Hyaluronan-carnosine conjugates inhibit Aβ aggregation and toxicity

Alzheimer’s disease is the most common neurodegenerative disorder. Finding a pharmacological approach that cures and/or prevents the onset of this devastating disease represents an important challenge for researchers. According to the amyloid cascade hypothesis, increases in extracellular amyloid-β (Aβ) levels give rise to different aggregated species, such as protofibrils, fibrils and oligomers, with oligomers being the more toxic species for cells. Many efforts have recently been focused on multi-target ligands to address the multiple events that occur concurrently with toxic aggregation at the onset of the disease. Moreover, investigating the effect of endogenous compounds or a combination thereof is a promising approach to prevent the side effects of entirely synthetic drugs. In this work, we report the synthesis, structural characterization and Aβ antiaggregant ability of new derivatives of hyaluronic acid (Hy, 200 and 700 kDa) functionalized with carnosine (Car), a multi-functional natural dipeptide. The bioactive substances (HyCar) inhibit the formation of amyloid-type aggregates of Aβ 42 more than the parent compounds; this effect is proportional to Car loading. Furthermore, the HyCar derivatives are able to dissolve the amyloid fibrils and to reduce Aβ-induced toxicity in vitro. The enzymatic degradation of Aβ is also affected by the interaction with HyCar.

Synthesis of Hy conjugates with Car Several methods of conjugating bioactive substances to hyaluronic acid have been reported in the literature 58 . We did not use any spacer linker to increase the resistance of the conjugate to carnosinase activity 60 ; therefore, we chose to link hyaluronic acid to Car through an amide bond between the carboxylic group of the d -glucuronic unit of the polysaccharide and the N-terminal amino group of Car. Accordingly, hyaluronic acid-Car conjugates were prepared following a synthetic route that involves the activation of the Hy-linked carboxylate groups and their subsequent conversion into amide bridges by reaction with the amino group of carboxyl-protected Car. The final products differ in terms of MW of the starting Hy (200 or 700 kDa) and Car loading (7, 10, 14, 35%), i.e., the percentage of D-glucuronic units drafted by the Car residues. Subsequently, NMR analysis of the HyCar conjugates was performed to ascertain the chemical conjugation and to calculate the Car loading percentage of each new HyCar derivative. Finally, the medium molecular weight (MMW) and the intrinsic viscosity (I.V.) (Supplementary Table S1 ) of each conjugate were determined, and the results were in accordance with those obtained from the NMR analysis 59 . The greater the loading percentage of Car was, the higher the MMW, as expected. Moreover, Car conjugation clearly affects the final viscosity: all the HyCar derivatives showed an I.V. value lower than that of the parent polymer compounds. This difference clearly increased as the MMW increased after Car derivatization. The structural characterization also confirmed that no fragmentation of the polysaccharide skeleton occurred as a consequence of the synthesis process.

