Introduction

Alzheimer’s disease (AD) is a neurodegenerative disease and progressive disorder of the brain that gradually decreases thinking, memory, and language abilities1. The amyloid cascade hypothesis suggests that abnormal aggregation and accumulation of amyloid β (Aβ) causes neurodegeneration in the aged brain2. Aβ consists of 39–43 amino acid residues and is derived from amyloid precursor protein that is cleaved by β- and γ-secretase3. These short peptides start to aggregate and exhibit neurocytotoxicity4.

To appreciate the pathological feature of AD, it is important to trace the dynamic behavior of Aβ aggregation to neurons because the aggregation of Aβ is a key step in the development of AD. Several studies on the polymerization and fibrillation of Aβ have been performed using transmission electron microscopy5,6,7,8. Although high resolution image analysis of Aβ42 aggregation is possible, it is necessary to dry the sample for observation. Therefore, observation of Aβ aggregation in a physiological state cannot be performed. Recently, Querol-Vilaseca et al. succeeded in the three-dimensional (3D) observation of Aβ present in amyloid plaques of AD patients using stimulated emission depletion microscopy, one type of super resolution microscope9. Furthermore, five categories of processes in the aggregation of cellular Aβ were reported using fluorescence-lifetime imaging and 3D structural illumination microscopy10. Thus, although the in situ analysis of Aβ aggregation is increasing annually, it is difficult to perform the spatiotemporal high resolution analysis of Aβ dynamics under physiological conditions. In particular, the molecular mechanism of interaction between Aβ aggregates and the cell surface has not been clearly elucidated.

It is well known that the interaction between membrane lipids and Aβ is responsible for the modulation of Aβ fibrillation and the expression of neurocytotoxicity11,12,13. The aggregation of Aβ on the cell membrane is an important step in the formation of diffuse plaques14. Aβ preferentially accumulated in gangliosides and cholesterol domains of the cell membrane of PC12 cells, and these aggregates exhibited cytotoxicity15. Recently, using scanning electron transmission microscopy and an electron tomogram, Han et al. reported a 3D interaction between fibrils and the membrane of cells and that fibrils disrupted the membrane and caused leakage of lysosomes16. Thus, it has been shown that the cell membrane and its components can function as scaffolds for Aβ aggregation. Moreover, cell membrane morphology and the size of lipid vesicles are also important factors that modulate Aβ aggregation. Sugiura et al. reported that Aβ aggregation was enhanced on lipid vesicles displaying high curvature17. Small liposomes promoted the formation of Aβ fibrils, whereas huge liposomes inhibited the growth of Aβ fibrils, suggesting that Aβ fibrillation is dependent on liposome size18. Diverse morphological changes in the cell membrane are caused by cell motile processes such as cell division and migration, which are physiologically important phenomena. One cytoskeleton protein, actin, plays a particularly important and indispensable role in regulating the morphology of eukaryotic cells. Actin is especially necessary for the formation of filopodia and lamellipodia in motile cells19. Actin dynamics, including modulation of its assembly and disassembly, is essential for synaptic plasticity involved with memory skills20. Particularly, dynamics of the spine are regulated by the actin cytoskeleton21. Small GTPase proteins have important roles in the morphological regulation of neuron cells20. Rac1 activity is required for the maintenance of spine shape22,23. On the other hand, activation of RhoA exhibits inhibitory function in formation of spines. Overexpression of constitutively active RhoA in hippocampal neurons of mice caused the spine to retract and decrease24. Thus, several studies have indicated the importance of the function of actin cytoskeleton in homeostasis of neural tissue. Indeed, F-actin disassembly of the dendritic spine was caused in neuron cells of AD model mice25. Aβ affected the architecture of actin and the tubulin network in AD pathology26. However, it is obscure whether changes in cell morphology derived from cell motility influence the polymerization and fibrillization of Aβ that exists in culture medium or on the cell membrane. Furthermore, it is unknown whether morphological changes of the cellular membrane, which occur after the formation of actin-dependent protrusions, affect Aβ aggregation in the environment close to the cell surface. This implies that the direct interaction of cell membrane dynamics through actin assembly and Aβ aggregation is also unclear.

