TECHNICAL FIELD

The present disclosure relates to the field of treatment of neurodegenerative disorders (NDs).

BACKGROUND

NDs are characterized by degradation of neurons which in turn impacts the cognitive or motor function, and examples include Alzheimer’s disease (AD), Frontotemporal Dementia (FTD), Amyotrophic Lateral Sclerosis (ALS), and Parkinson’s Disease (PD), all lacking efficient therapeutic treatment. The most prevalent is AD which affects around 60 million people in the world today. The disease-related cost reached in 2015 $818 billions and is expected to continue to increase with an expected aging population (Alzheimer’s Disease International, 2016).

WO2O16/131987 discloses a class of engineered polypeptides having a binding affinity for amyloid P (AP) peptides as well as use of such AP peptide binding polypeptides as therapeutic, prognostic and/ or diagnostic agents. In WO2O16/131987, the highest affinity for AP is demonstrated for an engineered polypeptide denoted ABPP095. This polypeptide is referred to as ZSYM 73 below.

SUMMARY

[0001] The objective of the present disclosure is to provide engineered polypeptides having binding affinity for AP that can be produced in higher yield than ZSYM 73. This objective is met by an AP peptide binding polypeptide that comprises an amino acid sequence selected from i) and ii), wherein: i) is SEQ ID NO:i - [LINK] - SEQ ID N0:2; ii) is SEQ ID NO:3 - [LINK] - SEQ ID NO:q; and [LINK] is a linker .

BRIEF DESCRIPTION OF THE FIGURES

Figure 1:

A) Affibody structure with randomized positions marked in black (pdb:2b89)

B) ZAb3 head-to-tail linked dimer in complex with amyloid beta (in black) (pdb:20TK). A cystine bridge between the two subunits and a linker between the subunits is shown.

C) Zseq_lib with randomized positions shown in black for the two subunits, and a connecting linker.

D) The different allowed randomisations in the rational design of Zseq_lib, as described per group position in each amino acid. The same distribution is allowed for both subunits.

Figure 2: Amino acid sequences discussed in the present disclosure.

Figure 3: Amino acid sequences of new selected sequestrins for the amyloid beta peptide. The amino acid sequence for ZSYM ₇ 3 (SEQ ID NO: 18) is included for comparison

Figure 4: Sensorgrams from surface plasmon resonance (SPR) measurements for the highest affinity clones A) ZSYM ₇ 3, B) Zseq_ab22, and C) Zseq_ab23. All three clones show a slow off rate from the target amyloid beta. Y-axis shows the relative response units (RU) and x-axis shows the time of the interaction. The protein is injected in duplicate in an 8-step titration from 342 nM to 30 nM in grey curves. The blank lines represent the fitting of the curves in order to calculate the kinetics. The experiment was conducted at 25°C with immobilized amyloid beta on different surfaces for each analyte. Figure 4D shows affinity data from fitting of the curves from duplicate readings of the injections from each sample and melting temperature (Tm) as determined by circular dichroism spectroscopy.

Figure 5: Circular Dichroism (CD) spectral scans between 195-260 nm for proteins in equimolar concentrations of 15.7 yM for A) ZSYM ₇ 3 B) Zseq_ab22 and C) Zseq_ab23, with and without co-incubation with amyloid beta (1-40). The black diamond dotted lines represent the change in signal upon co-incubation of amyloid beta and the sequestrin, demonstrating structural rearrangements upon binding. Zseq_ab22, Zseq_ab23, and ZSYM ₇ 3 shows complete refolding after T _m measurements both with and without amyloid beta. Figure 5D shows thermal melting point measurements of ZSYM ₇ 3, Zseq_ab22 and Zseq_ab23 with/without amyloid beta co-incubated with the sequestrins. For all samples, the T _m increases significantly when amyloid beta is present, indicating that the complex is stabilized in bound state. The T _m curves are normalized between unstructured and structured state at 221 nm. Figure 6:

A) Analysis of aggregation after 92 hours at 20 p,M for amyloid beta 1-40 or 1-42 with or without the addition of equal molar ratios of respective sequestrin, normalized to the signal of amyloid beta 1-42 at end point. The sequestrins are added separately and at two different pH. All constructs demonstrate aggregation inhibition.

B) Kinetic reading for 92 hours where end point is given in figure A. No aggregation is observed for samples with only sequestrins.

DETAILED DESCRIPTION

The present disclosure provides an AP peptide binding polypeptide, which comprises an amino acid sequence selected from i) and ii).

