from 6 to 18 amino acids, and may be defined by the sequence: X₅9X60X61X62X63X64X65X66X67X68X69X70X71X72X73X74X75X76X77X78X₇9X80X81X82 (SEQ ID NO:20), wherein

X59 to X78 are independently optional or are selected from any amino acid residue;

X79 is selected from any amino acid residue;

Xso is G, N, P, Q, or Y, preferably Q, or Y;

Xsi is L, V, or I, preferably V, or I; and X82 is L or I, preferably L.

The fusion protein or conjugate may comprise a linker between the first and second part. It is also conceivable that a linking or spacer function is provided by a portion of the second part of the fusion protein or conjugate.

As demonstrated in the tests performed by the inventors (see Example section hereinbelow, and e.g. figure 3), the very C-terminal-most amino acids of PGRN are sufficient to convey sortilin-binding in fusion with the anti-sortilin affibody A3, but not alone (see figure 18-19).

Even though a longer peptide yields a higher affinity to sortilin (as is observed with the fusion construct A3-PGRNcl5*), the results indicate that the N-terminal amino acids of the PGRN derived peptides mainly serve as spacers, with the PGRNc3 peptide harbouring the key residues for binding. This is evidenced in Figure 3, where PGRNc3 with a flexible linker conveys a similar binding contribution to PGRNc6* (A3-(G4S)3-PGRNc3 vs. A3-(G₄S)3-PGRN_C6*).

This is furthermore in agreement with the A588G mutation not affecting the PGRN affinity for sortilin [33], and indicates that the N-terminal amino acids of the optimal fusion peptide PGRNcl5* might be amenable to further mutations, for instance to increase stability of the fusion construct in vivo (see Example section).

In figure 3, and in table 3 hereinbelow, the fusion proteins, and PGRN peptides, respectively, denoted with * represent PGRN peptides comprising an A588G mutation. With reference to SEQ ID NO:20 hereinbefore, the A588G mutation corresponds to the substitution of A in position X77 with G.

Preferably, the peptide of the second part comprises from 4 to 18 amino acids. If the peptide of the second part is too long, the stability of the fusion protein or conjugate may be impaired.

In exemplary embodiments the peptide of the second part is a peptide (PGRNc) derived from the C-terminus of progranulin. The peptide may be defined by any one of SEQ ID NO: 21-26.

Table 3: Amino acid sequences of the C-terminal end of progranulin

Preferably, the first part of the fusion protein or conjugate as described hereinabove comprises the sortilin-binding polypeptide as defined in claim 3.

Hence, the first part of the fusion protein or conjugate comprises a sortilin- binding polypeptide derived from or based on the affibody molecule A3.

The inventors have found that when the fusion protein or conjugate comprises the affibody referred to as A3 (or a sortilin-binding peptide derived from the affibody A3), significant improvements in target-affinity for sortilin are observed.

Fusion of the affibody A3 with the PGRN C-terminus led to avidity effects and a dramatically decreased dissociation rate, which indicates that the A3 and PGRNc epitopes are compatible for simultaneous binding to sortilin (see figure 2C, and Example section).

Furthermore, optimization of the peptide moiety of the A3 -PGRNc fusion construct enabled an up to 380-fold reduction in apparent KD of the fusion construct compared to the parental affibody alone (see Examples).

The A3 -affibody -based fusion constructs were also shown to increase extracellular PGRN levels in vitro for a sortilin-expressing, PGRN-secreting cell line, with an ECso comparable to that of a latozinemab biosimilar (see figures 5 and 10).

Preferably, the first part of the fusion protein or conjugate comprises a sortilin- binding polypeptide comprising the sequence as defined in SEQ ID NO:7, SEQ ID NO:29- 38, or a corresponding sequence having at least 70%, preferably at least 80 %, more preferably at least 90% sequence homology with SEQ ID NO:7 or SEQ ID NO:29-38.

The sortilin-binding polypeptide may be extended by SEQ ID NO: 8, SEQ ID NO:9, SEQ ID NO: 10 and/or SEQ ID NO: 11 as described hereinbefore.

Alternatively, the first part of the fusion protein or conjugate may be an affibody as defined in SEQ ID NO: 16, SEQ ID NO:42-51, or a corresponding sequence having at least 70%, preferably at least 80 %, more preferably at least 90% sequence homology with SEQ ID NO: 16 or SEQ ID NO:42-51.

In embodiments where the first part of the fusion protein or conjugate comprises a sortilin-binding polypeptide derived from or based on the affibody A3, the at least one peptide of the second part of the fusion protein or conjugate may be a sortilin- binding polypeptide as defined in claim 4.

Accordingly, the fusion protein or conjugate may comprise two sortilin-binding polypeptides, wherein the first sortilin-binding polypeptide is derived from or based on the affibody referred to as A3 (or any of the variants thereof), and wherein the second sortilin- binding polypeptide is derived from or based on the affibody referred to as G11, F6 or Cl.

The first part comprising the first sortilin-binding polypeptide (derived from A3) may be arranged N-terminally or C-terminally of the second part comprising the second sortilin-binding polypeptide (derived from G11, F6 or Cl).

The peptide of the second part may comprise the sequence as defined in any one of SEQ ID NO:4-6, SEQ ID NO:39-41, preferably SEQ ID NO:4 or SEQ ID NO:39-41, or a corresponding sequence having at least 70%, preferably at least 80 %, more preferably at least 90% sequence homology with SEQ ID NO:4-6 or SEQ ID NO:39-41.

The inventors have found that when the affibody A3 (or a variant thereof) is fused with the affibody G11 or truncated variants of the affibody G11, significant improvements in sortilin-binding affinity are observed (see figures 16 and 17).

The second part may be the affibody defined in any one of SEQ ID 13-15, or SEQ ID NO: 39-41, preferably SEQ ID NO: 13, SEQ ID NO: 39-41, or a corresponding sequence having at least 70%, preferably at least 80 %, more preferably at least 90% sequence homology with said sequence.

The fusion protein or conjugate may be a dimer or a multimer of the sortilin- binding polypeptide described hereinbefore.

It is also conceivable that the fusion protein or conjugate comprises at least one sortilin-binding polypeptide, or affibody as described hereinbefore, and a fragment of an antibody, e.g. a Fc fragment.

The antibody fragment may be fused to the sortilin-binding polypeptide or affibody at the N-terminal end thereof (leaving the C-terminal end of the peptide/affibody free to bind).

The fusion protein or conjugate may further comprise an albumin-binding domain (ABD).

The albumin-binding domain may extend the half-life of the fusion protein or conjugate. ABD binds to serum albumin, a naturally abundant protein with a long half-life. By binding to albumin, the fusion protein “piggybacks” on the long half-life of albumin, and thereby obtains an extended circulation time. Alternative peptides may be added to the sortilin-binding fusion proteins, described herein, to increase the half-life, including PAS-ylation and XTENylation [43],

The albumin-binding domain may e.g. comprise the sequence defined in SEQ ID NO:27.

However, the present disclosure is by no means limited to this particular sequence.

According to a fourth aspect, there is provided a sortilin-binding polypeptide according to the first aspect, an affibody according to the second aspect or a fusion protein or conjugate according to the third aspect for use as a medicament.

According to a fifth aspect, there is provided a pharmaceutical composition comprising the sortilin-binding polypeptide according to the first aspect, an affibody according to the second aspect or a fusion protein or conjugate according to the third aspect.

The pharmaceutical composition is preferably adapted for intravenous or subcutaneous injection.

According to a sixth aspect, there is provided a sortilin-binding polypeptide, an affibody, a fusion protein, conjugate or a pharmaceutical composition for use in a method of treatment of a disorder.

According to a seventh aspect, there is provided a method of treatment or prevention of a disorder in a subject, wherein the method comprises administration of a polypeptide, affibody, fusion protein, conjugate or pharmaceutical composition of any one of the preceding aspects.

The disorder may be any disorder involving a decreased extracellular progranulin (PGRN) level, and wherein the addition of a sortilin-binding polypeptide can inhibit the sortilin-mediated PGRN clearance such that the extracellular PGRN levels can be increased.

For example, the disorder of the sixth or seventh aspect may be dementia, frontotemporal lobar degeneration (FTLD), amyotrophic lateral sclerosis (ALS), Parkinson’s disease (PD), Alzheimer’s disease (AD), limbic-predominant age-related transactive response DNA-binding protein 43 encephalopathy (LATE), neuropathic pain, or cancer.

Preferably, the dementia is frontotemporal dementia (FTD).

The method of the sixth or seventh aspect typically comprises injection, such as intravenous or subcutaneous injection, of the sortilin-binding polypeptide, affibody, fusion protein, conjugate or pharmaceutical composition.

The subject of the method of the sixth or seventh aspect is preferably human. According to an eighth aspect, there is provided the use of the polypeptide, affibody, fusion protein, or conjugate as described hereinbefore for the in vitro diagnosis of the presence and/or level of sortilin in a sample. This may be of interest for example in oncology where high levels of a soluble form of sortilin correlates with poor prognosis [44], Further features of, and advantages with, the present disclosure will become apparent when studying the appended claims and the following description. The skilled addressee realizes that different features of the present disclosure may be combined to create embodiments other than those described in the following, without departing from the scope of the present disclosure.

BRIEF DESCRIPTION OF THE DRAWINGS

The various aspects of the present disclosure, including its particular features and advantages, will be readily understood from the following detailed description and the accompanying drawings, in which:

Figure l is a schematic overview of the affibody scaffold and initial characterization of anti-sortilin affibodies. (A) Schematic of a three-helical bundle affibody molecule, highlighting the 14 randomized positions in the phage display library. (B) Circular dichroism data for the anti-sortilin affibody A3, showing alpha helical secondary structure with complete refolding after thermal denaturation (left). Variable temperature measurements (VTM) at 221 nm (right) indicate a melting temperature of 46°C. (C)-(D) SPR sensorgrams showing the interaction between the anti-sortilin affibodies G11, F6, Cl and A3 and human sortilin (C) or murine sortilin (D). Affibody -ABD fusion proteins were captured on a sensor chip surface functionalized with human serum albumin (HSA), followed by injection of variable sortilin concentrations, as indicated (right). The displayed data shows sensorgrams from which a reference capture of affibody -ABD with a 0 nM sortilin injection has been subtracted.