Previously, we reported a real-time imaging method of Aβ42 aggregation using quantum dot (QD) nanoprobes and developed a microliter-scale high-throughput screening (MSHTS) system for Aβ42 aggregation inhibitors that applied this imaging method27,28,29,30. QD, one of the fluorescent nanoparticle, is a cluster of semiconductor materials composed of hundreds to thousands of selenium and cadmium, and emits fluorescence of various wavelengths according to the particle diameter. Its size is about 10–20 nm, which is almost the same size as the protein. It is known that the advantages of QD are that excitation with a single wavelength light source is possible, and that it is very bright and hardly discolors. Since the extremely high photostability of such QD enables long-term fluorescence observation, it is most suitable for continuous observation of Aβ42 aggregation in vitro, which progresses on the order of several hours to several days. In MSHTS system, we utilized QD-conjugated Aβ40, which had a binding ratio (Aβ40/QD) of 6. We synthesized QDAβ by chemically cross-linking QD with NH2 group attached to the surface and Aβ40 with N-terminal cysteine, an amino acid having a thiol group in the side chain. By mixing QDAβ with unlabeled Aβ42 at a ratio of 1: 1000, we succeeded in visualizing Aβ42 aggregates formed by incubation at 37 °C. These QDAβ were incorporated into Aβ42 fibrils with a similar efficiency as unlabeled Aβ42. One advantage of this method is that it allows observation of the dynamics of Aβ42 aggregation under physiological conditions, which has been difficult until now.

In this study, we succeeded in visualizing the process of Aβ42 aggregation with living cells by applying the QD imaging method in real time. In other words, 2D fluorescence microscopy confirmed Aβ42 aggregation imaging method by QDAβ are utilized in culture medium. Further, 3D real-time observations using confocal laser microscopy revealed that Aβ42 preferentially started to aggregate on the dynamic cell surface where membrane protrusions frequently formed.

Results

Observed Aβ42 aggregation around PC12 cells

We initially examined whether Aβ42 aggregation was observed in culture medium. Rat adrenal pheochromocytoma PC12 cells, which were differentiated by 4.5 ng/ml nerve growth factor (NGF), were then incubated with 20 μM Aβ42 and 30 nM QDAβ for 24 h. After incubation, we observed Aβ42 aggregates using a conventional fluorescence microscope (Supplementary Fig. S1) and confocal laser microscope (Fig. 1A). These observations showed that Aβ42 aggregates were particularly concentrated around cells. We found cells in which the entire cell periphery was covered by Aβ42 aggregates and other cells in which only a part of the cell was covered (Fig. 1B,C). We confirmed that NG108–15 cells from a mouse neuroblastoma × rat glioma hybrid also exhibited Aβ42 aggregation at the peripheral region (Supplementary Fig. S2). This indicates that the aggregation of Aβ42 is a ubiquitous phenomenon in living cells. Then, we evaluated the neurocytotoxic effects of Aβ42 and QD nanoprobes using the 3-(4, 5-dimethylthial-2-yl)-2, 5-diphenyltetrazalium bromide (MTT) assay, and confirmed that the addition of Aβ42 reduced cell viability (Supplementary Fig. S3). The combinations of QD or QDAβ and Aβ42 also exhibited almost the same result as Aβ42 alone. These results indicate that the addition of QDAβ had little effect on the viability of Aβ42-treated cells. 3D real-time imaging revealed that early aggregation at the cell periphery occurred after 8 h of incubation (Fig. 2A,B and Supplementary Movies S1 and S2). After 16 h of incubation, there was a dramatic increase in the accumulation of Aβ42 aggregates on the cell side. As shown in Fig. 2C, we suggested the model of Aβ42 aggregation at the cell periphery. The amount of aggregates increased over time. Interestingly, these 3D observations showed that Aβ42 aggregation was not localized on the top of the cell. The amount of Aβ42 aggregates surrounding the cell periphery increased over time (Fig. 2D). To clarify the mechanism by which Aβ42 binds to cells, cell membranes were stained with a fluorogenic membrane dye, CellBrite green (Fig. 2E). Aβ42 and the cell membrane colocalized, indicating that Aβ42 might aggregate on the cell surface by specifically interacting with the cell membrane.