Sequence i) is SEQ ID NO:i - [LINK] - SEQ ID NO: 2. [LINK] is a linker, typically an amino acid sequence connecting the C-terminus of SEQ ID NO: 1 to the N-terminus of SEQ ID N0:2. SEQ ID NO:i and SEQ ID N0:2 are subsequences of Zseq_ab22 (SEQ ID NO:16), which is shown in the Examples section below to bind AP with high affinity. Further, Zseq_ab22 is shown in the Examples section to be expressed in a significantly higher yield than the prior art AP binder ZSYM ₇ 3 (SEQ ID NO: 18).

When sequence i) is SEQ ID NO:i - [LINK] - SEQ ID N0:2, the linker maybe hi) AEAKKLNDAQAPASSSSGSSSSGRASAGGE (SEQ ID NO: 13) or iv) an amino acid sequence that has at least 90% identity to hi). In the present disclosure, sequence identity is defined as in WO2O16/131987.

In one embodiment, sequence i) is SEQ ID NO:5 - [LINK] - SEQ ID NO: 6. Again, [LINK] is a linker, typically an amino acid sequence connecting the C-terminus of SEQ ID NO:5 to the N-terminus of SEQ ID NO:6. SEQ ID NO:s and SEQ ID NO:6 are longer subsequences of Zseq_ab22 (SEQ ID NO:16) than SEQ ID NO:i and SEQ ID N0:2 (see figure 2). In this embodiment, the linker maybe v) AEAKKLNDAQAPASSSSGSSSSGRAS (SEQ ID NO: 14) or vi) an amino acid sequence that has at least 90% identity to v).

In yet another embodiment, sequence i) is SEQ ID NO:9 - [LINK] - SEQ ID NO:io. Again, [LINK] is a linker, typically an amino acid sequence connecting the C- terminus of SEQ ID NO:9 to the N-terminus of SEQ ID NO:io. SEQ ID NO:9 and SEQ ID NO: 10 are longer subsequences of Zseq_ab22 (SEQ ID NO: 16) than SEQ ID NO:5 and SEQ ID NO:6 (see figure 2). In this embodiment, the linker maybe vii) SSSSGSSSSGRAS (SEQ ID NO: 15) or viii) an amino acid sequence that has at least 90% identity to vii).

Sequence ii) is SEQ ID NO:3 - [LINK] - SEQ ID NO:q. Again, [LINK] is a linker, typically an amino acid sequence connecting the C-terminus of SEQ ID NO:3 to the N-terminus of SEQ ID NO:q. SEQ ID NO:3 and SEQ ID NO:q are subsequences of Zseq_ab23 (SEQ ID NO:iy), which is shown in the Examples section below to bind AP with high affinity. Further, Zseq_ab23 is shown in the Examples section to be expressed in a significantly higher yield than the prior art AP binder ZSYM ₇ 3.

When sequence ii) is SEQ ID NO:3 - [LINK] - SEQ ID NO:q, the linker maybe hi) or iv) (as defined above).

In one embodiment, sequence ii) is SEQ ID NO:y - [LINK] - SEQ ID NO:8. Again, [LINK] is a linker, typically an amino acid sequence connecting the C-terminus of SEQ ID NO:? to the N-terminus of SEQ ID NO:8. SEQ ID NO:y and SEQ ID NO:8 are longer subsequences of Zseq_ab23 (SEQ ID NO:iy) than SEQ ID NO:3 and SEQ ID NO:4 (see figure 2). In this embodiment, the linker maybe v) or vi) (as defined above).

In yet another embodiment, sequence ii) is SEQ ID NO: 11 - [LINK] - SEQ ID NO: 12. Again, [LINK] is a linker, typically an amino acid sequence connecting the C- terminus of SEQ ID NO:n to the N-terminus of SEQ ID NO:12. SEQ ID NO:n and SEQ ID NO: 12 are longer subsequences of Zseq_ab23 (SEQ ID NO: 17) than SEQ ID NO:7 and SEQ ID NO:8 (see figure 2). In this embodiment, the linker maybe vii) or viii) (as defined above).

The present disclosure further provides a fusion protein or conjugate comprising at least a first part and a second part. The first part consists of an AP peptide binding polypeptide as defined above. The second part comprises a half-life-extending moiety, such as an Fc-binding moiety or an albumin-binding moiety. Various half- life-extending strategies for proteins are described in a review article by Kontermann (EXPERT OPINION ON BIOLOGICAL THERAPY, 2016 VOL. 16, NO. 7, 903-915). A specific example of an albumin-binding moiety is an albumin-binding affibody.

