Coassembly of Macrocyclic Amphiphiles for Anti-β-Amyloid Therapy of Alzheimer’s Disease

Abstract

Open AccessCCS ChemistryRESEARCH ARTICLE1 Sep 2021Coassembly of Macrocyclic Amphiphiles for Anti-β-Amyloid Therapy of Alzheimer's Disease Yu-Chen Pan†, Hui Wang†, Xinxin Xu†, Han-Wen Tian, Hong Zhao, Xin-Yue Hu, Yu Zhao, Yang Liu, Guihua Ding, Qingbin Meng, Bart Jan Ravoo, Tao Zhang and Dong-Sheng Guo Yu-Chen Pan† College of Chemistry, Key Laboratory of Functional Polymer Materials, Ministry of Education, State Key Laboratory of Elemento-Organic Chemistry, Nankai University, Tianjin 300071 , Hui Wang† College of Life Sciences, Key Laboratory of Bioactive Materials, Ministry of Education, Nankai University, Tianjin 300071 , Xinxin Xu† College of Life Sciences, Key Laboratory of Bioactive Materials, Ministry of Education, Nankai University, Tianjin 300071 , Han-Wen Tian College of Chemistry, Key Laboratory of Functional Polymer Materials, Ministry of Education, State Key Laboratory of Elemento-Organic Chemistry, Nankai University, Tianjin 300071 , Hong Zhao College of Chemistry, Key Laboratory of Functional Polymer Materials, Ministry of Education, State Key Laboratory of Elemento-Organic Chemistry, Nankai University, Tianjin 300071 , Xin-Yue Hu College of Chemistry, Key Laboratory of Functional Polymer Materials, Ministry of Education, State Key Laboratory of Elemento-Organic Chemistry, Nankai University, Tianjin 300071 , Yu Zhao College of Chemistry, Key Laboratory of Functional Polymer Materials, (Ministry of Education, State Key Laboratory of Medicinal Chemical Biology), Nankai University, Tianjin 300071 , Yang Liu College of Chemistry, Key Laboratory of Functional Polymer Materials, (Ministry of Education, State Key Laboratory of Medicinal Chemical Biology), Nankai University, Tianjin 300071 , Guihua Ding State Key Laboratory of Toxicology and Medical Countermeasures, Beijing Institute of Pharmacology and Toxicology, Beijing 100850 , Qingbin Meng State Key Laboratory of Toxicology and Medical Countermeasures, Beijing Institute of Pharmacology and Toxicology, Beijing 100850 , Bart Jan Ravoo Organic Chemistry Institute and Center for Soft Nanoscience (SoN), Westfälische Wilhelms-Universität Münster, Münster 48149 , Tao Zhang *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] College of Life Sciences, Key Laboratory of Bioactive Materials, Ministry of Education, Nankai University, Tianjin 300071 and Dong-Sheng Guo *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] College of Chemistry, Key Laboratory of Functional Polymer Materials, Ministry of Education, State Key Laboratory of Elemento-Organic Chemistry, Nankai University, Tianjin 300071 https://doi.org/10.31635/ccschem.020.202000561 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail Based on the amyloid hypothesis, anti-β-amyloid (Aβ) therapy has dominated clinical trials for the prevention and treatment of Alzheimer's disease (AD) in recent years. A key element of this strategy is the interaction between therapeutic agents and Aβ. However, the design and development of artificial receptors that may render selective and strong recognition toward Aβ remains a huge challenge because of the complexity and size of peptide guests and their flexible conformation. To address this issue, we fabricated a coassembly of two kinds of macrocyclic amphiphiles, guanidinium-modified calixarene (GCA) and cyclodextrin (CD). The GCA-CD coassembly exhibited extraordinarily strong binding affinity to Aβ with exquisite specificity benefiting from the heteromultivalent recognition. As a result, GCA-CD was capable of completely inhibiting the fibrillation of Aβ and disintegrating mature fibrils, and moreover, effectively mitigated the cytotoxicity induced by Aβ fibrils and improved the cognitive deficit of AD mice. The potential of GCA-CD as a therapeutic agent for the treatment of AD was witnessed both in vitro and in vivo. Download figure Download PowerPoint Introduction Alzheimer's disease (AD), as the most prevalent type of dementia, is the most concerning form of amyloidoses.1 According to statistics, 1 out of every 85 