Shedding Light on Alzheimer’s β-Amyloidosis: Photosensitized Methylene Blue Inhibits Self-Assembly of β-Amyloid Peptides and Disintegrates Their Aggregates

Abnormal aggregation of β-amyloid (Aβ) peptides is a major hallmark of Alzheimer’s disease (AD). In spite of numerous attempts to prevent the β-amyloidosis, no effective drugs for treating AD have been developed to date. Among many candidate chemicals, methylene blue (MB) has proved its therapeutic potential for AD in a number of in vitro and in vivo studies; but the result of recent clinical trials performed with MB and its derivative was negative. Here, with the aid of multiple photochemical analyses, we first report that photoexcited MB molecules can block Aβ42 aggregation in vitro. Furthermore, our in vivo study using Drosophila AD model demonstrates that photoexcited MB is highly effective in suppressing synaptic toxicity, resulting in a reduced damage to the neuromuscular junction (NMJ), an enhanced locomotion, and decreased vacuole in the brain. The hindrance effect is attributed to Aβ42 oxidation by singlet oxygen (1O2) generated from photoexcited MB. Finally, we show that photoexcited MB possess a capability to disaggregate the pre-existing Aβ42 aggregates and reduce Aβ-induced cytotoxicity. Our work suggests that light illumination can provide an opportunity to boost the efficacies of MB toward photodynamic therapy of AD in future.

ability and memory, which is characterized by abnormal accumulation of β-amyloid (Aβ) peptides of 39-43 amino acids 13 . Decades of studies have revealed that Aβ aggregation is a central pathological hallmark of AD, but the original function of Aβ and the mechanism by which Aβ self-assembly induces neurotoxicity have not been clearly elucidated 14 . Previous studies have shown that the aggregation of Aβ into β-sheet-rich oligomers or fibrils is a key pathogenic event in the onset of AD 15 . In this regard, the prevention of the self-assembly of Aβ monomers into aggregate states has been deemed vital for the treatment of AD. Over the years, researchers have made numerous efforts to screen small molecules that can inhibit Aβ aggregation 16 . Recently, photosensitizing chemicals have been explored for light-induced inhibition of Aβ assembly 17,18 . For example, photosensitized riboflavin and water-soluble porphyrin molecules significantly suppressed Aβ aggregation by oxidizing the peptides in the early stage of Aβ assembly 17,19 . MB is also known for its excellent photosensitizing property and has been extensively used for photodynamic treatment of cancer cells and microbes due to its high quantum yield of 1 O 2 generation (φ Δ ~ 0.5) under red light 20,21 . Based on the photochemical property of MB, here we explore light-induced inhibition of Aβ 42 aggregation by MB in vitro as well as the suppression of synaptic toxicity in Drosophila AD model under light illumination, as depicted in Fig. 1. Furthermore, we investigated the possibility of disintegrating pre-formed Aβ 42 aggregates by photo-excited MB molecules. One of the remarkable merits of MB as a photo-induced therapeutic agent for treating neurodegenerative diseases is its ability to cross BBB, which is regarded as a major difficulty for the development of brain-targeting drugs 22 . Furthermore, MB can be excited upon the absorption of red light (>630 nm), of which tissue penetration is better than that of green or blue light 23 . The higher tissue penetration depth of red light is a potential advantage of MB over previously reported, light-driven anti-amyloid aggregation agents of metal oxides and organic compounds, the absorption maxima of which lie at much lower wavelengths 17,19,24 .