Figure 1
figure 1

Localization and dynamics of Aβ42 aggregation with PC12 cells. (A) PC12 cells were co-incubated with 20 μM Aβ42 and 30 nM QDAβ, and observed by a conventional inverted microscope (DIA) and a confocal microscope (QDAβ). Bar = 50 μm. (B) Magnified images of boxes a and b in panel A. Bar = 20 μm. Cells were covered by Aβ42 aggregation on all (box a) or half (box b) of the cell periphery. (C) Schematic model of the Aβ42 aggregation with PC12 cells.

Full size image
Figure 2
figure 2

3D dynamics of Aβ42 aggregation with PC12 cells. (A) PC12 cells were co-incubated with 20 μM Aβ42 and 30 nM QDAβ, and then imaged by confocal microscopy with three-dimensional real-time imaging. White arrowhead indicates early aggregation at 8 h. Approximate cell morphology is shown as dotted lines at 16 h in the side view image. Yellow arrowheads indicate the accumulation of Aβ42 aggregates on the cell side. Note that Aβ42 aggregation did not localize on top of the cell. Bar = 20 μm. (B) 3D reconstruction of 48 h time point in Panel (A). (C) Schematic model of the Aβ42 aggregation around PC12 cells. Top and side view were illustrated. (D) Histogram of Aβ42 aggregated perimeter (%) after 0, 16, and 24 h incubation. The percentage of Aβ42-aggregated perimeter was calculated as the ratio of Aβ42-aggregated perimeter to the entire perimeter. n = 470, 268, 235 cells in 0, 16, 24 h incubation, respectively. Error bars represent ±SDs of the mean values of six areas in the well. Note that PC12 cells display time-dependent increase of Aβ42 aggregation. (E) 3D reconstruction of PC12 co-incubated with 30 μM Aβ42 and 30 nM QDAβ. Cells stained with CellBrite green (green). Images were captured by a confocal microscope. Note that Aβ42 aggregates (red) colocalized on the cell membrane.

Full size image

Interaction between Aβ42 aggregation and protrusive cell membrane

To reveal the process of Aβ42 aggregation at the cell’s periphery, we performed multi-channel live cell imaging using QDAβ. 2D time lapse observation clarified the formation of Aβ42 aggregates with movement of PC12 cell protrusion (Fig. 3A and Supplementary Movie S3). After 24 h of incubation, Aβ42 aggregation was observed in the region where membrane protrusions were frequently formed. A kymograph at the edge of the cell showed that Aβ42 aggregation first occurred around the protrusions (Fig. 3B, white arrow). Here, we suggested the model of hypotheses that active cell protrusion enhanced Aβ42 aggregation (Fig. 3C). In the 3D time lapse observation, the concentration of Aβ42 was reduced to 16 μM to facilitate analysis of Aβ42 dynamics on the cell membrane and PC12 cells were transiently transfected to a plasmid with enhanced green fluorescent protein (EGFP) to distinguish cell morphology (Fig. 4). Still images of a maximum fluorescence intensity projection movie showed a time-dependent increase of Aβ42 aggregates around the cell (Fig. 4A and Supplementary Movie S4) and the incorporation of Aβ42 and QDAβ to intracellular regions by phagocytosis27. Furthermore, still images of a 3D reconstruction movie revealed that Aβ42 preferentially started to aggregate at the region where membrane protrusions frequently formed (Fig. 4B and Supplementary Movie S5). In Fig. 4C, we suggested 3D model of Aβ42 aggregation around the PC12 cell.