The present disclosure further provides a composition comprising:

- an AP peptide binding polypeptide, fusion protein or conjugate as defined above; and

- and at least one pharmaceutically acceptable excipient or carrier.

The AP peptide binding polypeptide, fusion protein, conjugate or composition described herein is typically used as a medicament.

In one embodiment, the AP peptide binding polypeptide defined above is used to clear AP from the blood of a patient having one of the conditions described below in hemodialysis, hemofiltration or hemodiafiltration. Accordingly, the present disclosure provides a dialyser or “hemofilter” comprising an AP peptide binding polypeptide defined above.

The AP peptide binding polypeptide, fusion protein, conjugate or composition defined above may be used in a method of treatment of an AP peptide associated condition selected from the group consisting of dementia, cognitive impairment, Alzheimer’s disease, cerebral amyloid angiopathy, Down’s syndrome, type II diabetes, primary and secondary systemic amyloidosis, familial amyloid polyneuropathy 1, glaucoma and age-related macular degeneration. The preferred condition to be treated is Alzheimer’s disease.

Similarly, the present disclosure provides a method of treatment of an AP peptide associated condition selected from the group defined above, wherein said method comprises administration of the AP peptide binding polypeptide, fusion protein, conjugate or composition according to any one of claims 1-8 to a human patient in need thereof, preferably a human patient suffering from Alzheimer’s disease or having an increased risk of developing Alzheimer’s disease.

EXAMPLES

METHODS

Library design and cloning

A library was designed based on the ZSYM ₇ 3 scaffold, consisting of two head-to-tail linked affibody subunits. Two randomized double stranded DNA oligonucleotides, encoding each of the two sequestrin subunits, were purchased from TWIST Bioscience (United States) for assembly into one library oligonucleotide by hybridization of overlapping bases in each subunit. The genes were flanked by BamHI and Sall restriction sites for subcloning into the modified pAffn phagemid. Each subunit of the library was amplified by PCR in 12 cycles using Phusion DNA polymerase (New England Biolabs, United States) and primers introducing ~6o overlapping bases into the two subunits for subsequent hybridization. The PCR products were purified using a PCR purification kit (Qiagen GmbH, Germany). Equimolar amounts of the oligo subunits were hybridized into one long library gene, which was subsequently PCR amplified in 10 cycles using Phusion DNA polymerase (New England Biolabs, United States). PCR products of correct length were purified by preparative gel electrophoresis (2% agarose gel) followed by purification using a QIAquick gel purification kit (Qiagen GmbH). Purified PCR products were digested by BamHI and Sall (New England Biolabs, United States) enzymes. The modified pAffii vector was digested by the same enzymes and purified by preparative gel electrophoresis.

Purified sequestrin-encoding DNA insert were ligated to the phagemid DNA vector using T4 DNA ligase (New England Biolabs, United States) in a 1:6 molar ratio of vector to insert. Ligated phagemid vector was purified by QIAquick PCR purification kit (Qiagen GmbH, Germany) before transformed into TG1 electrocompetent cells (Lucigen, United States). Transformed cells were directly amplified by cell growth after electroporation.

Phage stocks were created by standard procedures with superinfection from M13K07 phages (5X excess, New England Biolabs, United States) and precipitated using PEG / NaCl to yield phage titers around 10 ¹³ pfu/ml.

Library validation

An aliquot of transformed cells were individually analyzed with respect to: library size by titration, insert length by PCR amplification using DreamTaq DNA Polymerase and gel electrophoresis (Thermo Scientific, United States), and DNA sequence identity by Sanger sequencing (Microsynth SeqLab, Germany).

To validate the expression of sequestrins on the phage surface, a monoclonal phage ELISA was employed in which the C-terminal ABD fusion protein was monitored in an human serum albumin (HSA) capture assay. Individual clones were cultivated in 96-deep well format [TSBY, carbenicillin [100 pg/ml], 3O°C, 250 rpm, ON] and where re-inoculated [37°C, 250 rpm] until ODeoo reached 0.5-0.8. Superinfection with o.3xio ⁶ pfu/clone M13K07 (New England Biolabs, United States) at 37°C without rotation for 30 minutes proceeded induction with 1 mM IPTG and further cultivation ON [37°C, 250 rpm], under additional antibiotic pressure from kanamycin [30 pg/ml]. Harvested phage supernatants where incubated for 1 hour in wells with precoated HSA [5 pg/ml] or BSA [1 %w/v] in a 384-well format (Nunc, PS, Low binding, Hi-Edge, clear). The ELISA plate was blocked with 1% BSA prior to phage incubation to minimize background. The plate was washed with PBST (0.05% Tween- 20) before incubation with a mouse monoclonal M13 Bacteriophage Antibody (HRP) (Sino biological, China) according to manufacturer’s recommendations. Signal development was performed with Pierce™ TMB Substrate Kit (Thermo Scientific, United States), as per manufacturer’s instructions. After sufficient colorimetric development, the reaction was terminated by addition of 2 M H ₂ SO ₄ . Absorbance was measured at 450 nm on a CLARIOStar Plus (BMG Labtech, Germany). DNA sequences were identified by Sanger sequencing (Microsynth SeqLab, Germany), and further analyzed with the Geneious software (version 11.2, Biomatters LTD).