Figure 2 illustrates SPR screening of first-generation affibody-PGRNc fusion proteins. (A) Schematic overview of the evaluated constructs. In addition to Z-ABD- PGRN_C21* and ABD-Z-PGRN_C21* format constructs, Z-ABD and ABD-PGRN_C21* constructs were included as controls. (B)-(C) SPR sensorgrams showing the interaction between representative constructs and human sortilin. Fusion of anti-sortilin affibody F6 (B) with the PGRN C-terminus does not lead to avidity effects, whereas fusion of the anti-sortilin affibody A3 (C) with the PGRN C-terminus leads to avidity effects and a marked decrease in sortilin dissociation rate compared to the affibody or the PGRNc21* peptide alone. Figure 3 illustrates SPR-based affinity determination of second-generation A3- PGRNc fusion proteins. Representative sensorgrams showing the interaction of the second- generation A3 -PGRNc fusion constructs, as well as A3-ABD and ABD-PGRNc21* controls with varying concentrations of human sortilin. ABD-containing constructs were captured on an HSA-coated sensor chip, followed by sortilin injection, and reference subtraction.

Figure 4 illustrates the characterization of sortilin binding on cancer cells by flow cytometry. The binding of anti-sortilin affibody A3, the PGRN C-terminus, and A3- PGRNc constructs to sortilin-expressing U-251 cells (A) and very low-expressing PC-3 cells (B) was evaluated in comparison to the non-binding affibody control Zwt. 100 nM of ABD fusion proteins were pre-incubated with HSA-Alexa Fluor 647 prior to addition to the cells, enabling fluorescent detection of binding. ABD-A3-PGRNcl5* is shown as a representative example of the high-affinity A3 -PGRNc fusion proteins.

Figure 5 illustrates the characterization of the biological activity of sortilin- binding constructs in a PGRN clearance assay. Addition of a suitable sortilin binder inhibits the sortilin-mediated PGRN clearance, increasing extracellular PGRN levels. Extracellular PGRN levels were measured by ELISA after 72 h incubation of sortilin-expressing, PGRN- secreting U-251 cells with increasing concentrations of A3-PGRN fusion constructs or controls. PGRN levels are reported as mean+SD fold change versus untreated cells from n=3 biological replicates. The data shown is a representative example of N=3 independent experiments.

Figure 6 illustrates secondary structure characterization of anti-sortilin affibodies G11, F6, and Cl. (A)-(C) Circular dichroism spectroscopy data for anti-sortilin affibodies Gi l (A), F6 (B), and Cl (C) shows mainly disordered structure. CD spectra are shown before (pre-VTM) and after (post-VTM) heating to 95°C.

Figure 7 illustrates SPR sensorgrams of first-generation anti-sortilin affibody- PGRNc fusions for affibodies G11 and Cl (A) SPR sensorgram showing the interaction of the PGRNc21* peptide with varying concentrations of human sortilin. (B)-(C). SPR sensorgrams showing the interaction of affibodies G11 (B) and Cl (C), and their fusions with PGRNc21* in the formats ABD-Z-PGRNc21* and Z-ABD-PGRNc21* with human sortilin. ABD-containing constructs were captured on an HSA-coated sensor chip, followed by sortilin injection, and reference subtraction.

Figure 8 illustrates flow cytometric data for binding of control constructs to cancer cells. (A)-(B) The negative control affibody Zwt shows no binding to U-251 (A) or PC-3 (B) cells compared to the secondary reagent HSA-Alexa Fluor 647 alone. (C)-(D) Binding of a positive control anti-sortilin antibody confirms the presence of sortilin on U-251 cells (C), but shows very low signals on PC-3 cells (D) compared to secondary antibody only. Primary constructs were pre-incubated with their respective secondary constructs prior to addition to the cells.

Figure 9 illustrates the flow cytometric analysis of constructs binding to cancer cells. Overview of median fluorescence intensity signals normalized to Zwt signal for the respective cell line for all tested ABD-containing constructs or secondary reagent only. 100 nM of primary construct was pre-incubated with 200 nM of secondary HSA-Alexa Fluor 647 prior to addition to the cells. The displayed data shows the mean ± SD of n=2 independent experiments.

Figure 10 illustrates a summary of ECso values in the PGRN clearance assay. U-251 cells were incubated with varying concentrations of A3-PGRNc fusion proteins, affibody controls, or a latozinemab biosimilar in triplicates for 72 h. Supernatant PGRN levels were measured by ELISA and converted to PGRN fold change versus untreated cells (see Fig. 5 for an example). ECso values were calculated as the inflection point of a 4- parameter curve fit to the PGRN fold change vs construct concentration curve. ECso values for constructs ABD-A3-PGRNcl5* and latozinemab are given as mean±SD for three independent experiments, values for ABD-A3-PGRNc6*, ABD-A3-PGRNc9*, and ABD- A3-(G4S)3-PGRNC3 are given as mean±SD for two independent experiments, and remaining constructs were screened in one experiment only. Constructs A3-ABD, ABD-PGRNc21*, ABD-A3-PGRN_C3, ABD-PGRN_C21*, G11-ABD, and Zwt-ABD were also included in the screening, but failed to reach a plateau within the tested concentration range, making ECso determination impossible.

Figure 11 illustrates human and murine sortilin binding of unique affibody clones from FACS3, selected from the A. coli display libraries based on the anti-sortilin affibody A3. E. coli cells displaying the indicated affibody in fusion with ABD were analyzed by flow cytometry. Target-binding signal, as quantified by SAPE fluorescence intensity, was normalized for surface expression level, as quantified by HSA-Alexa Fluor 647 fluorescence intensity, yielding the normalized MFI. This was further normalized to the signal from Zwt (negative control), yielding relative sortilin binding values for all clones.

Figure 12 illustrates human and murine sortilin binding of unique affibody clones from FACS3T, selected from the E. coli display libraries based on the anti-sortilin affibody A3. E. coli cells displaying the indicated affibody in fusion with ABD were analyzed by flow cytometry. Target-binding signal, as quantified by SAPE fluorescence intensity, was normalized for surface expression level, as quantified by HSA-Alexa Fluor 647 fluorescence intensity, yielding the normalized MFI. Relative normalized MFI was obtained by defining the highest value to be equal to 1. Clones M533, HM49, HM50, and M51 were included for comparison.

Figure 13 illustrates SPR-based human sortilin affinity determination of affibody variants selected from the E. coli display libraries based on the anti-sortilin affibody A3. All affibody constructs were captured on an HSA surface via a C-terminal ABD035, followed by injection of varying concentrations of human sortilin, as indicated. Following reference subtraction, a 1 : 1 Langmuir model (dotted lines) was used to estimate kinetic constants.

Figure 14 illustrates SPR-based murine sortilin affinity determination of affibody variants selected from the E. coli display libraries based on the anti-sortilin affibody A3. All affibody constructs were captured on an HSA surface via a C-terminal ABD035, followed by injection of varying concentrations of murine sortilin, as indicated. Following reference subtraction, a 1 : 1 Langmuir model (dotted lines) was used to estimate kinetic constants.

Figure 15 illustrates SPR-based sortilin affinity determination of affibody variants selected from the E. coli display library based on the anti-sortilin affibodies G11, F6 and Cl. All affibody constructs were captured on an HSA surface via a C-terminal ABD035, followed by injection of varying concentrations of either human or murine sortilin, as indicated. Following reference subtraction, a 1 : 1 Langmuir model (dotted lines) was used to estimate kinetic constants.

Figure 16 illustrates affinity determination of anti-sortilin affibody fusion A3- G11 by surface plasmon resonance. (A-B) The affinity of anti-sortilin affibodies A3 and G11 as well as the fusion protein A3-G11 for human sortilin was determined by SPR, using HSA capture to immobilize the affibody constructs on the sensor chip surface via their ABD, followed by injection of varying concentrations of human sortilin. Sensorgrams were obtained at (A) 25°C and (B) 37°C.

Figure 17 illustrates SPR-based affinity screening of first-generation anti- sortilin affibody variants A3 and G11, as well as C-terminally truncated variants of the fusion protein A3-G11 for human sortilin. HSA capture was used to immobilize the ABD- containing affibody constructs on the sensor chip surface, followed by injection of 50 nM of human sortilin at 37°C. Figure 18 illustrates affinity determination of peptides derived from the PGRN C -terminus for human sortilin by surface plasmon resonance. Constructs with N-terminal His6-Zwt-G4S-ABD035 were captured on an HSA chip, followed by injection of varying concentrations of human sortilin, as indicated. The constructs vary in their C-terminal composition, as indicated above each sensorgram.

Figure 19 illustrates affinity determination of peptides derived from the PGRN C -terminus for murine sortilin by surface plasmon resonance. Constructs with N-terminal His6-Zwt-G4S-ABD035 were captured on an HSA chip, followed by injection of varying concentrations of murine sortilin, as indicated. The constructs vary in their C-terminal composition, as indicated above each sensorgram.

DETAILED DESCRIPTION

The present invention provides novel sortilin-binding polypeptides which bind sortilin with high affinity.

Accordingly, the sortilin-binding polypeptides can be used to increase the extracellular levels of progranulin (PGRN), which makes these polypeptides promising candidates in any disorder involving decreased levels of PGRN.

According to a first aspect of the present disclosure, there is provided a sortilin- binding polypeptide comprising the amino acid sequence: X9X10X11X12X13X14X15X16X17X18X19X20X21X22X23X24X25X26X27X28X29X30X31X32X33X34X35 (SEQ ID NO: 1), wherein

X₉ is I, N, R, H, T, K, or Q, preferably N, R, Q, K, or I;

X10 is E, A, V, K, T, H, Y, D, Q, R, S, or I, preferably E, H, Q, R, Y, A, or V;

Xu is E, G, W, H, M, R, K, Y, A, I, Q, T, V, F, L, N, or D, preferably A, E, I, Q, V, W, R, G, or T;

X12 is A, V or T, preferably A;

X13 is G, R, S, K, Y, W, F, or Q, preferably G, F, K, Q, R, or Y;

Xi₄ is A, W, N, H, K, R, Y, F, Q, or G, preferably A, F, G, or R;

X15 is E, D or V, preferably E;

Xi6 is I, N or T, preferably I;

X17 is I, T, R, M, V, W, Y, L, or K, preferably I, R, V, W, or T;

Xi8 is Q, F, H, V, or Y, preferably F, Y or Q;

X19 is L, P or Q, preferably L;

X20 is P, S, Q or T, preferably P; X21 is N, K, S or Y, preferably N;

X22 is L, M or Q, preferably L;

X23 is N or T;

X₂₄ is R, K, S, Q, L, E, or N, preferably E, K, N, S, or R;

X₂₅ is W, N, K, R, H, F, L, or Q, preferably F, H, K, N, R, or W;;

X26 is Q, P or H, preferably Q;

X27 is K, G, H, Q, S, T, A, W, or absent, preferably K, H, Q, or G;

X₂₈ is G, W, Q, H, Y, M, or absent,;

X29 is A or T, preferably A;

X30 is F, I or Y, preferably F;

X31 is I, K, H, Y, W, F, S, T, or D, preferably I, W, D, H, K, S, or W;

X32 is V, H, M, F, R, or K, preferably H, M, K, or V;

X33 is S, G, T or I, preferably S;

X34 is L, S, F or V, preferably L, and

X35 is K, W, A, E, F, H, T, M, D, Y, G, Q, or R, preferably D, K, W, R, Q, M, or F.