Figure 3
figure 3

2D real-time imaging of cell protrusion and Aβ42 aggregation. (A) PC12 cells were co-incubated with 20 μM Aβ42 and 30 nM QDAβ, and observed by a conventional inverted microscope (DIA) and confocal microscope (QDAβ). Time series of images show the gradual steps of Aβ42 aggregation at the cell periphery. Bar = 20 μm. (B) Kymograph of cell periphery in enlarged image of panel A (white boxes) show that Aβ42 starts to aggregate after formation of the protrusion. The cell is localized on the top side of the panel. White arrow and yellow arrowheads indicate membrane protrusion and Aβ42 aggregates, respectively. (C) Schematic model of the Aβ42 aggregation around protrusions of PC12 cells.

Full size image
Figure 4
figure 4

3D real-time imaging of cell protrusion and Aβ42 aggregation. (A) Image of maximum fluorescence intensity projection of 0 and 24 h time point in 3D real-time imaging. PC12 cells were co-incubated with 16 μM Aβ42 and 30 nM QDAβ. Aβ42 aggregates at the cell periphery at the 24 h time point. Bar = 10 μm. (B) 3D reconstruction images of Panel (A). Active protrusions formed from the cell surface at 11 h 45 min (upper left). Aβ42 started to aggregate in the part indicated by white arrowheads. Images were captured by a confocal microscope. (C) Schematic 3D model of the Aβ42 aggregation around protrusions of PC12 cells.

Full size image

To elucidate whether Aβ42 aggregations were formed around the cell protrusion specifically, we performed wound healing assay (WHA) with Aβ42 and observed Aβ42 aggregation processes. WHA is simple and inexpensive method to evaluate the cell motility such as migration rate and directionality31. In this method, a confluent cell monolayer was scratched by sharpen tip. Then, it was observed closure process of artificial gap by cell motility. Wu et al. reported that PC12 cell showed active protrusions and recovery wounded gap during migration in this assay32. As shown in Fig. 5A, we confirmed PC12 cells continued to migrate toward newly created space for 24 h observation. In 24 h time point, PC12 cell exhibited active protrusions (Fig. 5B: white arrows). Incubation with Aβ42 and QDAβ during cell migration, we succeeded in observation of preferential formation of Aβ42 aggregates at the wounded edge (Fig. 5C and Supplementary Movie S6). To clarify where aggregations were occurred, we analyzed fluorescence intensity profiles along the boxed region in Fig. 5C (+Aβ42) and Supplementary Fig. S4 (+DMSO). Fluorescence intensity profiles indicated that Aβ42 treated PC12 cells showed characteristic peak around cell edge, suggesting that Aβ42 aggregates were formed at the wounded edge (Fig. 5D and Supplementary Fig. S4B). The aggregates showed about 100 μm width across the wounded edge. In other words, Aβ42 actively aggregates in the anterior group where cell motility is active (the width of a few cells). We suggested the schematic illustration of preferential aggregation at the wounded edge where frequently protrusions were formed (Fig. 5E). These results demonstrate that active membrane protrusions are essential for the promotion of Aβ42 aggregation at the cell periphery.

Figure 5
figure 5

Aβ42 aggregation induced at edge of migrating PC12 cells. (A) The monolayer of PC12 cells was scratched by toothpick and defined as 0 h. Images of live cell migrated to wounded gap were captured every 10 min for 24 h. Bar = 100 μm. (B) Boxed region of panel A. Image of 24 h time point showed that migrating PC12 cells elongated protrusions (white arrows). Bar = 20 μm. (C) The monolayer of PC12 cells was scratched by toothpick. Then, PC12 cells were co-incubated with 25 μM Aβ42 and 30 nM QDAβ, and observed by a conventional inverted microscope. Time series of images show the gradual steps of cell migration and Aβ42 aggregation at wounded edge (white arrowhead). Images of live cell migrated to wounded gap were captured every 10 min for 24 h. Images of QD channel were false-colored in red using image J software. Bar = 100 μm. (D) Profile plot of boxed region of Panel (C) (QD channel at 24 h) indicates mean gray value per pixel. Plotted value displays a column average plot in boxed region (red line). The line was represented with error bar means ± SD (pink line). Black arrowhead and black dot line indicate the edge of cell group. (E) Schematic model of the Aβ42 aggregation at edge of migrating PC12 cells.