individuals worldwide is expected to live with AD by 2050.2,3 Someone develops dementia every 3 s, with the current annual cost of dementia being estimated at 1 trillion dollars, a figure set to double by 2030 ( www.alz.co.uk, World Alzheimer Report 2019). Although the pathogenesis of AD has not been clearly verified, the amyloid hypothesis is by now the most widely accepted.4–6 The amyloid hypothesis postulates that β-amyloid (Aβ) peptides of 39–43 amino acid residues, in a variety of forms, trigger a cascade that results in the harming of synapses and ultimately neurons, thereby producing the pathological presentations of Aβ plaques, tau tangles, synapse loss, and neurodegeneration, eventually leading to dementia.7 With respect to the amyloid hypothesis, inhibition of Aβ fibrillation, disintegration of Aβ fibrils, and promotion of the clearance of Aβ to maintain the balance of the metabolism and catabolism of Aβ appear to be quite significant for the prevention and treatment of AD. For a few years, anti-Aβ therapies have dominated clinical trials that await approval from the United States Food and Drug Administration ( www.clinicaltrials.gov). Several kinds of inhibitors have been developed as potential therapeutic agents. For example, natural small molecules, such as curcumin and congo red, have been reported to interact with the fibril-related conformation, and thus affect protein aggregation.8 Artificial receptors, such as molecular tweezer,9 cyclodextrin,10,11 calixarene,12 resorcinarene,13,14 and cucurbituril,15 have been designed to bind with one or several certain amino acids and inhibit the fibrillation of proteins. Certain peptides, either designed directly based on the Aβ sequence or designed without relying on the sequence, may also be utilized as inhibitors to interfere with the formation of fibrils or disrupt the aggregation process.16,17 With the rapid development of nanomedicine, various nanoparticles have also been engaged in recognition, fibrillation inhibition, fibril disintegration, and clearance of Aβ, thereby relying on the unique functionalized surface and size of nanoparticles.18–21 Despite these significant achievements, selective and strong binding toward Aβ, which is the prerequisite element for constructing an inhibitor, remains a huge challenge because of the complexity and size of peptide guests and their flexible conformation.22–26 Inhibitors with weak binding capabilities have been shown to only increase the lag phase of fibrillation, and then delay the assembly kinetically, but fail to inhibit and thermodynamically disrupt fibrillation.27,28 Poor selectivity would give rise to interference binding of biologically coexisting species, such as proteins. Moreover, an ideal inhibitor for anti-Aβ therapy should also incorporate the following features: (1) disintegrate mature fibrils in vitro and more importantly in vivo. Considering the complex physiological environments, in vivo disintegration would be much more challenging than that in vitro; (2) promote the clearance of Aβ in vivo. Accumulation of Aβ due to the imbalance of its production and clearance has been considered to be an indispensable process in the pathogenesis of AD19,29; (3) reduce the population of oligomers, which are considered as the major toxic forms.30–32 Oligomer toxicity is a major difficulty in anti-Aβ therapies. On the one account, it is better to inhibit Aβ fibrillation at an early stage before oligomerization. On the another account, antiamyloiddogenic agents are expected to not disaggregate amyloid plaques into more neurotoxic oligomers. Otherwise, treatment agents would demonstrate slight therapeutic effects or even aggravate the disease during the actual application.33 To overcome these challenges, we developed a new type of Aβ inhibitor by coassembling two kinds of macrocyclic amphiphiles, guanidinium-modified calixarene (GCA) and cyclodextrin (CD), for the treatment of AD. A key element of anti-Aβ therapy is the interaction between therapeutic agents and Aβ. The coassembled heteromultivalent recognition strategy has been shown to be suitable in recognition of complex