The amyloid-type aggregation of Aβ starts with the formation of soluble and toxic oligomers; their dimensions increase during the process, and fibrillary and insoluble structures are formed at the end of this pathological pathway. Reducing or preventing the aggregation process of Aβ is one of the main challenges associated with attenuating or preventing the onset of AD. Therefore, all HyCar derivatives were tested for the in vitro self-induced aggregation of Aβ 42 by using thioflavin T (ThT), a fibril-sensitive dye. As a first step, the extent of dose-dependent aggregation of the HyCar(200) derivatives and their parent compounds was monitored at the end of the incubation period (40 h). As reported in Fig. 4 , the maximum ThT fluorescence variation due to amyloid aggregation was significantly reduced in the presence of Hy(200). In contrast, Car alone did not show any significant effect on the formation of ThT-sensitive species at the lower concentrations used (0.5 and 1 µM). However, both the underivatized Hy(200) polymer and the dipeptide (Car) were clearly able to partially inhibit the total amyloid species formation in a dose-dependent manner. The concentrations of Car were equivalent to the level of Car units in the samples containing HyCar(200)14. Figure 4 Final antiaggregant activity of HyCar derivatives towards the amyloid-type aggregation of Aβ. Relative fluorescence values of the samples containing Aβ 42 (15 µM) alone (CTRL), in the presence of the Hy(200) derivatives of Car (HyCar(200)7, HyCar(200)10, HyCar(200)14) or their parent compounds (Hy(200) or Car), incubated at 37 °C for 40 h. The concentration of Hy(200) and that of its Car conjugates (HyCar(200)7, HyCar(200)10, HyCar(200)14) varied from 0.5 to 10 µM, whereas the concentration range of Car (35–700 µM) is equivalent to the amount of Car in the samples containing HyCar(200)14. HyCar(200) derivatives inhibit the formation of amyloid species, and the decrease in fluorescence with respect to that of Aβ alone (CTRL sample) is directly proportional to their concentration. Interestingly, the higher the Car content covalently linked to the Hy(200) scaffold was, the better the inhibition activity. As the main result, HyCar(200)14 showed the best performance among the HyCar(200) derivatives in terms of reducing the amyloid-type aggregation of Aβ 42, and the higher concentrations of this compound yielded a final fluorescence increment as low as 10% compared to that shown by Aβ 42 alone. Then, kinetic measurements of amyloid aggregation were carried out in order further study the antiaggregant capabilities of the HyCar(200) derivatives. The maximum fluorescence gain ( F max – F 0 ) is proportional to the aggregation extent of Aβ, whereas the lag time ( t lag ) value represents the starting point of the amyloid-type aggregation of Aβ. As a consequence, the lower F max – F 0 is and the higher t lag is, the better the antiaggregant activity. When ThT is incubated with Aβ 42 , the amyloid-type aggregation mechanism results in a sigmoid-shaped fluorescence response (Fig. 5 a). The lag phase lasts 5.9 h (Supplementary Table S2 ), and the maximum fluorescence gain ( F max − F 0 ) reaches a value of 51.2. Figure 5 Kinetic trends of amyloid aggregation in the presence of HyCar derivatives. Kinetic profiles of Aβ 42 aggregation alone (CTRL, 15 µM), incubated with the (a) Hy(200) derivatives of Car (HyCar(200)7, HyCar(200)10, HyCar(200)14) or their parent compounds (Hy(200), Car or a mixture) or (b) with the Hy(700) derivative of Car (HyCar(700)35), their parent compounds (Hy(700), Car or a mixture). The concentrations of the tested compounds were as follows: (a) 10 µM (Hy(200), HyCar(200)7, HyCar(200)10, HyCar(200)14), 700 µM (Car), (b) 0.4 µM (Hy(700), HyCar(700)35) or 270 µM (Car). The concentration values of Hy and Car in the Hy(200) + Car mixture are equivalent to the amount of Hy and Car units in the sample containing HyCar(200)14, whereas the concentration values of Hy and Car in the Hy(700) + Car mixture are equivalent to the amount of Hy and Car units in the sample containing HyCar(700)35. For the treatment with Car, the kinetic parameters of the self-induced amyloid aggregation do not significantly change. When Hy(200) or a mixture of Hy(200) and Car was co-incubated with Aβ, the corresponding F max − F 0 values (17.4 and 16.4, respectively) were significantly lower than those fitted in the case of Aβ alone. Conversely, Hy(200) and Hy(200) + Car do not show any effect on the lag phase of the amyloid aggregation process. For HyCar(200)7, the fitted kinetic parameters do not significantly differ from those obtained in the case of the Hy(200) + Car mixture. HyCar(200)10 is able to delay amyloid aggregation with respect to HyCar(200)7. HyCar(200)14 greatly outperforms the antiaggregant activity exhibited by the Hy(200) derivatives with lower Car loading (HyCar(200)7 and HyCar(200)10). The ThT fluorescence signal over time becomes almost flat and thus cann