The delivery of light into the brain tissue through the skull has been a major obstacle for the application of light in neuroscience and neuroengineering fields. Recent progresses in optogenetics, a technology to control a specific neural activity in the brain circuit using light 25 , facilitate the delivery of light to the target brain areas much feasible. To activate (or silence) a specific neural circuit, the researches illuminate the confined area using light guides such as fiber optics 26 . The optic fibers allows the light to be transferred to the deep brain areas, retaining its power density; they can easily be implanted in the head of freely moving animals. Moreover, recently development of wireless, implantable microLED platforms provide a minimal restriction in the behavior 27 . We envision that these recent advances in the implantable optoelectronic devices may lower the existing barrier in future applications of phototherapies to the neurodegenerative disorders.  monomers (40 μM) incubated with MB (10 μM) under dark conditions, CD profile showed a negligible change, while the peaks completely disappeared in the presence of MB under light illumination. This result indicates that photosensitized MB molecules strongly affect the conversion of Aβ 42 monomers into β-sheet rich aggregates. CD spectra of Aβ 42 recorded at different times show that the secondary structure of the peptides are changed from random coil structure to β-sheet structure. ( Figure S1) The diminished CD peaks monitored in Aβ 42 treated with photo-excited MB implies that the unstructured Aβ 42 monomers were remained after the incubation. Thioflavin T (ThT) fluorescence assay and atomic force microscope (AFM) analysis also support the photo-induced inhibitory effect of MB. MB-treated Aβ 42 under dark conditions showed an insignificant decrease in ThT fluorescence compared to the native Aβ 42 (Fig. 2b). While dense networks of mature fibrils were observed in the AFM images after incubation of Aβ 42 monomers for 24 hours (Fig. 2d,e), numerous short Aβ 42 fibrils were found when MB was treated (Fig. 2f). The short length of the fibrils is attributed to the accelerated rate of nucleation and fibril formation 28 . According to the previous study 29 , MB promotes the progress of Aβ 42 fibrillation by stabilizing pre-nuclear intermediates that favor Aβ 42 nucleation. In contrast, upon light illumination, substantially decreased ThT emission was monitored in the presence of MB, and only a limited number of aggregates were observed (Fig. 2g). The effect of photo-excited MB on the result of ThT assay was negligible according to the control experiment (Fig. S2). Native gel electrophoresis results revealed that the contents of Aβ 42 monomers (4.5 kDa) increased significantly with photo-excited MB, implying that the considerable amount of monomers did not assemble into the aggregates of high molecular weight (Fig. 2c). Note that effect of MB on the reduction of Ag + ion during the silver staining was negligible. The results obtained from sedimentation assay also demonstrate that the insoluble aggregates of Aβ 42 were significantly reduced when the monomers were incubated with MB under light. ( Figure S3) We further verified that the degree of photo-induced inhibition increased with the increasing MB concentration (Figs S4, S5). These photochemical analysis results clearly show that MB effectively suppressed the self-assembly of Aβ 42 monomers into neurotoxic, β-sheet-rich aggregates under light. Further studies to investigate the effect of photo-excited MB on the oligomerization of Aβ 42 or the equilibrium between various intermediates are needed. According to the literature, the distribution of Aβ 42 oligomers at the certain time point can be assessed using photoinduced cross-linking of unmodified proteins (PICUP), which provides "snapshots" of the size distribution of various intermediates existing during the assembly 30 . We further monitored the photo-induced Aβ 42 aggregation inhibition effect by changing light wavelength, power density, and illumination time. We investigated the effect of the light wavelength using red (λ max = 630 nm), green (λ max = 520 nm), and blue (λ max = 450 nm) LEDs. The maximum degree of inhibition was observed under red light, and the effect decreased with LEDs that had shorter light wavelengths (Fig. S6a). We attribute the result to the unique optical property of MB; as shown in Fig. S6b, the absorbance spectrum of MB overlaps mostly with the emission spectrum of red LED, but MB shows only a weak absorption in the shorter wavelength region (<550 nm). Moreover, we verified that the hindrance effect of photo-excited MB correlates with the power density of the light source (Fig. S7). In addition, when we shortened the illumination time from 24 h to 15 min, we could observe almost a similar degree of the inhibition effect on Aβ 42 aggregation (Fig. S8), which indicates that light illumination for a very brief period is sufficient to induce a full capacity of photo-excited MB against Aβ 42 aggregation. These results show that the efficacy of light-induced inhibition of MB can be easily controlled by the modulation of the light illumination system.

Photo-excited MB suppresses Alzheimer's Defects in Drosophila.