systems, such as peptides, proteins, and membranes.34 The GCA-CD coassembly has been employed as a novel heteromultivalent recognition platform, where GCA tends to bind negatively charged amino acids, glutamic acid (E) and aspartic acid (D); CD favors encapsulation of aromatic amino acids, such as tyrosine (Y) and phenylalanine (F).35 These four amino acids are significant components of Aβ peptides. Guanidinium groups modified at the upper rim of calixarene are known to contribute to cellular uptake,36–41 and therefore, the complexation of Aβ with GCA-CD is expected to promote Aβ uptake by microglia for clearance. Moreover, both components of the coassembly, GCA and CD, have been previously demonstrated to be biocompatible.42–45 Consequently, we comprehensively investigated the potential of GCA-CD as a therapeutic agent in vitro and in vivo, including the recognition capability and selectivity of GCA-CD toward Aβ, inhibition of fibrillation and disintegration of fibrils, reduction of the population of oligomers, mitigation of the cytotoxicity of Aβ fibrils, promotion of microglia phagocytosis of Aβ, and improvement of cognitive deficits in an AD mice model (Scheme 1). Scheme 1 | Illustration of GCA-CD disintegrating Aβ fibrils, mitigating cytotoxicity, and improving cognitive deficits in AD mice. Download figure Download PowerPoint Experimental Methods Preparation of the GCA-CD coassembly and Nile red-loaded GCA-CD coassembly GCA and CD were dissolved in methanol and chloroform, respectively, at a concentration of 1.00 mM. The mixed organic solution of GCA and CD in a ratio of 1∶1 was dried under high vacuum for 4 h to yield a thin film in a glass vial. HEPES (2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid) buffer (10 mM, pH 7.4) was added, and the solution was sonicated at 80 ° C for 2 h to make the GCA-CD coassembly. Nile red (NiR) was dissolved in alcohol at a concentration of 5.00 mM. 4.00 μL of NiR alcohol solution was added into 1.00 mL GCA-CD stock solution (100/100 μM) to obtain the NiR-loaded GCA-CD coassembly. Fluorescence titrations The complexation of the assembly with amino acids, peptides, and proteins was measured by competitive fluorescence titrations. A mixed solution containing known amounts of reporter dye, assembly, and competitive guest was sequentially injected into 2.50 mL reporter dye and assembly solution in a quartz cuvette. Care was taken to keep the concentrations of dye and assembly constant in the course of titrations. The fluorescence intensity was measured before the first addition and after every additions on a Varian Cary Eclipse spectrometer equipped with a Varian Cary single-cell peltier accessory to control temperature. The association constant was obtained by fitting fluorescence intensity according to 1∶1 or n∶1 competitive binding model.46,47 Thioflavin T fluorescence assays The Aβ42 solutions in the absence and presence of different concentrations of GCA-CD were incubated by continuous orbital shaking at 200 rpm and 37 ° C. At different time points, 60.0 μL solution was mixed with 25.0 μM Thioflavin T (ThT) stock solution (10 mM HEPES buffer, pH 7.4) in a volume of 540 μL. The fluorescence was measured with excitation and emission at 440 and 480 nm, respectively. The obtained data were normalized as a percentage of the group without GCA-CD, whose maximum value was taken as 100%. Cell viability assays Cell viability was determined by MTT (3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2-H-tetrazolium bromide) assays. Briefly, PC-12 cells were plated in 96-well culture plates at a density of 3 × 104 cells per well and differentiated for 48 h. After treating with Aβ42 in the absence and presence of different concentrations of GCA-CD, MTT solution (20.0 μL, 0.500 mg/mL) was added to each well and cultured at 37 ° C for 4 h. The medium was removed carefully, and the formazan product in the cells was solubilized with 150 μL dimethyl sulfoxide (DMSO). Finally, formazan levels were measured by using a microplate reader (Multiskan Mk3; Thermo Labsystems, Helsinki, Finland) with a wavelength of 492 nm. Animals The 5xFAD