To characterize the morphological features of the aggregated amyloid forms during Aβ aggregation as a function of the Hy derivatives, samples from the aggregation reactions (Supplementary Figure S5 ) were analysed by using Atomic Force Microscopy (AFM). Figure 7 shows the AFM height micrographs of Aβ assemblies for freshly dissolved and 24 h-incubated solutions in MOPS buffer at 37 °C, either in the absence or presence of the different Hy derivatives. A mixed population of small oligomers was found for freshly dissolved Aβ 42 samples (Fig. 7 a). The formation of oligomer species could reasonably be attributed to the interaction of the monomer forms with the mica surface, as previously reported 62 . After 24 h of incubation, bundled assemblies of fibrils were visible for Aβ 42 alone (Fig. 7 b) and Aβ incubated in the presence of Hy(200) (Fig. 7 c) or Hy(700) (Fig. 7 d). In contrast, an evident fibril antiaggregant effect was detected for Aβ 42 co-incubated with HyCar conjugates, since Aβ 42 + HyCar(200)14 samples exhibited only oligomers (Fig. 7 e), while few coiled fibrils and oligomers were found for Aβ 42 + HyCar(700)35 (Fig. 7 f). Figure 7 AFM analysis of Aβ aggregation in the presence of HyCar. AFM topography images of Aβ 42 (15 μM) incubated at 37 °C alone for 0 h (a) or 24 h (b) or co-incubated for 24 h with 10 µM Hy(200) (c) , 0.4 μM Hy(700) (d) , 10 μM HyCar(200)14 (e) or 0.4 μM HyCar(700)35 (f) . AFM analyses were also performed to test the capability of the HyCar compounds to dissolve the preformed Aβ fibrils (Supplementary Figure S6 ). Indeed, for Aβ 42 alone, after 24 h of ageing after the aggregation period (40 h), amyloid structures were clearly visible, with individual fibrils that formed a tightly packed, highly ordered matrix of twisted fibrils laterally self-associated in large, micrometre-sized aggregates (Supplementary Figure S6 a). On the other hand, the preformed Aβ fibrils incubated with HyCar(200)14 (Supplementary Figure S6 b) or with HyCar(700)35 (Supplementary Figure S6 c) displayed disjointed fibrils and mixed networks/separated fibrils, thus confirming the results obtained by ThT measurements.