Phage selections

Amyloid beta (1-40) phage selections with sequestrins

The prepared phage stock from Zseq_lib was used in selections against C-terminally biotinylated API- ₄₀ peptide (AnaSpec, United States) in five rounds with decreasing amount of soluble target antigen (50 nM, 40 nM, 20 nM, 10 nM, 1 nM). The incubation temperature varied between the rounds (4°C ON for first round, 1 h RT for round 2-4, 1 h 45°C for the fifth round) before antigen was captured by Dynabeads M-280 Streptavidin beads (Invitrogen, United states). All rounds were preceded with a negative selection towards beads pre-blocked with BSA [5% w/v]. The captured phages were washed with PBSTB (0.1% Tween-20, 3% BSA) and the number of washing steps was increased with each selection round (2x1 min, 4x1 min, 5x3mm, 5x6 min, 4x6 min + lx 2h+ix6min), where the last wash was done in PBS buffer. Phage eluates were obtained by incubation for 10 min in Glycine-HCl [0.1 M, pH 3.0], followed by neutralization by Tris-HCl [1 M, pH 8.0]. Eluates were amplified by infecting IOOX excess of XLiBlue E. coli cells and plated on Aquare BioAssay Dish (Corning, United States) with TBAB agar supplemented with D-glucose [2%], carbenicillin [100 pg/ml], and tetracycline [10 pg/ml] and amplified by cell growth at 37°C ON. The agar plate colonies were recovered by scraping the plate and dissolving the cells in TSBY before continuing cultivation in a suspended format. The selection process was continuously monitored by titration of all steps of the selection procedure.

Phage ELISA and clone selection

To monitor the enrichment of amyloid-beta binding sequestrins during the selection, a phage-based ELISA was performed on polyclonal phage stocks as well as on individual phage clones after each selection round. All phage stocks were diluted to the same concentration before incubating with target antigen.

The ELISA plates were prepared in a 384-well format (Nunc, PS, Low binding, Hi- Edge, clear), with four wells per sample. Two wells per clone were coated with HSA [5 pg/ml] or BSA [1 % w/v], and two wells per clone were precoated with streptavidin [5 pg/ml] and thereafter coated with biotinylated amyloid-beta peptide [1 pg/ml] or BSA [1 % w/v]. Signals were normalized to the HSA signal and the background signals from streptavidin/BSA-coated wells were subtracted. The DNA sequence for clones showing binding to amyloid-beta peptide in the ELISA was determined by Sanger sequencing (Microsynth SeqLab, Germany), and sequences were analyzed with the Geneious software (version 11.2, Biomatters LTD, New Zealand).

Expression and purification of soluble sequestrins

Clones representing different clusters in the phylogenic tree were selected for further characterization. DNA encoding the selected clones were amplified from phagemids with primers designed for the In-Fusion HD cloning kit (Takara Bio Europe), as per manufacture’s recommendations into the pET-26b(+)-vector for periplasmic production with a C-terminal Hise-tag in E. coli. Subcloned sequestrins in the pET- 26b(+)-vector were transformed to the E. coli strain BL2i(DE3) by heat shocked for protein expression. Cultivations were started at an ODeoo of 0.1 AU in TSBY with kanamycin [25 pg/ml] and grown until ODeoo reached 1.0 AU at 37°C before induction with 1 mM IPTG and cultivation at 25°C for 16 hours. Cell were harvested and resuspended in IMAC running buffer [50 mM NaP, 300 mM NaCl, 15 mM imidazole, pH 7.4] before proceeding with purification.