transgenic mice overexpressing family AD mutation of human Aβ precursor protein (the Swedish mutation, K670N/M671L; the Florida mutation, I716V; and the London mutation, V717I) and PS1 (M146L/L286V) were obtained from the Jackson Laboratory (Bar Harbo, ME, USA). The animals were housed in the animal house of College of Life Sciences, Nankai University (Tianjin, China). All experiments were carried out according to the protocol approved by the Committee for Animal Care at Nankai University (20160004). Behavioral experiments For novel object recognition (NOR) test, during the habituation stage, the mice were placed in the open field, and let them explore the field for 5 min. The training stage begins 24 h after the habituation stage, and the mice were placed in the same box, which contained two same objects (objects 1 and 2) in two opposite corners. The mice were allowed to explore objects for 5 min. After 2 h, object 1 was replaced with a new one (object 3), and the mice were allowed to explore objects for 5 min (test 1 stage). After 24 h, object 3 was replaced with a new one (object 4), and the mice were allowed to explore objects for 5 min (test 2 stage). The time for mice to explore the object 2 was defined as T1, and the time to explore the object 1 (in training stage), or object 3 (in test 1 stage), or object 4 (in test 2 stage) was defined as T2. The recognition indices were calculated through the following formula.48 T 2 / ( T 1 + T 2 ) × 100 % The Morris water maze (MWM) test was performed in a circular pool (90 cm in diameter and 60 cm in height), filled with opaque water maintained at 20−22 ° C. The tank was divided into four equal quadrants by two imaginary vertical axes. During the initial training stage, a platform (10 cm in diameter) located in a target quadrant, submerged 1 cm below the water surface. The mice were placed into the tank facing the side wall and near the edge at one of four points randomly. The mice were allowed to find a hidden platform for 60 s, which if not found, guided to the platform. After each training, mice were allowed to stay on the platform for about 10 s. The training stage lasted for 5 days, recording the time needed to find the platform (escape latency) and swimming speed. The escape latency of mice that could not find the platform was recorded as 60 s. In the space exploring test stage (day 6), the platform was removed, and the mice were placed in the pool from the opposite side of the platform. The time spent in the target quadrant and the number of crossing platform were examined in the test stage. Results and Discussion Construction and characterization of the GCA-CD coassembly of heteromultivalent recognition Two coassembling components, GCA and CD (Figure 1a), were synthesized according to previous procedures.43,49 The GCA-CD coassembly was prepared by hydrating a mixture of GCA and CD in HEPES buffer (10 mM, pH 7.4) under sonication at 80 ° C for 2 h. The GCA-CD showed a hydrated diameter of 27.1 nm and a surface potential of 46 mV according to the results obtained from dynamic light scattering (DLS) and zeta potential (Figures 1c and 1e). Transmission electron microscopy (TEM) images showed consistent result with DLS (Figure 1d). In particular, DLS and static light scattering measurements were combined to provide a profile of the assembling morphology, as reflected by the Rg/RH ratio.50 The Rg/RH ratio was calculated and found to be 1.14 ( Supporting Information Figure S1), thereby indicating a vesiclular morphology. We compared the zeta potentials of GCA-CD with the mixture of GCA and CD (GCA + CD, 62 mV, Figure 1e), and the apparent difference between them indicated that the GCA-CD coassembly was precisely formed and it was not a physical mixture. We then employed solid-state NMR to further confirm that GCA and CD were coassembled. Compared with GCA + CD, the benzene 13C peak of GCA-CD underwent a downfield shift ( Supporting Information Figure S2), thereby indicating that GCA interacted with CD in the GCA-CD coassembly. Having confirmed the formation of the GCA-CD coassembly, we validated whether these two kinds of