According to a second aspect of the present disclosure, there is provided an affibody comprising the sortilin-binding polypeptide according to the first aspect.

According to a third aspect of the present disclosure, there is provided a fusion protein or conjugate comprising:

- a first part comprising the sortilin-binding polypeptide as described hereinbefore or the affibody as described hereinbefore; and

- a second part comprising a peptide adapted to enhance the binding to sortilin.

According to a fourth aspect, there is provided a sortilin-binding polypeptide according to the first aspect, an affibody according to the second aspect or a fusion protein or conjugate according to the third aspect for use as a medicament.

According to a fifth aspect, there is provided a pharmaceutical composition comprising the sortilin-binding polypeptide according to the first aspect, an affibody according to the second aspect or a fusion protein or conjugate according to the third aspect.

According to a sixth aspect, there is provided a sortilin-binding polypeptide, an affibody, a fusion protein, conjugate or a pharmaceutical composition for use in a method of treatment of a disorder.

According to a seventh aspect, there is provided a method of treatment or prevention of a disorder in a subject, wherein the method comprises administration of a polypeptide, affibody, fusion protein, conjugate or pharmaceutical composition of any one of the preceding aspects.

The disorder of the sixth or seventh aspect may be dementia, frontotemporal lobar degeneration (FTLD), amyotrophic lateral sclerosis (ALS), Parkinson’s disease (PD), Alzheimer’s disease (AD), limbic-predominant age-related transactive response DNA- binding protein 43 encephalopathy (LATE), neuropathic pain, or cancer.

EXAMPLES

Materials and Methods

Phage display selection of anti-sortilin affibody molecules

Phage display selections of affibodies binding to sortilin were performed essentially as described by Giang et al [37] using a previously described library [38], Briefly, murine (R&D Systems, 2934-ST) and human (R&D Systems, 3154-ST) sortilin proteins were biotinylated for 1 h using EZ-Link Sulfo-NHS-LC-Biotin (Thermo Fisher Scientific) in a 25-fold molar excess followed by dialysis to PBS using a Slide-A-lyzer 10K MWCO (Thermo). Four cycles of selection were performed with decreased target concentration and increased washing in each round. In the first two cycles, biotinylated sortilin was preimmobilized onto paramagnetic streptavidin beads (M-280, Thermo) and in the later cycles phages were incubated with biotinylated target in solution before capture of target-phage complexes on beads. Parallel selections were performed using either murine or human sortilin as target, and bound phages were eluted by either trypsin cleavage (2.5 mg/ml #15090-046, Gibco by Life Technologies) or acid (0.3 M HAc, pH 2.8). Starting in the third round, two additional selection tracks were initiated from pooled phage stocks from the two elution strategies for each target (murine/human sortilin) and subjected to competitive elution (1 h incubation) using a solution of 5 pM neurotensin (Sigma, N6383) in PBST.

Screening of affibody candidates

47 random clones from each of the six selection tracks were assayed for binding to human and murine sortilin (immobilized in separate wells at 2.5 pg/ml in 0.1 M carbonate buffer pH 9.6) by phage-ELISA. Binding signal from an albumin-binding domain (ABDwt) tag fused to the C-terminus of phage p3-displayed Affibody to immobilized HSA was used to normalize binding signals in ELISA. Clones with positive ELISA data and expected insert size in PCR screening were sent for sequencing. Production and purification of recombinant proteins

Construct genes containing an ABD035 [39] were ordered in a pET21 vector (Twist Bioscience, South San Fransisco, CA, USA) with an N-terminal hexahistidine tag for immobilized metal affinity chromatography (IMAC) purification. Constructs of the format Hise-Z for circular dichroism were cloned into a pT7 vector. The plasmids were transformed into Escherichia coli BL21 Star cells (Thermo Fisher Scientific) using a standard heat shock transformation protocol. For each construct, a single colony was inoculated to 10 mL tryptic soy broth (TSB, Merck) supplemented with 100 pg/ml of carbenicillin (pET21) or 50 pg/ml of kanamycin (pT7), and incubated O/N at 37°C and 150 rpm. After 16 h, the O/N cultures were diluted 1 : 100 into TSB supplemented with yeast extract (TSB+Y, Merck) media supplemented with 100 pg/ml of carbenicillin (pET21) or 50 pg/ml of kanamycin (pT7). When ODeoo reached 0.7-1, the cultures were induced with isopropyl P-D-l- thiogalactopyranoside (IPTG) to a final concentration of 1 mM, and incubated O/N at 25°C and 150 rpm. After 16 h, cells were harvested by centrifugation (5000xg, 10 min, 4°C) and stored at -20°C. Subsequently, the cells were resuspended in IMAC equilibration buffer (20 mM Tris-HCl, 300 mM NaCl, 10 mM imidazole, pH 8.0) and lysed by sonication for 1.5 minutes (1 sec on: l sec off). Cell debris was removed by centrifugation (25 OOOxg, 15 min, 4°C), and the protein-containing supernatant was filtered (0.45 pm), followed by batch IMAC purification on HisPur Cobalt resin (Thermo Fisher Scientific) at 4°C. Briefly, the IMAC resin was washed with 2x10 CV equilibration buffer, followed by sample binding for 30 min, after which the matrix was washed with 3x10 CV wash buffer (20 mM Tris-HCl, 300 mM NaCl, 30 mM imidazole, pH 8.0). Bound protein was eluted by incubation with 2.5 mL elution buffer (20 mM Tris-HCl, 300 mM NaCl, 500 mM imidazole, pH 8.0) for 10 min. Eluted proteins were buffer-exchanged to PBS (pH 7.4) using PD-10 columns (Cytiva), according to the manufacturer’s recommendations, and analyzed by SDS-PAGE (NuPAGE, Invitrogen), bicinchoninic acid assay (Pierce, Thermo Fisher Scientific) and mass spectrometry (MS, 4800 MALDI TOF/TOF, Applied Biosystems/MDS SCIEX).

Surface Plasmon Resonance for screening and affinity determination

Target binding was assessed by surface plasmon resonance (SPR) using Biacore 3000 (screening of affibody clones after phage display selections), Biacore T200 (screening of first-generation affibody-PGRN fusion constructs and affibody heterodimers), and Biacore 8k (all other SPR experiments) instruments (Cytiva, Uppsala, Sweden). In all cases, HSA was immobilized through amine coupling on a CM-5 sensor chip according to the manufacturer’s recommendations, using 10 mM sodium acetate pH 4.5 as the immobilization buffer, with a reference surface being only activated and inactivated. PBST (PBS supplemented with 0.05% Tween20, pH 7.4) was used as the running buffer in all binding experiments. 30-100 RU of ABD035 -containing constructs were captured on the HSA surface, followed by injection of human (R&D Systems 3154-ST) or murine (R&D Systems 2934-ST) sortilin at 30 pl/min and 25°C, unless otherwise specified. The surfaces were regenerated through a 30 s injection of 10 mM HC1 at 30 pl/min. Kinetic constants were estimated using 1 : 1 Langmuir curve fits of reference subtracted sensorgrams.

Circular dichroism spectroscopy for secondary structure and melting temperature determination

Circular dichroism (CD) spectroscopy was performed on affibody molecules in a Hise-Z format in PBS (pH 7.4), using a Chirascan Circular Dichroism Spectrometer (Applied Biophysics Ltd, Leatherhead, UK). All analyses were performed at a concentration of 0.1-0.2 mg/ml and a 1 mm path length. Data on the secondary structure content was obtained by measuring ellipticity at 20°C from 195 nm to 260 nm. Thermal stability was assessed by measuring the change in ellipticity at 221 nm when heating from 20°C to 95°C at 5°C/min, after which another spectrum was obtained at 20°C (195-260 nm) to assess the molecules’ refolding capacity. The melting temperature (T_m) was obtained as the inflection point of a 4-parameter fit of the variable temperature measurement data in Prism version 10 (GraphPad, Boston, MA, USA).

Mammalian cell cultivation

U-251 (JCRB IFO50288) human glioblastoma cells were cultivated in Minimum Essential Medium with glutamine (Gibco MEM, Thermo Fisher Scientific 31095) supplemented with 10% fetal bovine serum. PC-3 (ATCC CRL-1435) prostatic adenocarcinoma cells were cultivated in Roswell Park Memorial Institute 1640 medium with glutamine (Gibco RPMI 1640, Thermo Fisher Scientific 21875) supplemented with 10% fetal bovine serum. Cells were grown at 37°C in a 5% CO2 atmosphere.

Fluorophore labeling of secondary reagents

Human serum albumin (HSA, 20 mg/ml) was labeled with Alexa Fluor™ 647 succinimidyl ester (Invitrogen A-20006, 0.5 mg/ml) in carbonate buffer (0.1 M, pH 8.5) for 3 hours at room temperature. The labeling reaction was quenched through the addition of an excess of glycine, followed by removal of free dye and glycine by gel filtration using PD-10 columns (Cytiva), according to the manufacturer’ s instructions, with phosphate-buffered saline (PBS, pH 7.4) as the target buffer.

Flow cytometric analysis of sortilin binding on cancer cells

Primary constructs were pre-incubated with secondary constructs in PBS+1% BSA on ice for a minimum of 1 hour prior to their addition to the cells. ABD035 -containing constructs (100 nM) were pre-incubated with HS A- Alexa Fluor 647 (200 nM), and a positive control Human Sortilin Antibody (R&D systems MAB31541, 0.625 pg/ml) was preincubated with Alexa Fluor 647 Goat Anti-Mouse Antibody (Invitrogen A21235, 2.86 pg/ml). Simultaneously, cells were detached from the culture flasks using TrypLE™ EXPRESS (Thermo Fisher Scientific 12605), according to the manufacturer’s instructions. 2 x 10⁵ cells per sample were washed once in 200 pl ice-cold PBS+1% BSA, followed by incubation with 200 pl primary+secondary reagents for 40 min at 4°C. Cells were then washed twice, followed by resuspension in 200 pl ice-cold PBS+1% BSA for analysis on a Cytoflex S flow cytometer (Beckman Coulter, Brea, CA, USA), using a 638 nm laser for excitation and a 660/10 BP filter for detection.