The harvested cells were lysed by sonication using a Vibra-Cell VCX 130 sonicator (Sonics, United States) and cell debris was removed by centrifugation and filtration. The lysate was loaded on an equilibrated HisPur Cobalt Resin (Thermo Scientific, United States) and washed with running buffer. The sample was isocraticly eluted with the running buffer supplemented with 150 mM imidazole. Fractions of eluates containing protein according to measurement with the Pierce™ BCA Protein Assay Kit (Thermo Scientific, United States), performed as per instructions of the manufacturer, were pooled and buffer exchanged to PBS using PD 10 desalting columns (Cytiva, United States). Protein concentration, size and purity was estimated using SDS-PAGE analysis (NuPAGE Bis-Tris 4-12 %, Invitrogen, United States). Molecular mass was determined by mass spectrometry using a Thermo Ultimatesooo Bruker Impact II system connected to a ProSwift RP-4H, 1x50 mm column (Thermo Fisher, United States) using a linear gradient elution with acetonitrile [3 % to 95 %], supplemented with 0.1 % formic acid.

Circular dichroism spectroscopy

The secondary structure content of the sequestrins was analysed by circular dichroism spectroscopy using the Chirascan system (Applied Photophysics, United Kingdom) with a 1 mm High precision cell (110-1P-40 cuvettes, Hellma Analytics, Germany). Five wavelength scans were recorded and averaged between 195 nm and 260 nm at 20°C on protein samples of 0.2 mg/ml in PBS.

The melting temperatures were determined by using a temperature gradient of i°C per minute for five average readings at 221 nm. Refolding was analyzed by repeating the spectral scan after the sample had been subjected to heat treatment and cooled down to 20 °C. The spectra from before and after thermal heating was compared to assess the refolding of each protein.

For analyzing the secondary structure during interaction with target peptide, equimolar concentrations of respective sequestrin and amyloid-beta peptide was coincubated and analyzed by circular dichroism spectroscopy. Secondary structure content was approximated by BeStSel algoritm (Micsonai et al., 2015). Biosensor analysis of interaction between API- _4O and sequestrins

Surface Plasmon Resonance for A fi- ₄₀ and sequestrins

Surface Plasmon Resonance (SPR) was used to analyse binding between sequestrins and AP1-40 target antigen on a Biacore 8K instrument (Cytiva, United States). Here, a Series S SA sensor chip (Cytiva, United States) was immobilized with 120 RU (response units) of API- ₄₀ (biotinylated) as per the manufacturer’s instructions, and run in PBST (0.05% Tween-20) at 25°C. Sequestrins were injected in a multi-cycle analysis using an 8-step 1:2 dilution series from 342 nM to 30 nM in duplicate.

Ztaq ₃ 638-(G ₄ S)2-Ztaq ₃ 638-ABD was used as control. Respective analyte was injected for 200 seconds at 30 pl/min and the dissociation was thereafter recorded for 600 seconds. Surfaces were regenerated with 10 mM HC1 for 35 seconds and stabilized for 30 sec. The results were evaluated using the Multi-cycle kinetics method - 1:1 binding, by the Biacore Insight Evaluation Software (Version 2.0.15, Cytiva, United States).

In a second analysis, the chip was immobilized using similar conditions but on coating densities of 80 RU and analyzed in a 1:1 dilution series ranging 300 nM to 75 nM at 37°C in duplicate.

Method aggregation assay

Aggregation inhibition assays were performed with amyloid beta 1-40 (#AS-24236, AnaSpec, Australia) and amyloid beta 1-42 (# AS24224, AnaSpec, Australia), using the sequestrins Zseq_ab22, Zseq_ab23, and ZSYM73 as inhibitors for aggregation at an equimolar ratio of 20 pM for respective amyloid-beta peptide and sequestrin. All proteins were thawed on ice and spun down at 10,000 xg for 5 min at 4°C to remove precipitates. Each sample was prepared in triplicate in a volume of 60 pl per reaction by adding PBS (ABI- ₄₀ : 10 mM Phosphate/150 mM NaCl, pH=7.2; ABI- ₄₂ : 10 mM Phosphate/150 mM NaCl, pH=8.o), 20 pM Thioflavin T (ThT) dye, and 20 pM sequestrin before vortexing for 3 sec and kept on ice. When all samples were prepared, respective amyloid-beta peptide was added at 20 pM concentration to the samples vortexed for 1 sec. The samples were added to a 384-well plate (Nunc, PS, Low binding, Hi-Edge, clear) with surrounding wells filled with PBS, and sealed to minimize evaporation during analysis. The fluorescence intensity was immediately analyzed in a CLARIOStar Plus (BMG Labtech, Germany) with excitation at 440 nm and emission at 480 nm. The readings were done during 999 cycles with a cycle time of 325 sec at 37°C with 50 flashes with 0.1 s settling time per well, and using 1500 gain adjustment. The plate was shaken in an orbital motion 500 rpm for 1 sec before each measurement. Analysis were done in the Mars analysis software and replicates were averaged, and background signals from wells containing only PBS were subtracted.