Full size image

Observed Aβ42 aggregation-induced apoptosis of PC12 cells

The aggregation and accumulation of Aβ42 on the surface of a cell membrane have various adverse effects on cellular physiological homeostasis. Actually, human neuroblastoma, SH-SY5Y cells exhibiting the accumulation of Aβ42 aggregates undergo apoptosis33. As shown in Fig. 6A,B and Supplementary Movie S7, we demonstrated that Aβ42 aggregation on the cell surface abruptly caused cell death. Additionally, we found that aggregates of Aβ42 cause abnormal cell morphology and the flow of cytoplasmic components out of the cell due to destruction of the cell membrane (Supplementary Fig. S5). Furthermore, we found that Aβ42 aggregation led to defects in maintenance of nuclear morphology (Supplementary Fig. S6). These phenomena are considered to be important processes in the mechanism of cell death caused by Aβ42 aggregation. Interestingly, after cell death, Aβ42 aggregates, which accumulated on the cell surface, remained there. As shown in Fig. 6 (time point at 11 h 59 min), even after the membrane blebbed and the cytoplasm flowed out, the Aβ42 aggregates maintained a rough cell morphology, like a “husk”. We suggested the schematic illustration of cell death caused by Aβ42 aggregations (Fig. 6C). Furthermore, additional Aβ42 aggregation occurred and continued to grow, so the basis of this “husk” consists of Aβ42 aggregates (Supplementary Movie S7: time point at 14 h 14 min).

Figure 6
figure 6

Cell death induced by Aβ42 aggregation. (A,B) Time series images of maximum fluorescence intensity projection (A) and images of 3D reconstruction (B) of real-time imaging. PC12 cells were co-incubated with 16 μM Aβ42 and 30 nM QDAβ. Note that the Aβ42-aggregated cell (white arrowheads) exhibited abnormal blebbing and cell death. Images were captured by a confocal microscope. After abnormal blebbing formation, cell death was suddenly induced by Aβ42 aggregation. Bars = 10 μm. (C) Schematic model of the cell death caused by Aβ42 aggregation around PC12 cells.

Full size image

F-actin polymerization enhanced Aβ42 aggregation on the cell surface via protrusion regulation

Here, we hypothesized that cytoskeletal actin, which is essential for the control of cell morphology and cell motility, is involved in the mechanism of Aβ42 aggregation around cells. To investigate the role of actin in protrusion-dependent Aβ42 aggregation, we visualized F-actin of PC12 cells using Alexa 647-phalloidin. Aβ42 and F-actin colocalized at the cell periphery (Fig. 7A). Many Aβ42 aggregates were localized in the region where F-actin had accumulated. To examine whether actin polymerization avidity in protrusions is necessary for Aβ42 aggregation, we inhibited actin polymerization of cells using 20 μg/ml cytochalasin D, which is an actin polymerization inhibitor that caps the barbed end of F-actin34. Before the experiment, we analyzed the effect of cytochalasin D on Aβ42 aggregation using MSHTS methods28,29. In MSHTS method, we quantified the amount of Aβ42 aggregates from standard deviation (SD) values of fluorescence intensity of each pixel. Addition of cytochalasin D (0.16-100 μg/ml) did not change the SD value compared to control sample (Supplementary Fig. S7), indicating that Aβ42 aggregation was not directly inhibited by 100 μg/ml cytochalasin D. Then, we confirmed that 20 μg/ml of cytochalasin D treatment suppressed Aβ42 aggregation around cells (Fig. 7B). These results indicate that the formation of F-actin polymerization-dependent protrusions is required for the preferential aggregation of Aβ42 at the cell periphery.