PGRN clearance assay

A PGRN clearance assay was performed essentially as described by Miyakawa et al [31], Briefly, 1 X 10⁴ U-251 cells in 100 pl of the appropriate medium were seeded in each well of a 96-well plate (Nunclon™ Delta Surface, Thermo Fisher Scientific). The cells were incubated for 24 h prior to the addition of fresh medium containing different concentrations of protein constructs, in triplicates. After 72 h, supernatants were collected, and PGRN concentrations were quantified using Human Progranulin DuoSet ELISA (R&D Systems DY2420), according to the manufacturer’s instructions. Absorbance was measured at 450 and 540 nm using a ClarioStar (BMG Labtech, Ortenberg, Germany) plate reader, and PGRN levels were normalized against untreated cells to obtain the PGRN level fold change upon treatment. ECso values were obtained from a 4-parameter fit in Prism version 10 (GraphPad, Boston, MA, USA). In addition to in-house-produced proteins, a latozinemab biosimilar (ProteoGenix PX-TA1676) was evaluated.

Mutational study of variable positions in the anti-sortilin affibodies A3, G1 L Cl and F6 In order to investigate the allowed amino acid variability in fourteen surface- exposed residues on helix 1 and 2 of the anti-sortilin affibodies A3, G11, Cl and F6, three affinity maturation libraries were constructed.

Affibody library generation

Three E. /z'-di splayed affibody libraries were prepared based on the pBAD vector [40], containing an affibody insertion site in fusion with an ABD035 and the AIDA-I autotransporter, under the control of an arabinose-inducible promoter. Three libraries were constructed from two synthetic oligonucleotides each, 75 (5’-CTC GAG GTG GAT AAC AAA TTC AAC AAA GAA X01 X02 X03 GCG X04 X05 GAA ATT X06 X07 CTG CCG AAC CTG AAC-3’, forward) and 84 (5’- AGC TAG CAG GTT CGC ACT CTG GCT CGG ATC ATC Z01 CAG GCT Z02 Z03 AAA CGC Z04 Z05 CTG Z06 Z07 GTT CAG GTT CGG CAG-3’, reverse complementary) nucleotides in length, spanning amino acids 1-46 of the affibody molecule, with a central complementary region (Ella Biotech, Furstenfeldbruck, Germany).

14 positions (X01-X07 in the forward oligonucleotide, Z01-Z07 in the reverse complementary oligonucleotide) had been randomized during the synthesis using mixtures of trinucleotides, according to table 4, with libraries 1 and 2 being based on anti-sortilin affibody A3, and library 3 being based on anti-sortilin affibodies G11, F6 and Cl. The two oligonucleotides of each library were combined to create double-stranded DNA libraries by PCR, using Phusion High-Fidelity DNA polymerase (New England Biolabs, Ipswich, MA, USA) for 20 cycles. This was followed by a second PCR reaction with 10 cycles, using primers Forward (5’-CGTACGTGCTCGAGGTGGATAACAAATTC-3’), defined by SEQ ID NO:56, and Reverse (5’-TAATTAGCTAGCAGGTTCGCACTCTGGCTC-3’), defined by SEQ ID NO:57, to introduce Xhol and Nhel restriction sites up- and downstream of the library sequence, respectively. Following gel extraction (Qiagen, Hilden, Germany), double digestion of the libraries and the vector was performed with XhoI-HF (New England Biolabs, Ipswich, MA, USA) and Nhel-HF (New England Biolabs, Ipswich, MA, USA) according to the manufacturer’s recommendations.

Subsequently, libraries were ligated into the pBad2.2 vector using T4 DNA ligase (New England Biolabs, Ipswich, MA, USA), at a molar ratio of vectorinsert of 1 :3 (library 1) or 1 :5 (library 2 and library 3). The resulting libraries were desalted using a PCR purification kit (Qiagen, Hilden, Germany), followed by electroporation into E. cloni EXPRESS BL21(DE3) electrocompetent cells (LGC Biosearch Technologies, Teddington, UK) according to the manufacturer’s recommendations. Library sizes were estimated by spreading dilutions of the transformation mix on carbenicillin agar plates after 1 h incubation in Expression Recovery Medium (LGC Biosearch Technologies, Teddington, UK). The remaining transformation mix was inoculated to 200 mL of Tryptic Soy Broth supplemented with 100 g/ml of carbenicillin. Following incubation at 37°C and 150 rpm overnight, the libraries were harvested by centrifugation at 4600xg, 10 min, 4°C. The cell pellets were resuspended in sterile 85% glycerol to a final concentration of 17% glycerol, followed by storage at -80°C.

Table 4 hereinbelow illustrate the amino acid composition of the anti-sortilin affibody A3, G11, Cl and F6 affinity maturation libraries.

Table 4: Amino acid composition of the anti-sortilin affibody A3 affinity maturation libraries (library 1 and 2), and the G11, F6 and Cl affinity maturation library (library 3). Selection procedure

Two rounds of magnetic-activated cell sorting (MACS) and three rounds of fluorescence-activated cell sorting (FACS) were performed in order to deplete the libraries of non-binding clones and enrich for higher-affinity variants. Selections from the A3 libraries (library 1 and 2) were performed in three separate main selection tracks, consisting of human only, murine only, or alternating human-murine-human-murine-human target. Furthermore, the murine-only and alternating human-murine tracks were each split into one track with and one track without trypsin treatment, starting at FACS2 and continuing into FACS3.

Selections from the G11/C1/F6 library (library 3) were performed in two tracks with alternating human-murine-human-murine-human, and human-murine-human-human-human target, respectively. For each round of selections, a number of cells from the appropriate library or selection output corresponding to at least 10 times the library size was used to inoculate Luria Bertani broth (LB, Sigma-Aldrich) supplemented with 100 pg/ml of carbenicillin to a starting ODeoo of no more than 0.1. Following incubation O/N at 37°C and 150 rpm, at least a lOx library coverage of cells was taken out from the cultures, and used to inoculate fresh LB supplemented with 100 g/ml of carbenicillin to a starting ODeoo of no more than 0.1. When ODeoo reached 0.5-0.8, the cultures were induced with 20% L- arabinose, to a final concentration of 0.6%, and incubated O/N at 25°C and 150 rpm. For MACS round 1 for the A3 selections, cells corresponding to lOx the library size for each of library 1 and library 2 were taken out and pooled, and used as selection input, whereas pure library 3 was used as selection input for the G11/C1/F6 selections. For subsequent rounds of selections, a number of cells corresponding to at least lOx coverage of the appropriate preceding selection outputs were used as input.

For MACS selections, Dynabeads MyOne Streptavidin Cl beads (Thermo Fisher Scientific) and cells were washed twice in PBS supplemented with 0.1% Pluronic F108 NF surfactant (BASF, PBSP). To remove any streptavidin-binding clones, cells were first incubated with non-coated beads for 30 min at room temperature (RT) under constant end-over-end (eoe) mixing. Non-bound cells were collected by magnetic separation, and added to beads coated with either site-specifically biotinylated human sortilin (Aero Biosystems, SON-H82E9) or in-house biotinylated (Biotin-XX Microscale Protein Labeling Kit, Invitrogen) murine sortilin (R&D Systems, 2934-ST), at a ratio of 1 : 10 beads:cells. For MACS1, cells were incubated with beads for 1.5 h (eoe) at RT, followed by 3 h (eoe) at 37°C, whereas cells were incubated with beads for 2 h (eoe) at 37°C for MACS2. Non-bound cells were subsequently removed by magnetic separation, and beads were washed thrice with PBSP. Finally, beads with remaining cells were resuspended in TSB supplemented with 100 pg/ml of carbenicillin, followed by incubation at 37°C and 150 rpm O/N. The number of cells in washes and output were estimated by plating dilutions on carbenicillin agar plates immediately after selections.

For FACS selections, cells were washed twice in PBSP, followed by resuspension in PBSP with 10 nM biotinylated human or murine sortilin, and incubation for 2 h at 37°C (eoe). In FACS round 1 (FACS1), cells were then washed twice with ice-cold PBSP, and incubated with 81 nM HSA-Alexa Fluor 647 and 2 pg/ml Streptavidin-R- phycoerythrin conjugate (SAPE, Invitrogen) for 20 min at 4°C. Cells were then washed twice and resuspended in PBSP, after which target-binding cells were sorted into TSB using a CytoFlex SRT cell sorter (Beckman Coulter, Brea, CA, USA), detecting affibody expression by L638-F660-10, and target binding by L561-F585-42.

In FACS round 2 without trypsination (FACS2), cells from the FACS1 output were treated as in FACS1 until incubation with biotinylated target, after which cells were washed twice and incubated with 50 nM non-biotinylated human (R&D Systems, 3154-ST ) or murine (R&D Systems, 2934-ST) sortilin for 30 min at 37°C (eoe). This was repeated once for the A3 selections and twice for the G11/C1/F6 selections, after which cells were again washed twice, and further treated in the same way as in FACS1. In FACS round 3 without trypsination (FACS3), FACS2 output cells were treated as in FACS2, except that cells were incubated 4x with non-biotinylated sortilin.

In FACS round 2 with trypsination (FACS2T), FACS1 output cells were washed twice in PBSP, followed by resuspension in PBSP with 1 pM trypsin-EDTA (Gibco, 25200072) and incubation for 10 min at 37°C prior to washing twice in PBSP. Cells were then resuspended in PBSP with 10 nM biotinylated human or murine sortilin, and treated as in FACS1. In FACS round 3 with trypsination (FACS3T), FACS2T output cells were treated as in FACS2T, except that 0.5 pM trypsin-EDTA was used with a 5 min incubation at 37°C. Furthermore, after incubation with the biotinylated target, cells were washed twice in PBSP and incubated with 50 nM non-biotinylated human or murine sortilin for 30 min at 37°C (eoe), followed by washing and incubation with HSA-Alexa Fluor 647 and SAPE as previously described.

Flow cytometric evaluation of selection outputs

The outputs from all selection rounds were evaluated for affibody expression and target binding by flow cytometry, alongside the unsorted libraries, parental affibody clones A3 and G11, and negative control affibody Zwt expressed on the surface of E. coli. In essence, samples were prepared according to the FACS1 protocol described above, but with analysis being performed on a CytoFlex S flow cytometer (Beckman Coulter, Brea, CA, USA), using L638-F660-10 for detection of surface expression level, and L561-F585-42 for detection of target binding.

Following confirmation of binding, 96 single colonies from each of the FACS3 A3 tracks, 48 single colonies from each of the G11 FACS3 tracks, and 96 single colonies from each of the FACS3T tracks were sent for Sanger sequencing (Microsynth Seqlab GmbH, Balgach, Switzerland).

Analysis of individual clones from FACS3 and FACS3T outputs

The sortilin-binding properties of unique clones identified by Sanger sequencing were screened in a single-clone format by E. coli display as described for selection outputs, but with cultures being inoculated from single colonies. Based on sortilin binding and sequence prevalence, a total of 15 clones from the A3 FACS3 output and 3 clones from the G11 FACS3 output were chosen for characterization by SPR. Variants were subcloned into a pT7 vector for expression in a Hise-Z-ABD035 format, followed by protein production, IMAC purification, and sortilin affinity evaluation by SPR, as previously described.