RESULTS

Scaffold design and functionality of Zseq_lib

The sequestrin library was based on the amino acid sequence of the previously described head-to-tail dimeric amyloid beta binding affibody molecule, denoted ZSYM ₇ 3 (Lindberg et al., 2015). In resemblance to ZSYM ₇ 3, the sequestrin library scaffold was designed to comprise (1) a disulphide bond between the two subunits, (2) an eleven amino acid truncation of the unstructured N-terminus, as it has been shown to increase binding to amyloid beta (Lindgren 2010; Lindberg 2013; Lindberg 2015), as well as (3) a flexible (S ₄ G) ₂ linker between the two subunits to allow for correct folding with minimal steric hindrance.

The positions that were targeted for randomization, constitute the binding surface of ZSYM ₇ 3 in complex with Af (see figure 1B,1C). In total, 22 positions (11 in each subunit) were subjected to diversification. In principle, each randomized position was designed to contain the codon for the wild type amino acid together with a mix of codons for a pre-determined pool of the remaining amino acids. In 18 of the diversified positions (nine in each subunit), an average proportion of 61% of the original codon was combined with a mix of approximately 39% of mutation codons. Diversification of residues that contribute to the hydrophobic core of the complex was restricted to codons for hydrophobic amino acids in order to preserve the hydrophobic interface of the subunits. Residues that were subjected to randomization in the beta strands were allowed to vary between all remaining amino acids except for cysteine. The remaining diversified residues were allowed to vary between any amino acid except for proline, glycine or cysteine. In addition to these 18 randomized positions, four residues (two in each subunit) were subjected to partial randomization and allowed to vary between the original amino acid and the codon for one other amino acid to a 50:50 proportion (see figure 1D). The two library subunits were purchased from TWIST Bioscience (United States) as separate DNA oligonucleotides and amplified to contain approximately 60 overlapping bases for hybridization into one long assembled library oligonucleotide.

The constructed library Zseq_lib in E. coli yielded 5x10 9 individual clones.

Phage display selection of sequestrins towards amyloid beta 1-40

The phage-displayed library was used in selections towards C-terminally biotinylated AP1-40 peptide. The selection process was monitored by titration of phage outputs and inputs for each round. After the third selection round, a clear enrichment was seen where 50 % of the input phages were recovered after extensive washing. At this round, only full-length sequestrins were observed in PCR screening of the phage eluate, indicating a successful selection.

Enrichment of target-binding clones during the selection was monitored by a polyclonal phage ELISA. The phage pool showed an increased signal towards target AP1-40 peptide with increasing selection rounds, and correlated to the observed phage enrichment. A monoclonal phage ELISA performed after the third and fourth selection round showed positive target binding signal, supporting the results from phage titrations and the polyclonal ELISA. The amino acid sequence of individual clones from the selections was analyzed by DNA sequencing. The sequences showed an average of around 4 mutations per sequestrin and the mutations were predominantly located in the second helix. None of the new clones had the same amino acid sequence as the original binder ZSYM ₇ 3, of which the library is based on. Based on the reoccurrence of clones from the sequencing, 13 clones were selected for further characterization.

Expression and cloning of sequestrins targeting amyloid beta

Selected clones from the selection towards the API- ₄₀ peptide were subcloned into a pET26b(+)-vector with a C-terminal hexa-histidine tag for purification (further denoted as Zseq_ab-His6). The yield after production in E. coli BL21* cells was compared to the yield for the original binder ZSYM ₇ 3 and most of the new variants showed a trend towards higher protein yields (see table 1). Zseq_abi which is closest in identity to ZSYM ₇ 3, with only two amino acid substitutions in the second subunit, yielded only a small increase in production yield. The protein purity and size were analyzed using SDS-PAGE. A single band was seen at around 14 kDa corresponding with the expected molecular weight of approximately 13 kDa for all loaded proteins. Zseq_ab8 had a relatively low yield and showed a faint band at the expected size in SDS-PAGE. The mass of the proteins was analyzed using MS and correlated well to the expected sizes for the majority of the new variants.

Table 1, Production yields, expected mass and observed mass.