We further investigated the in vivo efficacy of photosensitized inhibition of Aβ 42 aggregation by MB using Drosophila AD model. Animal models of human diseases are vital for understanding pathogenesis and for developing potential therapeutic agents. Drosophila melanogaster is one of most popular animal models due to its well-studied anatomical features that enable quantitative analysis of various phenotypes 31 . The AD model of Drosophila achieved by the overexpression of Aβ 42 shows several neurodegeneration phenotypes, such as morphological defects of neuromuscular junction (NMJ), locomotion defects, and brain vacuolization 32,33 . Here, we tested the suppression of the Aβ 42 -induced phenotypes by feeding MB to the Drosophila AD model under red LED light. The postsynaptic overexpression of Aβ 42 (Mhc > Aβ42) leads to a significant loss in the synaptic bouton number 34 . We found that the number of bouton was reduced by ~30% in Mhc > Aβ42 larvae compared to the control (Mhc-GAL4/+) according to the confocal images of muscle 6/7 of abdominal segment 3 (Fig. 3a,e). A negligible improvement was monitored with the treatment of MB under dark or light illumination without MB. (Fig. 3f,g) The defect, however, was significantly rescued in the NMJ of larvae treated with 100 nM MB as well as light. (Fig. 3h,i) Reduced NMJ synaptic connection in the larvae is known to cause defective movement of the muscles related to the crawling behavior of the larvae and brain vacuolization 32,35 . Figure 3j,k show that the larvae with postsynaptic overexpression of Aβ 42 exhibited the reduction in crawling distance by more than 30% compared to the Mhc-GAL4/+ control. In contrast, when MB was fed to Mhc > Aβ42 larvae and was illuminated by red light, a significant improvement in the crawling behavior was observed in a dose-dependent manner (Fig. 3l), which coincides with the NMJ analysis results. When cultured under dark conditions, however, we could monitor only slight increase in the locomotion even when a highly concentrated MB was treated. We attribute the mild recovery of the phenotype to the therapeutic activity of MB itself without the aid of light, as reported previously 7,9 . Yet, the photo-excited MB showed notably higher efficacy in a low dosage than static MB. We also found that photo-excited MB can reduce Aβ 42 toxicity-induced brain vacuolization in the fly's brain. In the brain of 30-days-old flies, overexpression of Aβ 42 (elav > Aβ42) showed severe brain vacuolization in the cell body region compared to the elav-GAL4 control (Figs 4a, and S9). In contrast, the number of vacuoles of photo-excited-MB-treated elav > Aβ42 was reduced compared to that of non-MB-treated elav > Aβ42 fly brains. The lost area of the photo-excited-MB-treated elav > Aβ42 fly brains were reduced by ~20% compared to non-treated flies (Fig. 4b). This significant restoration of Aβ 42 -induced toxicity in the elav > Aβ42 fly brains coincided with the results of the locomotion defect experiment. Taken together, these results suggest that photo-excited MB can suppress defects of NMJ morphology, locomotion defects, and Aβ 42 -induced toxicity in the Drosophila AD model. Photo-excited MB dissociate the pre-exisiting aggregates. While numerous studies have focused on the inhibition of the Aβ 42 assembly pathway, recent studies have attempted to reverse the progress by dissociating pre-formed Aβ 42 aggregates [36][37][38] . Previous studies have demonstrated that the clearance of pre-existing amyloid deposits could reverse AD pathology, including behavioral deficits, in transgenic mouse models 36,39 .
To examine the possibility of disassembling Aβ 42 aggregates by photo-excited MB, we incubated pre-formed aggregates with MB under dark or light conditions and monitored the changes in ThT fluorescence, morphology, and cytotoxicity. For the experiment, Aβ 42 monomers were incubated for 48 h to produce fully-grown, fibrillar aggregates. According to the ThT assay result (Fig. 5a), ThT fluorescence was drastically diminished (~50%) when photo-excited MB was applied, while MB under dark conditions caused a negligible decrease. The corresponding AFM images also confirmed that the density of the fibril networks decreased in the presence of photo-excited MB (Fig. 5f). Both the results of the ThT assay and the AFM images clearly indicate that light triggers disassembly of existing Aβ 42 aggregates when incubated with MB. Additional researches such as size-exclusion chromatography (SEC) and in vivo studies using the brain of mouse AD models are required to further investigate the efficacy of photo-excited MB against pre-formed aggregates.

Discussion
We attribute the light-induced hindrance effect of MB to its high binding affinity to Aβ 42 and oxidative stress generated from photochemical reactions. To investigate the interaction between MB and Aβ 42 , we monitored the change of MB fluorescence in the presence of Aβ 42 . We observed enhanced fluorescence of MB with a blue shift with an increasing concentration of Aβ 42 peptides (Fig. 6a). According to the literature 40 , the fluorescence enhancement can be attributed to the reduction in the non-radiative decay of photo-excited MB due to the suppressed rotation and vibration upon binding to other chemicals. The blue shift of the fluorescence of fluorophore occurs when a dye exists in a more non-polar environment because the energy difference between the excited and ground state increases 41,42 . This implies that MB may bind to the hydrophobic C terminus of Aβ 42 monomers 43 . Further studies, including computational simulations, are required to predict the exact binding site of MB to    Aβ 42 . The binding constant (K d ) of MB to Aβ 42 , estimated from the changes in the fluorescence intensity at 670 nm for various MB/Aβ 42 ratios, was 48.7 ± 3.6 μM (Fig. 6b). This is comparable to the K d of curcumin, a well-known small molecular inhibitor of Aβ 42 aggregation (K d = 46 μM) 44 .