Results

Isolation of sortilin-binding affibody molecules through phage display

In order to obtain sortilin-binding affibody molecules, phage display selections were performed against human and murine sortilin. A previously described phage library [38] was subjected to four rounds of panning against decreasing concentrations of either human or murine sortilin, using acid, trypsin cleavage, or competition with the sortilin ligand neurotensin for elution. Target-binding clones were identified by phage ELISA screening of 47 randomly selected clones from each of the six selection tracks. DNA sequencing data from 54 target-binding clones with inserts of the correct size showed 31 unique clones. Among these, 12 clones representing major clusters were chosen for subcloning to a Hise-Z- ABDwt format, expressed in the E. coli cytoplasm, and purified by IMAC. After screening by surface plasmon resonance (SPR, data not shown), clones G11 (mSort track, trypsin elution), F6 (mSort track, acid elution), Cl (mSort track, acid elution), and A3 (hSort track, acid elution) were chosen for further characterization.

Production and characterization of sortilin-binding affibody molecules

The anti-sortilin affibody clones G11, F6, Cl and A3 were produced in soluble format in E. coli BL21 Star cells, with an N-terminal hexahistidine tag, with (Hise-Z-G4S- ABD035) or without (Hise-Z) a C-terminal albumin-binding domain (ABD035) [39], Following IMAC purification, the affibody molecules’ secondary structure and thermal stability were investigated using circular dichroism spectroscopy of Hise-Z format proteins. Interestingly, only variant A3 displayed the expected alpha helical structure. Variable temperature measurement showed its melting temperature (T_m) to be 46°C, with complete refolding after heating to 95°C (Fig. IB). Variants Gi l, F6, and Cl showed a CD spectrum largely consistent with a random coil conformation (Fig. 6).

The affibody molecules’ binding to human and murine sortilin was assessed by SPR, by capturing Z-ABD constructs on a sensor chip immobilized with HSA, followed by injection of five concentrations (ranging from 1 nM to 200 nM) of human or murine sortilin. No sortilin binding was observed to the negative control Zwt-ABD construct, confirming the absence of any interaction between sortilin and the ABD or the affibody scaffold. Interestingly, the three affibody candidates selected against murine sortilin were cross-reactive for human sortilin, whereas clone A3 only displayed binding to human sortilin (Fig. 1C-D). While the interaction patterns differed between the affibodies, the equilibrium dissociation constants were in the relatively narrow range of 34 to 155 nM and 49 to 78 nM for human and murine sortilin, respectively.

Design of first-generation affibody-PGRN fusion constructs

The sortilin-progranulin interaction is known to be mediated by the C-terminal tail of PGRN [28,29], Specifically, the three C-terminal -most amino acids (QLL) have been shown to be essential for binding to sortilin [29], Thus, we hypothesized that, given a second interaction partner with a suitable epitope on sortilin, a short peptide derived from the PGRN C -terminus could be used as a binding moiety to obtain a biparatopic sortilin-binding fusion protein. To test this hypothesis, we created fusion proteins of the four sortilin-binding affibodies and the human PGRN C-terminus, in two different configurations. Previous work by Zheng et al has demonstrated that the 24 C-terminal -most amino acids of PGRN (here denoted PGRNc24) are fully sufficient for binding to sortilin, whereas PGRNc9 and PGRNc6 interact with sortilin, but to a lesser extent [29], Since amino acid 22 from the C-terminus is a cysteine, we opted for PGRNc21 (L573-L593) as the initial PGRN moiety in the affibody- PGRN fusion proteins, in order to avoid dimerization due to disulfide bond formation. In addition, it has been shown that introduction of the A588G mutation in PGRN increases proteolytic stability by disruption of a neutrophil elastase cleavage site, with no effect on PGRN uptake by cells [33], The A588G mutation was confirmed to not affect sortilin binding by SPR (figures 18-19), and the PGRNc21 peptide carrying the A588G mutation (henceforth PGRNc21*) was chosen as the initial PGRN peptide moiety for fusion. Given that the epitopes of the affibodies were not known, the ABD moiety was evaluated in two positions in the fusion proteins, either N-terminally or between the affibody and PGRN moiety, to serve as a spacer. As a free PGRN C-terminus has been shown to be required for sortilin binding [21,29], this established the set of constraints for the construct design (Fig. 2A).

Evaluation of first-generation affibody-PGRN fusion constructs

Constructs of the format Hise-Z-G4S-ABD035, His6-Z-G4S-ABD035- PGRNc21* and Hise-ABD035-G4S-Z-PGRNc21* for each of the four sortilin-binding affibodies (Z), as well as Hise-ABD035-PGRNc21* were expressed in the A. coli cytoplasm and purified by IMAC. The potential for simultaneous sortilin binding by the affibody and PGRNc21* moi eties was investigated by SPR, as described above. Fusion of affibodies G11, Cl and F6 to PGRNc21* did not lead to any dramatic improvements in apparent affinity compared to the parental peptide or affibodies (Fig. 2B and Fig. 7), indicating that the epitopes were incompatible for simultaneous binding. In contrast, fusion of the PGRNc21* peptide to anti-sortilin affibody A3 led to a dramatic decrease in the sortilin dissociation rate (Fig. 2C), indicating the presence of avidity effects and compatibility between the A3 and PGRN epitopes on sortilin. Notably, the dissociation rate of the A3-PGRNc21* constructs was too slow to enable accurate determination of kinetic constants within the tested dissociation time. Qualitatively, the position of the ABD in the A3-PGRNc21* constructs seemed to have no major effect on binding, indicating that the ABD moiety is not required as a spacer between A3 and PGRNc21*.

Optimization of an affibody-PGRN fusion construct

In order to investigate whether all 21 amino acids of the PGRNc21* peptide are required for simultaneous sortilin binding in the A3-PGRNc fusion construct, constructs of the format His₆-ABD035-G₄S-A3-PGRN_cX*, where X={3, 6, 9, 12, 15, 18, 21 } were produced and purified. In addition, the contribution of the peptide’s N-terminal amino acids to sortilin binding was investigated by replacing these amino acids with a flexible linker, in constructs of the format His6-ABD035-G4S-A3-(G4S)3-PGRNcX*, where X={3, 6}. An initial comparison between the 3 C-terminal-most amino acids of the sortilin-binding polypeptide neurotensin (YIL), known to bind to the same epitope as PGRN [27,28,41], and those of PGRN (QLL) was also included, in the form of the construct His6-ABD035-G4S-A3- (G₄S)3-NT_C3.

The constructs’ apparent affinities for human sortilin were evaluated by SPR, as described above. Notably, a dramatic decrease in dissociation rate and apparent KD is seen between the A3-PGRNc3 (KD=3.93 nM) and the A3-PGRNc6* (KD=289 pM) constructs (Fig. 3, Table 5), indicating that PGRNc6* is a sufficient peptide length to both bind to sortilin and serve as a spacer between the A3 and PGRNc epitopes. In agreement with previous work [29], PGRNc6 seems to be sufficient to mediate a certain extent of binding to sortilin, but a somewhat longer peptide mediates a stronger interaction, as showcased by the minimum apparent KD observed for A3-PGRNcl5* (185 pM). However, the very similar apparent KD values of the A3-(G4S)₃-PGRNc3 (310 pM) and A3-(G4S)3-PGRNc6* (319 pM) constructs indicate that the 3 N-terminal-most amino acids of the PGRNc6* peptide mainly serve as a spacer in the A3-PGRNc6* construct, rather than being essential for sortilin binding. In addition, the A3-(G4S)₃-NTc3 (KD=303 pM) construct displays similar binding to A3-(G₄S)₃-PGRN_C3. (Fig. 3, Table 5).

Table 5: Summary of kinetic constants for A3-PGRNc fusion proteins and controls. Apparent association (k_on), dissociation (k_og), and equilibrium dissociation (KD) constants, as determined by SPR, are given as mean±SD of n=3 technical replicates.

Binding to sortilin on cancer cells

The constructs’ binding to sortilin in the biologically relevant context of the cell surface was investigated by comparing binding to the sortilin high-expressing glioblastoma cell line U-251 and the very low-expressing prostate cancer cell line PC-3. Sortilin expression was verified using a positive control antibody (Fig. 8). ABD-containing constructs were pre-incubated with an excess of fluorophore-labelled human serum albumin (HSA- Alexa Fluor 647), and were then added to cells and analyzed by flow cytometry. All constructs were shown to bind U-251 cells (Fig. 4A and Fig. 9), except the negative control affibody Zwt, which did not show any binding to U-251 or PC-3 cells (Fig. 8). Notably, the highest-affinity constructs showed a small fluorescence shift relative negative control affibody for the negative control cell line PC-3 (Fig. 4B). This is in line with reports of some sortilin expression on PC-3 cells [42], and the ultra-high affinity of these constructs for sortilin.

Effect on extracellular PGRN levels in vitro

The affibody-PGRNc peptide fusions’ biological activity was evaluated in a progranulin clearance assay. U-251 cells secrete progranulin as well as expressing sortilin, making them an interesting model system for studying the effect of sortilin binders on extracellular PGRN levels. Based on it having the highest apparent affinity for sortilin among the affibody -peptide fusions in SPR, A3-PGRNcl5* was chosen as the lead candidate, and evaluated in comparison to A3 alone, the PGRN C-terminus alone, an affibody negative control (Zwt), and the anti-sortilin antibody latozinemab. A3-PGRNcl5* increased extracellular PGRN levels up to approximately 2.5-fold compared to untreated cells, with an ECso value of 1.3+0.30 nM. Notably, this is comparable to that of latozinemab (0.68+0.20 nM), indicating that A3-PGRNcl5* holds promise as a therapeutic candidate (Fig. 5 and Fig. 10).

Mutational study of variable positions in the anti-sortilin affibody A3

The allowed amino acids in each target-binding position of the anti-sortilin affibody A3 were investigated by selecting the sortilin-binding population from a soft randomization E. coli display library based on the anti-sortilin affibody A3. The library used for selections was a 1 : 1 mixture of library 1 (1.38 X 10⁹ variants, Table 4) and library 2 (1.04 X 10⁹ variants, Table 4), with a total library size of 2.42 X 10⁹ variants. Selections were performed in three main tracks, using either only human, only murine, or alternating human-murine-human-murine sortilin as target. After two rounds of MACS and one round of FACS, the murine-only and alternating tracks were each further split into one track with and one track without trypsination. 96 single colonies were sent for sequencing from each FACS3 track with and without trypsination (FACS3T and FACS3 tracks, respectively). From these, 27 unique sequences were identified from the FACS3 top tracks, and 33 additional unique sequences were identified from the FACS3T tracks. Their binding to at least one sortilin ortholog was confirmed by single-clone E. coli display using human and murine sortilin (Figures 11-12).