macrocycles were evenly distributed on the surface of the coassembly or self-sorted each other to generate microphase separation by using energy-dispersive spectrometry-elemental mapping analysis. Accordingly, we detected both nitrogen and sulfur, which are specific to GCA and CD, respectively. The signals of both elements were evenly distributed throughout the coassembly (Figure 1f), thereby indicating the molecularly homogeneous distribution of GCA and CD (Figure 1b). It is reasonable that microphase separation has a considerably less possibility because it is unfavorable in the entropy term. Figure 1 | (a) Chemical structures of GCA and CD. (b) Illustration of the formation of the GCA-CD coassembly. (c) DLS data and (d) TEM image of GCA-CD. (e) Zeta potentials of GCA-CD, GCA, CD, and GCA + CD. (f) Elemental mapping images of GCA-CD. DLS, TEM, and zeta potentials were examined in HEPES buffer (10 mM, pH 7.4). Elemental mapping was examined in water. Download figure Download PowerPoint Molecular recognition of GCA-CD toward Aβ42 The Aβ42 peptide (Scheme 2) was employed for the inhibition and disintegration studies, as it is known as the main constituent of amyloid plaques and as the most toxic species among Aβ peptides.51 Using competitive fluorescence titration with fluorescein as a reporter dye ( Supporting Information Figure S8), we measured the apparent association constant of GCA-CD with Aβ42 and it was found to be (3.29 ± 0.53) × 1013 M−1 (Figure 2a). After converting to a dissociation constant, the value was calculated to be up to (32.9 ± 3.3) fM. The binding affinity was demonstrated to be over five orders of magnitude stronger than that (7.9 × 107 M−1) reported in our previous coassembly of CD and carboxylated calixarene (CCA),34 thereby benefiting from the replacement of CCA with GCA. The association constants of the GCA assembly with E and D were (4.95 ± 0.08) × 104 M−1 and (1.09 ± 0.34) × 104 M−1 ( Supporting Information Figures S3, S4, and S5), respectively, similar to the association constant of the CCA assembly with lysine (K) [(0.882 ± 0.065) × 104 M−1; Supporting Information Figures S6 and S7]. However, Aβ42 is known to contain three E and three D, but only two K, which means that replacing CCA with GCA would allow the coassembly to interact with more sites of Aβ42. Moreover, Aβ42, as a negatively charged peptide, is known to have electrostatic attraction with GCA but electrostatic repulsion with CCA. These two factors were hypothesized to have given rise to the extraordinary improvement of the binding affinity, we the platform from to GCA-CD. The interaction between GCA-CD and Aβ42 was also validated using zeta potential the addition of Aβ42, the zeta potential of GCA-CD was shown to less ( Supporting Information Figure been by the of Aβ42. Scheme 2 | acid of Aβ42 and control peptides. Download figure Download PowerPoint The control peptide of Aβ42 with of D, and (Scheme binding with GCA-CD ( Supporting Information Figure thereby indicating that the recognition of GCA-CD to Aβ42 was due to these four amino The peptide (Scheme 2) with the binding sites of CD and removed exhibited a binding affinity of ± × M−1 with GCA-CD ( Supporting Information Figure thereby the of CD in the heteromultivalent recognition. The same was also for GCA. the binding sites of GCA and were removed, the binding constant ± × 104 of (Scheme 2) with GCA-CD was demonstrated to be much than that of Aβ42 ( Supporting Information Figure thus the of GCA. Figure 2 | (a) fluorescence titration of Aβ42 in the presence of fluorescein μM) with GCA-CD μM) fitting of the titration data to n∶1 competitive binding model was as (b) formation of Aβ42 (20.0 μM) by fluorescence in the absence and presence of different concentrations of is a specific widely to fibril formation with (c) of Aβ42 fibrils by GCA-CD. GCA-CD was added to an solution of μM Aβ42 at h after of (d) TEM images of μM Aβ42 incubated in the absence and presence of different concentrations of GCA-CD for h. (e) TEM images of Aβ42 (20.0 μM) fibrils by GCA-CD. (f) assays the after μM Aβ42 in the absence and presence of different