Construct Yield protein per Fold Mw _exp ected [Da] Mwobserved [Da]

Zseqlib clone 100 ml culture improved

(mg) expression compared to ZSYM73- His6

Zs¥M73-His6 2.8 ± 1.9

Zseq_abi5-His6 10.0 ± NA 2.9 12 538 12 537

Zseq_ab2O-His6 6.7 ± NA 1.9 12 543 12 542

Zseq_ab22-His6 6.7 ± 1.4 2.4 12 528 12 527

Zseq_ab23-His6 6.1 ± 1.3 2.2 12 606 12 605

Biosensor analysis of selected sequestrin variants targeting amyloid beta

The affinity between the selected sequestrins and the biotinylated amyloid beta 1-40 was evaluated in an SPR-based biosensor assay where the biotinylated amyloid-beta peptide was immobilized on a streptavidin chip. The original binder ZSYM ₇ 3 with an affinity of 340 pM in solution (Lindberg et al., 2015) was included for comparison. The equilibrium dissociation constant (KD) for ZSYM ₇ 3 was measured to around 3 nM which is lower compared to the previously reported affinity. However, in the previous study, the affinity was measured in solution without immobilizing the amyloid beta. Given the complex structural rearrangements of both the amyloid beta peptide and the sequestrin upon binding, it is likely that the affinity is underestimated in the SPR assay. Although the affinities for the amyloid beta variants are likely to be higher if measured in solution, the comparison to ZSYM ₇ 3 allows for a relative ranking of the candidates. Zseq_ab22 and Zseq_ab23 showed a similar affinity for the amyloid-beta peptide (KD around 1-3 nM) compared to ZSYM ₇ 3- Zseq_abi5 and Zseq_ab20 showed a significantly lower affinity for the amyloid-beta peptide (KD > 30 nM) compared to ZSYM ₇ 3. Based on the results, Zseq_ab22 and Zseq_ab23 were selected for further characterizations.

Circular dichroism for secondary structure determination of sequestrins

Structure and thermal stability ofZseq_ab

Circular dichroism (CD) spectroscopy was used to verify that the secondary structure content of Zseq_ab22 and Zseq_ab23. Both proteins showed the same characteristic alpha helical spectrum (see figure 5) as has been demonstrated previously for the ZAb3 binder (Hoyer and Hard, 2008) and ZSYM ₇ 3 (Lindberg et al., 2015). Moreover, both sequestrins demonstrated complete refolding after heat treatment. The thermal stability (T _m ) was estimated by variable temperature measurement, resulting in a T _m of 4i°C for Zseq_ab22 and a T _m of 48°C for Zseq_ab23 (see figure 5D).

When co-incubating the sequestrins with amyloid beta, an increase in T _m is observed, indicating that target binding stabilises the complex.

Results aggregation assay

Thioflavin T (ThT) binds to beta rich structures and is commonly used when staining for amyloid beta aggregation as well as fibril formation during aggregation. The aggregation of amyloid beta 1-40 and 1-42, respectively, were monitored for 92 hours at 37°C by measuring ThT fluorescence. The results verified the aggregation propensity of amyloid beta 1-40 and 1-42 (figure 6). Co-incubation of either amyloid beta peptide with a 1:1 molar ratio of Zseq_ab22 or Zseq_ab23, respectively, inhibited the aggregation process for both amyloid beta peptides. The analysis also demonstrated that the sequestrins are not aggregating.

REFERENCES

Alzheimer’s Disease International. World Alzheimer Report 2016 https://www.alz.co.uk/research/world-report-2o16 (2016).

Andersen, Jan Terje et al., “Extending Half-life by Indirect Targeting of the Neonatal Fc Receptor (FcRn) Using a Minimal Albumin Binding Domain,” The Journal of biological chemistry, vol. 286, no. 7, pp. 5234-5241, 2011, doi: 10.1074/jbc.Mno.164848. Ayalon, Gai et al., “Antibody semorinemab reduces tau pathology in a transgenic mouse model and engages tau in patients with Alzheimer’s disease,” Science translational medicine, vol. 13, no. 593, p. eabb2639, 2021, doi: 10.1126/ scitranslmed.abb2639.

Ballard, Clive et al., “Drug repositioning and repurposing for Alzheimer disease,” 2020, Nature reviews Neurology, vol. 16, no. 12, doi: 10.17863/CAM.55748.

Boutajangout A et al., (2019) Affibody-Mediated Sequestration of Amyloid P Demonstrates Preventive Efficacy in a Transgenic Alzheimer’s Disease Mouse Model.

Front. Aging Neurosci. 11:64. doi: 10.3389/fnagi.2Oi9.00064

Bauzon, Justin, Lee, Garam, and Cummings, Jeffrey, “Repurposed agents in the Alzheimer’s disease drug development pipeline,” Alzheimer's research & therapy, vol. 12, no. 1, pp. 1-98, 2020, doi: 10.1186/S13195-020-00662-X.