In table 6 hereinbelow, “Allowed amino acids” lists every amino acid observed in a particular position in at least 5% of the sequences in at least one of the FACS3 or FACS3T tracks.

Table 6: Allowed amino acids in variable positions and framework positions with observed deviations from the designed amino acid, for sortilin-binding affibodies derived from the anti- sortilin affibody A3 libraries. Position XSB indicates an insertion between position Xs and X9.

A total of 15 clones from the FACS3 tracks with improved binding properties were subcloned to an expression vector and evaluated by SPR. 10 of the tested variants displayed improved binding to both human and murine sortilin, as compared to the parental clone A3 (Figure 13-14). Kinetic constants for each variant’s interaction with human and murine sortilin, respectively, were obtained from a 1 : 1 Langmuir fit, as summarized in Table 7.

Table 7. Kinetic constants for the interaction between human (hSort) and murine (mSort) sortilin and improved affibodies selected from the A3 library, as compared to the parental affibody A3. NA indicates too low signals to enable quantification. A3HM35 is defined by SEQ ID NO:42. The sortilin-binding polypeptide defined by SEQ ID NO:29 represents the amino acid sequence in positions X9-X35 of A3HM35.

A3HM49 1S defined by SEQ ID NO:43. The sortilin-binding polypeptide defined by SEQ ID NO:30 represents the amino acid sequence in positions X9-X35 of A3HM49.

A3HM5O 1S defined by SEQ ID NO:44. The sortilin-binding polypeptide defined by SEQ ID NO:31 represents the amino acid sequence in positions X9-X35 of A3HMSO.

A3M51 is defined by SEQ ID NO:45. The sortilin-binding polypeptide defined by SEQ ID NO:32 represents the amino acid sequence in positions X9-X35 of A3MSL

A3M528 1S defined by SEQ ID NO:46. The sortilin-binding polypeptide defined by SEQ ID NO:33 represents the amino acid sequence in positions X9-X35 of A3M528.

A3M557 is defined by SEQ ID NO:47. The sortilin-binding polypeptide defined by SEQ ID NO:34 represents the amino acid sequence in positions X9-X35 of A3M557.

A3M559 is defined by SEQ ID NO:48. The sortilin-binding polypeptide defined by SEQ ID NO:35 represents the amino acid sequence in positions X9-X35 of A3M559.

A3M565 is defined by SEQ ID NO:49. The sortilin-binding polypeptide defined by SEQ ID NO:36 represents the amino acid sequence in positions X9-X35 of A3M565.

A3M569 is defined by SEQ ID NO:50. The sortilin-binding polypeptide defined by SEQ ID NO:37 represents the amino acid sequence in positions X9-X35 of A3M569.

A3M589 is defined by SEQ ID NO:51. The sortilin-binding polypeptide defined by SEQ ID NO:38 represents the amino acid sequence in positions X9-X35 of A3M589.

In SEQ ID NO:29, 31, 33-36 (as well as in SEQ ID NO:42, 44,46-49), the amino acid in position X28 is absent.

In table 8 hereinbelow, “Allowed amino acids” lists every amino acid observed in a particular position in either the parental affibody A3, or any of the 10 A3 -derived clones with improved human and murine sortilin binding as confirmed by SPR.

Table 8: Allowed amino acids in variable positions of sortilin-binding affibodies derived from the anti-sortilin affibody A3 libraries, with sortilin binding confirmed by SPR.

Possible sequence variability for anti-sortilin affibodies G1 L Cl and F6

The allowed amino acids in each target-binding position of the anti-sortilin affibodies G11, Cl and F6 were investigated by selecting the sortilin-binding population from an E. /z'-di splay library based on the anti-sortilin affibodies Gi l, Cl and F6 (1.03xl0⁹ variants, library 3 in Table 4). Selections were performed using alternating human-murine- human-murine-human and human-murine-human-human-human sortilin as target. After two rounds of MACS and three rounds of FACS, 48 clones from each of the two selection tracks were sent for sequencing, leading to the identification of a total of 11 unique sequences from 94 high-quality sequences. The sortilin binding of the 11 unique clones was evaluated in E. coli display, after which 3 clones with binding signal to human and murine sortilin exceeding that of the parental G11 clone were subcloned into an expression vector and evaluated by SPR. Binding to murine and human sortilin was confirmed by SPR (Figure 15), and the kinetic constants of the new variants’ interactions with sortilin were estimated from a 1 : 1 Langmuir fit, as summarized in table 9.

Table 9. Kinetic constants for the interaction between human (hSort) and murine (mSort) sortilin and affibodies selected from the G11 library, as compared to the parental G11 clone.

Gi l io is defined by SEQ ID NO:52. The sortilin-binding polypeptide defined by SEQ ID NO:39 represents the amino acid sequence in positions X9-X35 of G1110.

G1113 is defined by SEQ ID NO:53. The sortilin-binding polypeptide defined by SEQ ID NO:40 represents the amino acid sequence in positions X9-X35 of G1113.

G11 is is defined by SEQ ID NO:54. The sortilin-binding polypeptide defined by SEQ ID NO:41 represents the amino acid sequence in positions X9-X35 of G1118. Based on the newly identified variants, a sequence motif for allowed amino acids in each of the variable positions in the anti-sortilin affibodies G11, Cl and F6 could be constructed (table 10).

In table 10 hereinbelow, “Allowed amino acids” are either observed in at least one of the original affibody clones G11, F6, and Cl selected against murine sortilin by phage display, or in at least one of the 3 confirmed binders from the E. coli display library based on the G11 cluster.

Table 10: Allowed amino acids in variable positions and framework positions with observed deviations from the designed amino acid, for anti-sortilin affibodies in the G11/C1/F6 cluster.

Fusion of the anti-sortilin affibodies A3 and G11 to investigate avidity effects

Given the fact that fusion of anti-sortilin affibodies G11, Cl and F6, belonging to the same sequence family, with the PGRN C-terminus did not lead to avidity effects, it was hypothesized that these affibodies’ epitopes might be overlapping that of the PGRN C- terminus. If so, fusion of anti-sortilin affibodies G11, Cl and F6 with A3 in a heterodimer format might lead to avidity effects in a similar way to fusion of A3 with PGRNcX*. To test this hypothesis, we produced and purified a construct of the format His6-ABD035-G4S-A3- (G₄S)₃-G11, and evaluated its binding to human sortilin by SPR (Biacore T200), as previously described.

As demonstrated in figure 16, genetic fusion of anti-sortilin affibodies A3 and G11 leads to avidity effects, with improved apparent affinity for human sortilin compared to either affibody alone. Apparent kinetic constants were estimated using a 1 :1 Langmuir fit of the reference-subtracted sensorgrams (Table 11). Notably, the apparent equilibrium dissociation constant for the A3-G11 fusion protein was about 53-fold and 16-fold lower than that of affibody A3 or G11 alone, respectively, at 25°C. At 37°C, the equilibrium dissociation constant for the A3-G11 fusion protein was about 45-fold and 22-fold lower than that of A3 and G11 alone, respectively.

Table 11 : Apparent kinetic constants for the interaction between human sortilin and anti- sortilin affibody monomers or dimers

We further investigated if the improved affinity of the A3-G11 heterodimer was retained after C-terminal truncation of the G11 affibody. Two derivatives of the A3-G11 heterodimer were produced and purified, in which the last helix of the G11 affibody was completely or partially removed. The formats of the produced proteins were His6-ABD035- G4S-A3-(G4S)3-G1 Ivi-S39, and Hise-ABD035-G4S-A3-(G4S)3-Gl lvi-L45, respectively. Binding to human sortilin was analysed by SPR (Biacore T200) using 50 nM sortilin, in principle as previously described.

As demonstrated in figure 17, truncation of the C-terminus of the G11 affibody of the A3-G11 heterodimer resulted in retained avidity effects as compared to both the intact A3-G11 dimer or either affibody alone. Notably, dissociation from sortilin of the truncated heterodimer candidates appeared even slower than that of the intact A3-G11 dimer.

Mutational study of the PGRN C-terminal peptide

Allowed amino acid variations in the PGRN C-terminus were investigated by SPR. Constructs of the format His6-Zwt-G4S-ABD035-PGRNcl8A3-XXX, where A3 denotes a deletion of the 3 C-terminal-most amino acids of the PGRNc peptide and XXX are 3 amino acids chosen from the set of XXX={GGG, GLL, NLL, PLL, QGL, QIL, QLG, QLI, QLV, QVL, YIL, YLL} , as well as His6-Z_wt-G4S-ABD035-(G₄S)3-QLL and Hise-Zwt^S- ABD035- PGRNcl8 and His6-Zwt-G4S-ABD035- PGRNcl8* were evaluated for binding to human and murine sortilin. Kinetic constants were estimated using a 1 : 1 Langmuir fit of the reference-subtracted sensorgrams (Table 12).

In line with previous reports of the 3 C-terminal-most amino acids of PGRN (PGRNc3, QLL) being essential for sortilin binding [29], replacement of QLL with GGG in the PGRNc background (His6-Zwt-G4S-ABD035-PGRNcl8A3-GGG) abolishes sortilin binding. However, the lack of measurable sortilin binding of His6-Zwt-G4S-ABD035-(G4S)3- QLL demonstrates that PGRNc3 in itself is not sufficient for sortilin binding (Figure 18-19). A longer PGRNc peptide (e.g. PGRNcl8, as shown in figure 18-19) or another compatible sortilin-binding unit (e.g. the anti-sortilin affibody A3, as shown for ABD-A3-(G4S)3- PGRNc3 in figure 3) is required for sortilin binding of PGRNc3.

In a PGRNcl8 background, replacements in position 3 from the C-terminus (C3) do not abolish sortilin binding when C2 and Cl are conserved, as demonstrated in figures 18-19. Retained binding upon C3 replacements by various diverse amino acids (G, N, P, Y) indicates that C3 might serve as a spacing amino acid, tolerant to wide variation. In contrast, there are stricter requirements regarding position C2, with maintained binding shown after substitution of L to amino acids I or V with similar characteristics (QIL, QVL), but largely lost binding when replaced by the less similar G (QGL). The importance of L in position Cl is shown by the strong negative effect of binding when replacing this with the dissimilar G (QLG) and only certain retention of binding after replacement with a similar amino acid, V or I (QLV, QLI).

Table 12. Kinetic constants for the interaction between human (hSort) and murine (mSort) sortilin and peptides derived from the PGRN C-terminal, as indicated. All constructs share the N-terminal His6-Zwt-G4S-ABD035- part. NA indicates too low signals for accurate determination of kinetic constants.

Terms, definitions and embodiments of all aspects of the present disclosure apply mutatis mutandis to the other aspects of the present disclosure.

Even though the present disclosure has been described with reference to specific exemplifying embodiments thereof, many different alterations, modifications and the like will become apparent for those skilled in the art.