concentrations of GCA-CD for h. assays the during the disintegration of Aβ42 (20.0 μM) by GCA-CD. All experiments were performed in HEPES (10 mM, pH 7.4) buffer at ° C. The group in the absence of GCA-CD as the Download figure Download PowerPoint binding recognition selectivity was another key to be considered for the of GCA-CD as an Aβ ideal inhibitor should interact with Aβ it should interact with other biologically proteins. during the complexation of GCA-CD should not of proteins. we investigated the complexation of GCA-CD with a of proteins, including protein and The obtained association constants of ± × × and ± × were shown to be much than that of Aβ42, with the binding to being weak to obtain data by competitive fluorescence titration ( Supporting Information Figures We also that the recognition of Aβ42 control proteins were over orders of We further examined whether the complexation of GCA-CD rise to the of proteins. In the of even of was about of the was ( Supporting Information Figure in the of of was about of the was ( Supporting Information Figure These results that GCA-CD not give rise to on in of the we the GCA-CD coassembly as a Aβ inhibitor due to the selective and strong binding toward Aβ42, as well as the of other proteins. of Aβ42 fibrillation and disintegration of fibrils by GCA-CD in buffer solution We incubated Aβ42 (20.0 μM) in the absence and presence of different concentrations and μM) of GCA-CD and the formation of by fluorescence and using As shown in Figure GCA-CD the fibrillation of Aβ42, and μM or more GCA-CD was we detected a slight increase in the that indicated the inhibition of the aggregation of TEM results were consistent with of the fluorescence TEM images showed that of Aβ42 for h fibrils, in the presence of μM or μM GCA-CD, fibrils were much and and in the presence of μM or μM GCA-CD, fibrils were (Figure the of the treatment of disintegration of Aβ42 fibrils is considered to be more than inhibition of the formation of fibrils, and is thus more After of Aβ42 for h, to the formation of mature fibrils, GCA-CD was We found that the fluorescence as a of time and after only 5 h, fluorescence to the initial thereby indicating a of Aβ42 fibrils (Figure We also validated the of the Aβ42 fibrils by GCA-CD using TEM (Figure The (in 5 and disintegration of mature fibrils was a result of the extraordinarily strong binding between GCA-CD and Aβ42. we investigated whether GCA-CD could also reduce the of Aβ42 oligomers, which are considered to be the most toxic After without or with different concentrations of GCA-CD for h, Aβ42 were detected by using the specific As shown in Figure the were shown to be in the presence of GCA-CD, and μM or μM GCA-CD was only a few were result indicated that GCA-CD effectively the formation of from Aβ42 were also during the process of fibril The were demonstrated to be as as the GCA-CD was added, and after 10 h, a few were (Figure thereby indicating that GCA-CD not disintegrate fibrils into oligomers, but could disintegrate oligomers. of Aβ42 fibrils We employed the PC-12 to the effects of GCA-CD in an in vitro model of AD. According to the results of the MTT the GCA-CD was shown to increase the viability of cells with Aβ42 fibrils (20.0 μM) Figure μM GCA-CD was viability from to The GCA-CD exhibited cytotoxicity at concentrations up to μM (Figure Figure 3 | (a) of PC-12 cells with Aβ42 fibrils μM) in the absence and presence of different concentrations of GCA-CD for 24 h. (b) of PC-12 cells with different concentrations of GCA-CD for 24 h. (c) Fluorescence images PC-12 were with (d) percentage of PC-12 (e) of and (f) of of PC-12 of PC-12 of of PC-12 images phagocytosis of by cells in the absence and presence of NiR-loaded coassembly were with For and measurements of the levels of as well as levels of the concentrations of Aβ42 and GCA-CD were μM and respectively. For the concentrations of and NiR-loaded coassembly were 2.50 μM and respectively. cultured without Aβ42 fibrils and GCA-CD as levels were and as compared with control and as compared with Aβ42 Download figure Download PowerPoint showed that which is an of We