The capacity of MB as a light-driven 1 O 2 generator has been widely utilized in a number of studies 45,46 . Under the irradiation of red light (λ max = 666 nm), MB monomers produce 1 O 2 through the type II photochemical pathway, in which the energy from triplet state MB (i.e., 3 MB + ) is transferred to molecular oxygen 20 , and the generated 1 O 2 oxidizes organic compounds nearby 47 . To explore the possible photo-oxidation of Aβ 42 by MB, we conducted 2,4-dinitrophenylhydrazine (DNPH) assay, which is one of the most commonly used methods to assess the amount of carbonyl groups formed by oxidative stress [48][49][50] . As shown in Fig. 6c, we observed a new absorption band at 380 nm only when Aβ 42 was incubated with MB under light illumination, indicating that Aβ 42 peptides were oxidized by photo-excited MB. In addition, our ThT assay and CD analysis revealed that the hindrance effect of photo-excited MB on Aβ 42 aggregation decreases significantly under anaerobic conditions (Figs S10, S11). These results indicate that the generation of 1 O 2 is a major cause of light-induced inhibition of Aβ 42 assembly by MB. We suppose that the generated oxidative stress induces sulfoxidation of Aβ 42 's methionine, which is known to be a most readily oxidizable residue 51 . According to the literature 52 , the oxidation of Met35 causes structural alteration in the hydrophobic C-terminus of Aβ 42 and impedes the association and self-assembly between monomers. The oxidation of Aβ 42 by photo-excited MB was further studied with matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS). Figure S12 shows that the mass of the MB-treated Aβ 42 increases when the light was irradiated. We attribute a +14 Da-modification to the oxidation of His13 or His 14 residues, which generates a dehydro-2-imidazolone derivative 53 . According to the literature, the further increases in the mass of Aβ 42 by 16 Da are resulted from the oxidation of Met35 and Tyr10 19 . We monitored how static and photo-excited MB affect the aggregation kinetics of Aβ 42 differentially using ThT fluorescence assay. For the experiment, we pre-incubated MB-treated Aβ 42 solution for 30 min at 4 °C under dark or light conditions before the measurement was performed at 30 °C. According to our results (Fig. 6d), while static MB slightly affected the nucleation of Aβ 42 peptides with a decreased lag time from 202.0 min to 164.7 min, photo-excited MB completely blocked the progression of Aβ 42 aggregation in the early stage. We attribute the hindered aggregation to the oxidation of Aβ 42 monomers by localized 1 O 2 generated during light illumination. The negligible inhibitory activity of pre-illuminated MB supports that the hindrance effect of photo-excited MB was derived from the photodynamic reaction of MB under light. (Figure S13) For the therapeutic applications, it is vital to minimize the undesirable oxidative damages to the surrounding by limiting the 1 O 2 generation sites. Future studies to enhance the Aβ-specific targeting of photosensitizers are essential for the phototherapy of amyloidosis.
In summary, we demonstrated that photo-excited MB molecules exhibit a high degree of inhibition against β-amyloidosis in vitro and in vivo. Our kinetic study revealed that, while static MB accelerates Aβ 42 aggregation, MB under illumination thoroughly blocks the progress in the early stage by oxidizing the peptide. We examined the in vivo effect of light-induced inhibition of aggregation by MB using the Drosophila AD model. At the same dose, while static MB exhibited mild recovery in the locomotion defect, photoexcited MB almost fully rescued the AD phenotype in in vivo experiments performed with the Drosophila AD model; the loss in synaptic bouton, locomotion defect, and vacuolization in the brain were significantly reduced with the MB treatment under red LED light, indicating that photoexcited MB successfully prevented in vivo toxicity resulting from β-amyloidosis. We further verified that MB under illumination is also able to dissociate pre-existing aggregates and to suppress resulting cytotoxicity, while MB under dark conditions did not affect the aggregation state. Based on these results, we suggest that shining light on MB can be a breakthrough to enhance its efficacy beyond the conventional limit. While the recent report on clinical trials performed with MB was not satisfactory, this study hints at a new opportunity of inhibiting β-amyloidosis based on the photosensitizing property of MB, a therapeutic chemical that has been used for more than a century.