Figure 7
figure 7

F-actin polymerization-dependent cell membrane movement is required for Aβ42 aggregation on the cell surface. (A) PC12 cells were co-incubated with 20 μM Aβ42 and 30 nM QDAβ (red) for 24 h and were stained with Alexa 647-phalloidin (magenta), and then observed by a confocal microscope. Bar = 50 μm. Bottom panels indicates enlarged images of boxed region in the top panels. Bar = 20 μm. The white arrow points to colocalization of Aβ42 aggregates and F-actin accumulation in the cell periphery region. (B) PC12 cells were treated with DMSO (control) or 20 μg/ml cytochalasin D for 24 h and then co-incubated with 20 μM Aβ42 and 30 nM QDAβ (red) for 24 h. Bar = 50 μm. Images were captured using a confocal microscope (QDAβ) and a conventional inverted microscope (DIA). Right three panels indicate enlarged images of the boxed region in the left panel. Cytochalasin D inhibited Aβ42 aggregation at the cell periphery. Bar = 20 μm.

Full size image

Discussion

In this study, by adding Aβ42 and QDAβ to a culture model neuron, we succeeded in real-time imaging of the aggregation process of Aβ42 around cells. It was revealed that the aggregation of Aβ42 is not uniform on the cell surface and that preferential aggregation is promoted at the cell periphery where cell protrusions are intensely formed. We also demonstrated that actin polymerization-dependent cell motility is responsible for the promotion of Aβ42 aggregation at the peripheral region of PC12 cells.

Lu et al. reported the gradual aggregation of Aβ that exists inside culture cells and defined the assembly of Aβ fibrils as an aggresome that grows until several micrometers in diameter10. Here, we also visualized extracellular aggregation. We confirmed that Aβ42 aggregates formed a unit several micrometers in diameter and that these Aβ42 units interacted with each other, then aggregated in solution (M. Kuragano et al., manuscript in preparation). It is possible that Aβ42 aggregates might exhibit an upper size limit of several micrometers. By using a super-resolution microscope, new insight about classification of the process of aggregation at the cell periphery might be obtained.

Recently, it was reported that the aggregation of Aβ is promoted under the presence of surfactants, such as micelles, near the cell membrane16,35. Wakabayashi et al. reported that gangliosides present in sphingolipids and cholesterol-rich domains called “lipid rafts” on cell membranes function as scaffolds for Aβ aggregation15. Here, we hypothesized that the diffusion of Aβ42 transits from a three-dimensional area in culture medium to a two-dimensional plane on the surface of the cell membrane (Fig. 8A) and that the intense cell membrane protrusions driven by actin polymerization dramatically increase the frequency of collisions among Aβ42 molecules at the cell surface (Fig. 8B), which promotes additional Aβ42 aggregation. Observation of the maturation of early Aβ42 aggregates in vitro also supports the possibility that the aggregation nucleus initially formed on the cell surface, incorporated new Aβ42, and exhibited the growth of aggregates (Supplementary Fig. S8 and Supplementary Movie S8).

Figure 8
figure 8

Schematic model of novel Aβ42 aggregation process on the cell surface. (A) Transition of the Aβ42 movement state from three-dimensional to two-dimensional diffusion. Due to the frequency of collisions between each Aβ42 monomer on the cell membrane surface, aggregation is thought to increase there. (B) Relationship between the formation of cell protrusion and promotion of Aβ42 aggregation. The frequency of collisions between each Aβ42 monomer on the cell membrane might increase in the area where active protrusions formed, more than in the static region. Therefore, it is thought that Aβ42 aggregate formation is particularly promoted at the cell edge where movement is active.