Chen, Guo-Fang et al., “Amyloid beta : structure, biology and structure-based therapeutic development,” Acta pharmacologica Sinica, vol. 38, no. 9, pp. 1205- 1235, 2017, doi: 10.10387aps.2017.28.

Cummings, Jeffrey et al., “Alzheimer's disease drug development pipeline: 2018,” Alzheimer's & dementia : translational research & clinical interventions, vol. 4, no.

1, pp. 195-214, 2018, doi: 10.10167j.trci.2018.03.009.

Gronwall, Caroline et al., “Selection and characterization of Affibody ligands binding to Alzheimer amyloid P peptides,” Journal of biotechnology, vol. 128, no. 1, pp. 162- 183, 2007, doi: 10.10167j.jbi0tec.2006.09.013.

Hardy, John A. and Higgins, Gerald A., “Alzheimer's Disease: The Amyloid Cascade Hypothesis,” Science (American Association for the Advancement of Science), vol. 256, no. 5054, pp. 184-185, 1992, doi: io.H26/science.i566o67.

Hoyer, Wolfgang; Hard, Torleif., “Interaction of Alzheimer’s AP Peptide with an Engineered Binding Protein— Thermodynamics and Kinetics of Coupled Folding- Binding,” Journal of molecular biology, vol. 378, no. 2, pp. 398-411, 2008, doi: io.ioi6/j.jmb.2oo8.o2.o4o. W. Hoyer et al., “Stabilization of a P-Hairpin in Monomeric Alzheimer's Amyloid- P Peptide Inhibits Amyloid Formation,” Proceedings of the National Academy of Sciences - PNAS, vol. 105, no. 13, pp. 5099-5104, 2008, doi: io.iO73/pnas.O7ii73iiO5.

Jonsson, A. et al., “Engineering of a femtomolar affinity binding protein to human serum albumin”. Protein Eng. Des. Sei. 21, 515-527, 2008. doi: io.io93/protein/gzno28

Lashuel, Hilal A., “Rethinking protein aggregation and drug discovery in neurodegenerative diseases: Why we need to embrace complexity? ” Current opinion in chemical biology, vol. 64, pp. 67-75, 2021, doi: io.ioi6/j.cbpa.2O2i.O5.oo6.

Lindberg, Hanna et al., “A truncated and dimeric format of an Affibody library on bacteria enables FACS-mediated isolation of amyloid-beta aggregation inhibitors with subnanomolar affinity,” Biotechnology journal, vol. 10, no. 11, pp. 1707-1718, 2015, doi: io.ioo2/biot.2oi5ooi3i.

Long, Justin M and Holtzman, David M, “Alzheimer Disease: An Update on Pathobiology and Treatment Strategies,” Cell (Cambridge), vol. 179, no. 2, pp. 312- 339, 2019, doi: io.ioi6/j.cell.2oi9.o9.ooi.

Micsonai, Andras et al., “Accurate secondary structure prediction and fold recognition for circular dichroism spectroscopy,” Proceedings of the National Academy of Sciences - PNAS, vol. 112, no. 24, pp. E3095-E3103, 2015, doi: io.io73/pnas.i5oo85iii2.

Panza, Francesco et al., “Amyloid-P immunotherapy for alzheimer disease: Is it now a long shot?,” Annals of neurology, vol. 85, no. 3, pp. 303-315, 2019, doi: io.ioo2/ana.254io.

Swanson, Chad J et al., “A randomized, double-blind, phase 2b proof-of-concept clinical trial in early Alzheimer's disease with lecanemab, an anti-AP protofibril antibody,” Alzheimer's research & therapy, vol. 13, no. 1, pp. 80-80, 2021, doi: 10.1186/S13195-021-00813-8. Tolar, Martin et al., “Aducanumab, gantenerumab, BAN2401, and ALZ-801— the first wave of amyloid-targeting drugs for Alzheimer’s disease with potential for near term approval,” Alzheimer's research & therapy, vol. 12, no. 1, pp. 1-95, 2020, doi: 10.1186/S13195-020-00663-W.

Yadav, Daniela Bumbaca et al., “Widespread brain distribution and activity following i.c.v. infusion of anti-P-secretase (BACE1) in nonhuman primates,” British journal of pharmacology, vol. 174, no. 22, pp. 4173-4185, 2017, doi: io.im/bph.14021.

Zhong, X., D’Antona, A.M. “Recent Advances in the Molecular Design and Applications of Multispecific Biotherapeutics”. Antibodies, vol. 10, no. 13, 2021, doi.org/ io.339o/antibioo2ooi3