The potential role of Car in brain-related disorders has recently been reviewed 66 , suggesting that reduced plasma Car levels 67 characterize patients with probable AD, while the carnosinase level has been proposed as a CSF biomarker for early stages of AD 68 . To avoid or delay the carnosinase degradation effect on Car, different derivatives of the dipeptide have been synthesized by conjugating it with derivatives of vitamin E or small carbohydrates 69 – 71 . Our findings also show that two Hy chains with different molecular weights (200 and 700 kDa) are able to protect Car from the degradative action of carnosinase when the dipeptide is covalently drafted on the polymer to form HyCar conjugates. In contrast, the mixture of the parent compounds does not prevent or delay the enzymatic degradation of Car. Polymeric carbohydrates are among the most abundant components of the extracellular matrix, such as the GAG chains in proteoglycans (PGs) 72 . They are associated with different amyloid deposits 73 – 76 and are involved in the formation of fibrils. Sulphated GAGs stabilize mature fibrils against dissociation 77 and prevent their proteolytic degradation 78 ; at the same time, sulphated GAGs attenuate the neurotoxic effect of Aβ in cellular assays 79 . AD therapeutic approaches are usually designed to prevent Aβ fibrillogenesis through reduction of Aβ production, inhibition of fibril formation, or removal of fibril aggregates 80 . It has been suggested that GAG-mediated neuroprotection occurs due to the sequestration of Aβ and increased fibril formation 81 , 82 , consistent with the hypothesis that the generation of senile plaques in AD characterized by fibrils is a protective response aimed at reducing soluble Aβ neurotoxicity 79 . The promotion of Aβ fibril formation has been attributed to the negative charges of sulphated GAGs 76 and to other GAGs, including Hy, bearing carboxyl groups instead of sulphate groups 82 . Conversely, some findings indicate that no fibril formation occurs in the presence of Hy 21 . PGs and Aβs colocalize in senile plaques in the AD brain. MD simulation has shown that the GAGs interact by means of their negative charges with the Aβ domain, characterized by the presence of a basic residue cluster (HHQK) and a contiguous tyrosine, while chitosan, with its positive charges, interacts with the Aβ regions rich in Asp and Glu residues 82 . It has been suggested that low-molecular-weight heparins can perturb the interactions between PGs and Aβs, blocking or preventing the process of amyloidogenesis 21 . In on-pathway fibrillation, Aβ monomers give rise to fibrils through many metastable steps of oligomeric species formation characterized by β-rich structures 83 , 84 . Among natural inhibitors of Aβ amyloid formation, polyphenols such as epigallocatechin-3-gallate (EGCG) and curcumin have been extensively studied 22 , 85 , 86 . EGCG could interact with different residues of Aβ 42 (Phe4, Arg5, Phe19, Phe20, Glu22, Lys28, Gly29, Leu34-Gly37, and Ile41) and form hydrogen bonds with Aβ 85 , while curcumin was able to bind to the 12 th and 17 th to 21 st residues of Aβ 42 86 . Mono-curcumine, mono-EGCG and the dual-EGCG-curcumine conjugated with Hy (MW = 1 × 10 6 kDa) have been synthesized 87 , 88 . The mono-conjugated derivatives 87 inhibited Aβ fibril formation more strongly than free polyphenols, while Hy modified with the two polyphenols was more efficient than the single conjugates with EGCG and curcumine 88 . All the modified polyphenols decreased the amyloid cytotoxicity more efficiently when conjugated with Hy. The formation of Hy conjugate nanogels induced an isolation effect, hindering the interactions between Aβ molecules, while hydrophobic interactions of Aβ with EGCG or curcumin and electrostatic repulsion between the bound Aβ and like-charged Hy altered the Aβ monomeric species structures, giving rise to off-pathway aggregation. These two effects were correlated with the conjugated nanostructures that changed with the loading percentage of Hy. In this context, our results provide evidence that Hy(200) inhibits Aβ 42 aggregation in a dose-dependent manner (Fig. 5 ) and protects cells from the toxicity induced by amyloid oligomeric species. Hy(700) induces cell viability in the presence of amyloid species, while it is a modest antiaggregation agent, probably due to the lower concentrations employed (nanomolar range) in the ThT assay than those used for Hy(200) (micromolar range). Furthermore, both the low- and high-molecular-weight hyaluronan chains are able to dissolve fibril formation (Supplementary Figure S3 ). These protective effects of Hy(200) and Hy(700) found in our experiments are in contrast with those reported in recent years, where the hyaluronic acids did not affect Aβ aggregation and toxicity. The high MW of Hy (1 × 10 6 Da) employed to examine the Aβ antiaggregation effect of polyphenol conjugates with hyaluronic acids and the related difficulties in handling the n

Chemicals, reagents and common instrumentation All solvents and reagents, unless otherwise specified, were purchased from Sigma-Aldrich. Sodium hyaluronate (200 and 700 kDa) was a kind gift from Fidia Farmaceutici S.p. A (IT). 3-Hydroxy,1,2,3-benzotriazin-4(3H)-one (HOOBt) and trifluoroacetic acid (TFA) were purchased from AAPPTec and VWR, respectively. 3-( N -morpholino)-propanesulphonic acid (MOPS) buffer solution (1 mM, pH 7.4, supplemented with 0.003 M KCl and 0.14 M NaCl) was prepared from a powder (Amresco Biochemicals). Anhydrous methanol was kept for 24 h on 4 Å molecular sieves before use. Carnosine methyl-ester (Car-OMe) was synthesized in anhydrous methanol, as previously reported 94 . All the fluorescence and UV–Vis spectra were reordered by using a multimode microplate reader (Varioskan Flash, Thermo Scientific).

The medium molecular weight of each conjugate was determined by means of size exclusion chromatography on a Visctek GPCmax VE 2001 chromatographic system (Malvern) equipped with two TSK-GEL GMPWXL columns (7.8 mm ID × 300 mm; TOSOH BIOSCIENCE) installed in series. The system was coupled with three detectors placed in series: a refraction index detector, a light scattering detector and a four-capillary differential viscometer. Each sample was eluted using 0.1 M sodium nitrate in water containing 0.5 g/L sodium azide with a flow rate of 0.6 ml/min at 40 °C. Omnisec 4.1 software (Marvern Panalytical, https://www.malvernpanalytical.com/en/products/product-range/viscotek-range/viscotek-systems/viscotek-tdamax/accessories/omnisec-software ) was used for data acquisition and analysis. Triple-detector detection was performed on a Viscotek TDA 302 (Malvern).