Monomeric Aβ 42 solution was prepared by dissolving the peptide in hexafluoro-2-propanol (HFIP) followed by sonication for 30 min and keeping it overnight at room temperature. The solution was aliquoted into 1.5 ml protein LoBind tubes (Eppendorf) and vacuum-dried for 2~3 h. The tubes were then stored at −20 °C for further use. Atomic Force Microscopy (AFM). For the AFM measurement, 5 μl of Aβ 42 sample solutions were deposited onto a cleaved mica substrate for 10 min and were rinsed several times with DI water to remove remaining salts and unbound peptides. After they were fully dried, AFM images were acquired in a tapping mode with an NCHR silicon cantilever (Nanosensors Inc.) using a Multimode AFM instrument equipped with a Nanoscope III controller and "E"-type scanner (Digital Instruments Inc.).

Light-induced inhibition of Aβ
Native gel electrophoresis and silver staining. The Aβ 42 solutions were transferred to a loading buffer containing 50 mM Tris HCl, pH 6.8, 1% SDS, 1% β-mercaptoethanol, 10% (v/v) glycerol, and 0.01% bromophenol blue. The samples were loaded onto 10% Gradi-Gel ™ II gradient gel (Elpis Biotech) and peptide distribution was visualized by silver staining. Protein electrophoresis kit were purchased from Bio-rad.
Sedimentation assay. The sedimentation assay was performed according to the previous study 54 . Briefly, Aβ 42 monomers (40 μM) were centrifuged at 10,000 g for 10 min at 4 °C. The supernatant was collected and the optical density (OD) at 214 nm was measured using V/650 spectrophotometer (Jasco Inc.). The supernatant was then moved to the glass vials and MB (2 μM) was introduced to the vials. The samples were incubated under dark or light conditions for 24 h at 30 °C. Then the samples were ultracentrifuged at 100,000 g for 10 min at 4 °C. The OD 214 of collected supernatant was measured. The aggregation was derived from the difference between the OD 214 before and after the incubation as described by Yoshiike et al. Fly strains. The UAS-Aβ42 was provided by Dr. K Iijima-Ando 55 , the Mhc-GAL4 driver was provided by Dr. T Littleton; and the elav-GAL4 driver was provided by Bloomington Stock Center. For the pharmacological approach, either MB or PBS was added to fly food at 1, 10, 100, 1000, and 10000 nM concentration. All flies were reared at 25 °C.
Brain vacuole analysis. For analysis of brain vacuolization, 30-days-aged fly heads were fixed in 4% paraformaldehyde (Electron Microscopy Sciences) and were processed for paraffin sections as described 32 . Embedded paraffin was cut into 4 um-thick coronal sections. These sections were stained with hematoxylin and eosin (Vector laboratories). For quantification of vacuole phenotypes in the fly head section, we measured the area of the vacuoles in the cell body region using ImageJ. Five to ten hemispheres were analyzed for each genotype. For the pharmacological approach of the brain vacuole analysis, either MB or PBS was added to fly food at 100 nM concentration.
Immunohistochemistry. Third instar larvae were dissected in PBS, fixed in 4% formaldehyde (Ted Pella) in PBS for about 15 minutes and washed 3x in 0.1% Triton X-100 in PBS. FITC-conjugated anti-HRP (Jackson ImmunoResearch Laboratories) were used at 1:100 and were incubated for about 1.5 hours at room temperature. Laval samples were mounted in SlowFade Antifade kit (Invitrogen). Confocal images were collected from OLYMPUS FLUOVIEW FV-1000 confocal microscopes SP2 equipped with 40x UPlans FL N inverted oil lens. OLYMPUS Fluoview software was used to capture, process, and analyze the images. Analysis of the NMJ was performed essentially as described 34 . Crawling assay. Wandering 3rd instar larvae were briefly washed with PBS to remove residual fly food. Larvae were dried for a short time on clean filter paper and were placed on a 2% agar-grape juice coated petri dish. Each genotype was allowed to crawl freely for 90 sec. To quantify the crawling distance, we drew lines to track crawled larvae and measured the distance using Image J software. Approximately 10-20 animals were tested for each genotype.