Variations to the disclosed embodiments can be understood and effected by the skilled addressee in practicing the present disclosure, from a study of the drawings, the disclosure, and the appended claims. Furthermore, in the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality.

References

[1] Vieira RT. Epidemiology of early-onset dementia: a review of the literature. Clinical Practice & Epidemiology in Mental Health 2013;9:88-95. https://doi.org/10.2174/1745017901309010088.

[2] Rosen HJ, Gorno-Tempini ML, Goldman WP, Perry RJ, Schuff N, Weiner M, et al. Patterns of brain atrophy in frontotemporal dementia and semantic dementia. Neurology 2002;58: 198-208. https://doi.Org/10.1212/WNL.58.2.198.

[3] Bang J, Spina S, Miller BL. Frontotemporal dementia. The Lancet 2015;386: 1672-82. https://doi.org/10.1016/S0140-6736(15)00461-4.

[4] Rohrer JD, Guerreiro R, Vandrovcova J, Uphill J, Reiman D, Beck J, et al. The heritability and genetics of frontotemporal lobar degeneration. Neurology 2009;73: 1451-6. https://doi.org/10.1212/WNL.0b013e3181bf997a.

[5] Po K, Leslie FVC, Gracia N, Bartley L, Kwok JBJ, Halliday GM, et al. Heritability in frontotemporal dementia: more missing pieces? J Neurol 2014;261 :2170-7. https://doi.org/10.1007/s00415-014-7474-9.

[6] Seelaar H, Kamphorst W, Rosso SM, Azmani A, Masdjedi R, De Koning I, et al. Distinct genetic forms of frontotemporal dementia. Neurology 2008;71 : 1220-6. https://doi.org/10.1212/01.wnl.0000319702.37497.72.

[7] Moore KM, Nicholas J, Grossman M, McMillan CT, Irwin DJ, Massimo L, et al. Age at symptom onset and death and disease duration in genetic frontotemporal dementia: an international retrospective cohort study. Lancet Neurol 2020;19: 145-56. https ://doi . org/ 10.1016/S 1474-4422( 19)30394- 1.

[8] Pickering-Brown SM, Rollinson S, Du Plessis D, Morrison KE, Varma A, Richardson AMT, et al. Frequency and clinical characteristics of progranulin mutation carriers in the Manchester frontotemporal lobar degeneration cohort: Comparison with patients with MAPT and no known mutations. Brain 2008; 131 :721-31. https://doi.org/10.1093/brain/awm331.

[9] Cruts M, Gijselinck I, Van Der Zee J, Engelborghs S, Wils H, Pirici D, et al. Null mutations in progranulin cause ubiquitin-positive frontotemporal dementia linked to chromosome 17q21. Nature 2006;442:920-4. https://doi.org/10.1038/nature05017.

[10] Gass J, Cannon A, Mackenzie IR, Boeve B, Baker M, Adamson J, et al. Mutations in progranulin are a major cause of ubiquitin-positive frontotemporal lobar degeneration. Hum Mol Genet 2006;15:2988-3001. https://doi.org/10.1093/hmg/ddl241.

[11] Manuel e JM, Chen, Desguerre I, Marchant D, Abitbol M, Ricquier D, et al. Progranulin Null Mutations in Both Sporadic and Familial Frontotemporal Dementia. Hum Mutation 2006;28:846-55. https://doi.org/10.1002/humu.

[12] Bhandari V, Palfree RGE, Bateman A. Isolation and sequence of the granulin precursor cDNA from human bone marrow reveals tandem cysteine-rich granulin domains. Proc Natl Acad Sci U S A 1992;89: 1715-9. https://doi.Org/10.1073/pnas.89.5.1715.

[13] Bateman A, Bennett HPJ. Granulins: The structure and function of an emerging family of growth factors. Journal of Endocrinology 1998;158: 145-51. https://doi.Org/10.1677/joe.0.1580145.

[14] Bateman A, Belcourt D, Bennett H, Lazure C, Solomon S. GRANULINS, A NOVEL CLASS OF PEPTIDE FROM LEUKOCYTES. Biochem Biophys Res Commun 1990;173: 1161-8.

[15] Lee CW, Stankowski JN, Chew J, Cook CN, Lam YW, Almeida S, et al. The lysosomal protein cathepsin L is a progranulin protease. Mol Neurodegener 2017;12: 1-9. https://doi.org/10.1186/sl3024-017-0196-6. [16] Zhu J, Nathan C, Jin W, Sim D, Ashcroft GS, Wahl SM, et al. Conversion of Proepithelin to Epithelins. Cell 2002;111 :867-78. https://doi.org/10.1016/s0092- 8674(02)01141-8.

[17] Tanaka Y, Suzuki G, Matsuwaki T, Hosokawa M, Serrano G, Beach TG, et al. Progranulin regulates lysosomal function and biogenesis through acidification of lysosomes. Hum Mol Genet 2017;26:969-88. https://doi.org/10.1093/hmg/ddx011.

[18] Logan T, Simon MJ, Rana A, Cherf GM, Srivastava A, Davis SS, et al. Rescue of a lysosomal storage disorder caused by Gm loss of function with a brain penetrant progranulin biologic. Cell 2021;184:4651-4668. e25. https://doi.Org/10.1016/j.cell.2021.08.002.

[19] Van Damme P, Van Hoecke A, Lambrechts D, Vanacker P, Bogaert E, Van Swieten J, et al. Progranulin functions as a neurotrophic factor to regulate neurite outgrowth and enhance neuronal survival. Journal of Cell Biology 2008;181 :37-41. https://doi.org/10.1083/jcb.200712039.

[20] De Muynck L, Herdewyn S, Beel S, Scheveneels W, Van Den Bosch L, Robberecht W, et al. The neurotrophic properties of progranulin depend on the granulin E domain but do not require sortilin binding. Neurobiol Aging 2013;34:2541-7. http s : //doi . org/ 10.1016/j . neurobiol aging .2013.04.022.

[21] Gass J, Lee WC, Cook C, Finch N, Stetler C, Jansen-West K, et al. Progranulin regulates neuronal outgrowth independent of Sortilin. Mol Neurodegener 2012;7: 1-13. https://doi.org/10.1186/1750-1326-7-33.

[22] Baker M, Mackenzie IR, Pickering-Brown SM, Gass J, Rademakers R, Lindholm C, et al. Mutations in progranulin cause tau-negative frontotemporal dementia linked to chromosome 17. Nature 2006;442:916-9. https://doi.org/10.1038/nature05016.

[23] Meeter LHH, Patzke H, Loewen G, Dopper EGP, Pijnenburg YAL, Van Minkelen R, et al. Progranulin Levels in Plasma and Cerebrospinal Fluid in Granulin Mutation Carriers. Dement Geriatr Cogn Dis Extra 2016;6:330-40. https://doi.org/10.1159/000447738.

[24] Finch N, Baker M, Crook R, Swanson K, Kuntz K, Surtees R, et al. Plasma progranulin levels predict progranulin mutation status in frontotemporal dementia patients and asymptomatic family members. Brain 2009;132:583-91. https://doi.org/10.1093/brain/awn352.

[25] Ghidoni R, Benussi L, Glionna M, Franzoni M, Binetti G. Low plasma progranulin levels predict progranulin mutations in frontotemporal lobar degeneration. Neurology 2008;71 : 1235-9. https://doi.org/10.1212/01.wnl.0000325058.10218.fc.

[26] Petersen CM, Nielsent MS, Nykjaer A, Jacobsen L, Tommerup N, Rasmussen HH, et al. Molecular identification of a novel candidate sorting receptor purified from human brain by receptor-associated protein affinity chromatography. Journal of Biological Chemistry 1997;272:3599-605. https://doi.Org/10.1074/jbc.272.6.3599.

[27] Quistgaard EM, Madsen P, Groftehauge MK, Nissen P, Petersen CM, Thirup SS. Ligands bind to Sortilin in the tunnel of a ten-bladed P-propeller domain. Nat Struct Mol Biol 2009;16:96-8. https://doi.org/10.1038/nsmb.1543.

[28] Hu F, Padukkavidana T, Vsegter CB, Brady OA, Zheng Y, Mackenzie IR, et al. Sortilin-mediated endocytosis determines levels of the frontotemporal dementia protein, progranulin. Neuron 2010;68:654-67. https://doi.Org/10.1016/j.neuron.2010.09.034.

[29] Zheng Y, Brady OA, Meng PS, Mao Y, Hu F. C-terminus of progranulin interacts with the beta-propeller region of sortilin to regulate progranulin trafficking. PLoS One

2011;6. https://doi.org/10.1371/journal.pone.0021023.

[30] Carrasquillo MM, Nicholson AM, Finch N, Gibbs JR, Baker M, Rutherford NJ, et al. Genome-wide screen identifies rs646776 near sortilin as a regulator of progranulin levels in human plasma. Am J Hum Genet 2010;87:890-7. https://doi.Org/10.1016/j.ajhg.2010.l l.002.

[31] Miyakawa S, Sakuma H, Warude D, Asanuma S, Arimura N, Yoshihara T, et al. Anti- sortilinl Antibody Up-Regulates Progranulin via Sortilinl Down-Regulation. Front Neurosci 2020; 14. https://doi.org/10.3389/fnins.2020.586107.

[32] Kurnellas M, Mitra A, Schwabe T, Paul R, Arrant AE, Roberson ED, et al. Latozinemab , a novel progranulin - elevating therapy for frontotemporal dementia. J Transl Med 2023: 1-18. https://doi.org/10.1186/sl2967-023-04251-y.

[33] Lee WC, Almeida S, Prudencio M, Caulfield TR, Zhang YJ, Tay WM, et al. Targeted manipulation of the sortilin-progranulin axis rescues progranulin haploinsufficiency. Hum Mol Genet 2014;23: 1467-78. https://doi.org/10.1093/hmg/ddt534.

[34] Nilsson B, Moks T, Jansson B, Abrahmsen L, Elmblad A, Holmgren E, et al. A synthetic IgG-binding domain based on staphylococcal protein A. Protein Engineering, Design and Selection 1987;1 : 107-13. https://doi.org/10.1093/protein/L2.107.

[35] Uhlen M, Guss B, Nilsson B, Gatenbeck S, Philipson L, Lindberg M. Complete Sequence of the Staphylococcal Gene Encoding Protein A. J Biol Chem 1984;259: 1695-702.

[36] Nord K, Gunneriusson E, Ringdahl J, Stahl S, Uhlen M, Nygren P-A. Binding proteins selected from combinatorial libraries of an a-helical bacterial receptor domain. Nat Biotechnol 1997;15:772-7.

[37] Giang KA, Nygren P-A, Nilvebrant J. Selection of Affibody Affinity Proteins from Phagemid Libraries. In: Hust M, Lim TS, editors. Phage Display. Methods in Molecular Biology, vol 2702. Second edi, New York, NY: Humana; 2023, p. 373-92. https://doi.Org/https://doi.org/10.1007/978-l-0716-3381-6_19.