Monomeric Aβ 42 was obtained using a variant of Zagorsky’s protocol 95 – 97 . Briefly, 1 mg of peptide was dissolved in TFA (1 ml). The solution was sonicated for 10 min, and TFA was then evaporated under a gentle stream of N 2 . One millilitre of HFIP was added to the peptide. After 1 h of incubation at 37 °C, the peptide solution was dried under a stream of N 2. This procedure was repeated 2 times. The dried peptide was then dissolved in HFIP (1 ml) and fast-frozen at -80 °C before lyophilization (10 h). The lyophilized sample was re-suspended in 5 mM NaOH solution in ultrapure water (Millipore, 18.2 MΩ·cm resistivity at 25 °C) to a final concentration of 1.5 mM and stored at − 80 °C.

Solutions containing Aβ 42 (15 µM) and ThT (45 µM) were incubated in MOPS buffer (50 mM, pH 7.4) with a black 96-well plate (Nalge-Nunc, Rochester, NY). After 40 h at 37 °C, the compounds of interest were added into the solutions, and the plates were kept at 37 °C for 24 h in a multiplate plate reader. Amyloid disaggregation was followed by measuring the ThT fluorescence emission at 480 nm after excitation at 450 nm. Aβ dissolution was calculated as the percent decrease in ThT after incubation with the compounds of interest.

The stored Aβ 42 solution was diluted to 15 μM in MOPS buffer (1 mM, pH 7.4) containing Hy200, Hy200 supplemented with Car (1.84 mM), HyCar35 (200), Hy700, Hy700 supplemented with Car (0.26 mM), HyCar35 (700) and Car (1.84 mM) at a hyaluronic acid concentration of 10 μM for the low molecular weight (200 kDa) and 0.4 μM for the high molecular weight (700 kDa). All the samples were incubated in an orbital shaker at 37 °C and 190 r.p.m. for 24 h. To assess the effect of hyaluronic acid on Aβ 42 aggregation, each measurement was carried out immediately after dilution of the peptide and after 4 and 24 h of incubation. To register AFM images, we used a previously described procedure 96 . Briefly, 10 μL of each sample was placed on freshly cleaved muscovite mica (Ted Pella, Inc.) and incubated for 5 min at room temperature. Then, samples were washed with 1 ml of ultrapure water, dried under a gentle nitrogen stream and immediately imaged. A Cypher AFM instrument (Asylum Research, Oxford Instruments, Santa Barbara, CA) operating in tapping AC-mode was used. Tetrahedral tips made of silicon and mounted on rectangular 30-μm-long cantilevers were purchased from Olympus (AT240TS, Oxford Instruments). The probes had nominal spring constants of 2 N/m and driving frequencies of 70 kHz. Images were scanned, and the sizes of the aggregates were measured using a free tool in the Asylum Research offline section analysis software program.

The MTT method with the undifferentiated SH-SY5Y cell line was employed for the cell viability assays. The day before the experiment, cells were seeded at a density of 2.5 × 10 4 cells per well (50% confluence) in full medium on Corning tissue culture treated 48-multiwell plates (Sigma-Aldrich, St. Louis, MO) until cellular adhesion was attained. Samples of Aβ 42 (15 µM) were pre-incubated for 6, 24 and 48 h at 37 °C alone or in the presence of HyCar(200)14, HyCar(700)35, Hy(200) and Hy(700) with or without Car. Cells were treated for 24 h with pre-incubated Aβ 42 (5 µM) in the presence of the studied compounds in full medium supplemented with 1% FBS. Then, cell viability was determined at 37 °C by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (Sigma-Aldrich, St. Louis, MO) method (MTT assay) as previously described 101 . After 90 min, the formazan salts produced by succinate dehydrogenase activity in live cells were solubilized by the addition of DMSO, and absorbance was measured at 570 nm by a microplate plate reader. The results were expressed as % of viable cells. At least three replicate experiments were conducted, and the averaged data with standard deviations were reported. Student's t-test was performed for statistical comparisons to analyse the variance, and a p-value less than 0.05 was considered to be statistically significant.