[38] Giang KA, Boxaspen T, Diao Y, Nilvebrant J, Kosugi-Kanaya M, Kanaya M, et al. Affibody -based hBCMA x CD 16 dual engagers for NK cell-mediated killing of multiple myeloma cells. N Biotechnol 2023;77: 139-48. https://doi.Org/10.1016/j.nbt.2023.09.002.

[39] Jonsson A, Dogan J, Herne N, Abrahmsen L, Nygren PA. Engineering of a femtomolar affinity binding protein to human serum albumin. Protein Engineering, Design and Selection 2008;21 :515-27. https://doi.org/10.1093/protein/gzn028.

[40] Andersson KG, Persson J, Stahl S, Lofblom J. Autotransporter-Mediated Display of a Naive Affibody Library on the Outer Membrane of Escherichia coli. Biotechnol J 2019;14: 1-8. https://doi.org/10.1002/biot.201800359.

[41] Trabjerg E, Abu-Asad N, Wan Z, Kartberg F, Christensen S, Rand KD. Investigating the Conformational Response of the Sortilin Receptor upon Binding Endogenous Peptide- and Protein Ligands by HDX-MS. Structure 2019;27: 1103-1113. e3. https://doi.Org/10.1016/j.str.2019.04.006.

[42] Tanimoto R, Morcavallo A, Terracciano M, Xu SQ, Stefanello M, Buraschi S, et al. Sortilin regulates progranulin action in castration-Resistant prostate cancer cells. Endocrinology 2015;156:58-70. https://doi.org/10.1210/en.2014-1590.

[43] Zhang et al. J Controlled Release 370: 468-478, 2024

[44] Roselli et al., Oncotarget 6: 10473-10486, 2015

[45] Klint S, Feldwisch J, Gudmundsdotter L, Dillner Bergstedt K, Gunneriusson E, Hoiden Guthenberg I, et al. Izokibep: Preclinical development and first-in-human study of a novel IL-17A neutralizing Affibody molecule in patients with plaque psoriasis. MAbs [Internet], 2023; 15(1).

[46] Meister SW, Hjelm LC, Dannemeyer M, Tegel H, Lindberg H, Stahl S, et al. An affibody molecule is actively transported into the cerebrospinal fluid via binding to the transferrin receptor. Int J Mol Sci. 2020;21(8): 1- 14.

Claims

1. A sortilin-binding polypeptide comprising the amino acid sequence:

X9X10X11X12X13X14X15X16X17X18X19X20X21X22X23X24X25X26X27X28X29X30X31X32X33X34X35

(SEQ ID NO: 1), wherein

X₉ is I, N, R, H, T, K, or Q, preferably N, R, Q, K, or I;

X10 is E, A, V, K, T, H, Y, D, Q, R, S, or I, preferably E, H, Q, R, Y, A, or V;

Xu is E, G, W, H, M, R, K, Y, A, I, Q, T, V, F, L, N, or D, preferably A, E, I, Q, V, W, R, G, or T;

X12 is A, V or T, preferably A;

X13 is G, R, S, K, Y, W, F, or Q, preferably G, F, K, Q, R, or Y;

X_His A, W, N, H, K, R, Y, F, Q, or G, preferably A, F, G, or R;

X15 is E, D or V, preferably E;

Xi6 is I, N or T, preferably I;

X17 is I, T, R, M, V, W, Y, L, or K, preferably I, R, V, W, or T;

Xi8 is Q, F, H, V, or Y, preferably F, Y or Q;

X19 is L, P or Q, preferably L;

X20 is P, S, Q or T, preferably P;

X21 is N, K, S or Y, preferably N;

X22 is L, M or Q, preferably L;

X23 is N or T;

X₂₄ is R, K, S, Q, L, E, or N, preferably E, K, N, S, or R;

X₂₅ is W, N, K, R, H, F, L, or Q, preferably F, H, K, N, R, or W;;

X26 is Q, P or H, preferably Q;

X27 is K, G, H, Q, S, T, A, W, or absent, preferably K, H, Q, or G;

X₂8 is G, W, Q, H, Y, M, or absent,;

X29 is A or T, preferably A;

X30 is F, I or Y, preferably F;

X31 is I, K, H, Y, W, F, S, T, or D, preferably I, W, D, H, K, S, or W;

X32 is V, H, M, F, R, or K, preferably H, M, K, or V;

X33 is S, G, T or I, preferably S;

X34 is L, S, F or V, preferably L, and

X35 is K, W, A, E, F, H, T, M, D, Y, G, Q, or R, preferably D, K, W, R, Q, M, or F.

2. The sortilin-binding polypeptide according to claim 1, comprising the amino acid sequence:

X9X10X11AX13X14EIX17X18LPNLNX24X25QX27X28AFX31X32SLX35 (SEQ ID NO:2).

3. The sortilin-binding polypeptide according to claim 1 or claim 2, wherein X9 is I, N, K, Q, or R, preferably I, N, or K;

X10 is E, H, Q, or R;

Xu is A, Y, I, W, L, N, E, or Q, preferably A, E, I, Q, V, or W;

X13 is G, K, R, W, F, Y, or Q, preferably G, F, K, Q, R, or Y;

Xuis W, Y, A, F, G, H, R, or Q, preferably A, F, G, or R ;

X17 is I, L, W, K, V, or R, preferably I, R, V, or W;

Xi8 is H, Y, F, V, or Q, preferably F, Q, or Y;

X24 is K, R, L, E, or N, preferably E, K, or R;

X₂₅ is W, F, L, N, H, K, or Q, preferably F, H, K, N, or W;

X27 is A, K, Q, H, or absent, preferably H, or K;

X28 is G, H, Y, or absent;

X31 is I, or W, preferably W;

X32 is V; and

X35 is D, K, T, or W, preferably D, K, or W.

4. The sortilin-binding polypeptide according to claim 1 or claim 2, wherein Xg is N, R, or Q;

X10 is Y, A, V, or H;

Xu is A, R, G, W, T, or V;

X13 is K, R, or F ;

X14 is G;

X17 is I, T, or V;

Xi8 is F, or Y;

X₂₄ is N, K, or S;

X₂₅ is R, K, or N;

X27 is H, Q, or G;

X28 is M, Y, W, or Q;

X31 is D, H, K, S, or W; X32 is M, K, or H; and

X35 is R, Q, M, D, or F.

5. The sortilin-binding polypeptide according to any one of the preceding claims, comprising the amino acid sequence as defined in any one of SEQ ID NO:4-7, SEQ ID NO:29-41, or a corresponding sequence having at least 70%, preferably at least 80 %, more preferably at least 90% sequence homology with SEQ ID NO:4-7 or SEQ ID NO:29-41.

6. An affibody comprising the sortilin-binding polypeptide according to any one of claims 1- 5.

7. The affibody according to claim 6, wherein said affibody is defined by any one of SEQ ID NO: 13-16, SEQ ID NO:42-54, or a corresponding sequence having at least 70%, preferably at least 80 %, more preferably at least 90% sequence homology with SEQ ID NO: 13-16 or SEQ ID NO:42-54.

8. A fusion protein or conjugate comprising:

- a first part comprising a sortilin-binding polypeptide according to any one of claims 1-5 or an affibody according to any one of claims 6-7; and

- a second part comprising at least one peptide adapted to enhance the binding to sortilin.

9. The fusion protein or conjugate according to claim 8, wherein the peptide of said second part comprises at least four amino acids, and wherein the C-terminal end amino acid of said second part is leucine (L) or isoleucine (I).

10. The fusion protein or conjugate according to claim 8 or claim 9, wherein said second part is arranged C-terminally of said first part.

11. The fusion protein or conjugate according to any one of claims 8-10, wherein said peptide of said second part comprises from 4 to 24, preferably from 6 to 18 amino acids and is defined by the sequence: X₅9X60X61X62X63X64X65X66X67X68X69X70X71X72X73X74X75X76X77X78X₇9X80X81X82 (SEQ ID NO:20), wherein

X59 to X78 are independently optional or are selected from any amino acid residue; X79 is selected from any amino acid residue;

Xso is G, N, P, Q, or Y, preferably Q, or Y;

Xsi is L, V, or I, preferably V, or I; and

X82 is L or I, preferably L.

12. The fusion protein or conjugate according to any one of claims 8-11, wherein said peptide of said second part is a peptide (PGRNc) derived from the C-terminus of progranulin, wherein said peptide is defined by any one of SEQ ID NO: 21-26.

13. The fusion protein or conjugate according to any one claims 8-12, wherein said first part comprises the sortilin-binding polypeptide according to claim 3.

14. The fusion protein or conjugate according to any one of claims 8-13, wherein said first part comprises the sequence as defined in SEQ ID NO:7, SEQ ID NO:29-38, or a corresponding sequence having at least 70%, preferably at least 80 %, more preferably at least 90% sequence homology with SEQ ID NO:7 or SEQ ID NO:29-38.

15. The fusion protein or conjugate according to any one of claims 8-14, wherein said first part is an affibody defined by SEQ ID NO: 16, SEQ ID NO:42-51 or a corresponding sequence having at least 70%, preferably at least 80 %, more preferably at least 90% sequence homology with SEQ ID NO: 16 or SEQ ID NO:42-51 .

16. The fusion protein or conjugate according to claim 14 or claim 15, wherein said peptide of said second part is the sortilin-binding polypeptide according to claim 4.

17. The fusion protein or conjugate according to claim 16, wherein said peptide of said second part comprises the sequence as defined by any one of SEQ ID NO:4-6, SEQ ID NO:39-41, preferably SEQ ID NON or SEQ ID NO:39-41, or a corresponding sequence having at least 70%, preferably at least 80 %, more preferably at least 90% sequence homology with SEQ ID NO:4-6 or SEQ ID NO:39-41.

18. The sortilin-binding polypeptide, affibody, fusion protein or conjugate according to any one of the preceding claims for use as a medicament.

19. A pharmaceutical composition comprising the sortilin-binding polypeptide, affibody, fusion protein or conjugate according to any one of the preceding claims.

20. The sortilin-binding polypeptide, affibody, fusion protein, conjugate or pharmaceutical composition according to any one of the preceding claims for use in a method of treatment of dementia, frontotemporal lobar degeneration (FTLD), amyotrophic lateral sclerosis (ALS), Parkinson’s disease (PD), Alzheimer’s disease (AD), limbic-predominant age-related transactive response DNA-binding protein 43 encephalopathy (LATE), neuropathic pain, or cancer.

21. The sortilin-binding polypeptide, affibody, fusion protein, conjugate or pharmaceutical composition for use according to claim 20, wherein said dementia is frontotemporal dementia (FTD).