Full size image

We also showed that Aβ42 aggregates at the cell surface can cause cell death. Neuronal death is considered to be triggered by promoting the formation of Aβ42 aggregates on the cell membrane. In fact, apoptosis of SH-SY5Y cells, was induced by Aβ fibrillation33. After abnormal morphological changes of the cell and nucleus, the cell membrane was disrupted, then neuronal cells died. Moreover, we found that the reduced cell membrane plasticity caused by Aβ42 accumulation may be involved in the expression of its neurocytotoxicity36,37. Remarkably, Aβ42 aggregates on the surface of PC12 cells remained there, even after cell death. After frequent membrane blebbing and spillage of the cellular components due to cell death, the aggregated Aβ42 continued to stay in place as if it had left the cell outline. This “husk” also seems to play the function of an aggregation nucleus, i.e., new aggregates accumulated on this “husk”. The amyloid cascade hypothesis suggested that the emergence of an amyloid plaque (senile plaque) is major feature of the expression of AD2. Although the real-time imaging of plaque formation in local neuronal tissue of transgenic mice was observed using multiphoton microscopy38, details of the processes and molecular mechanism underlying its formation are still obscure39. In addition to the newly formed plaques in the microenvironment near the neuronal cells38, we hypothesized that Aβ42 aggregation remaining around dead cells might mature into amyloid plaques. Actually, it was reported that the size-order of amyloid plaques is about the same as neuronal cells and that amyloid plaques consist of proteins, carbohydrates, nucleic acids, lipids, and metal ions40,41. Among these major components, lipid rafts on the surface of PC12 cells are known scaffolds of Aβ42 aggregation, as mentioned above15. It is possible that the debris of dead cells arising from Aβ42 neurocytotoxicity accelerate the progression of AD.

F-actin is necessary to regulate neuronal morphology20 and maintain spine plasticity42. SynGAP, a GTPase-activating protein, is the main regulator of actin dynamics of the dendritic spine and is involved with cofilin, which caused disassembly of F-actin by its serving activity43. Drebrin is an F-actin-binding protein and is highly expressed in brain tissue44. In normal spines, drebrin protects F-actin from serving by cofilin activity. In other words, the drebrin-actin complex exhibits an important function by maintaining spine morphology. Separately, cofilin caused the disappearance of F-actin polymerization in the AD spine45. However, we demonstrated that Aβ42-treated PC12 cells accumulated F-actin in the cortical region (Fig. 7A). It was reported that cofilin mediated Aβ-induced apoptosis46. Our results imply that the cofilin-independent apoptosis pathway was stimulated by the aggregation of Aβ42 on the cell surface. In fact, it was reported that F-actin acts as a sensor of apoptosis47 and that the apoptosis of endothelial cells was caused by stimulation of TNF-α through changes in the actin cytoskeleton and an increase in phosphorylation of the regulatory light chain of non-muscle myosin II48. It is possible that Aβ42 aggregation directly suppresses flexibility of the cell membrane and decreases cell plasticity. In addition, the F-actin cytoskeletal fractions, which are supposed to be used for cell membrane protrusion, are thought to have nowhere to localize, allowing them to start to accumulate under the membrane. Consequently, the apoptosis pathway might be induced. It is essential to consider in the future whether the excessive accumulation of actin, which is induced by Aβ42, may affect myosin II activity. A well-known structure in the brain, the dendritic spine, protrudes from the dendrite of a nerve cell49. Changes in the number and shape of these structures, which play a role in receiving information from most excitatory synapses in the brain, are thought to be involved in the mechanism of neuroplasticity49,50. Dendritic spines are generally classified into two sizes, small and large spines, which exhibit different properties50,51. Particularly, immature small spines, with filopodia-like structures, exhibit more active movement and a shorter lifetime than large spines49. As shown in Figs. 3–5, dynamic membrane movement promoted Aβ42 aggregation around cells. It is possible that Aβ42 aggregation occurred preferentially on the surface of small spines than on large spines in the brain. In other words, it is reasonable to expect that small spines are susceptible to neurocytotoxicity due to Aβ42 aggregation. Although large spines have more F-actin than small spines in proportion to their surface area, its motility is lower than that of small spines. Currently, we are working on elucidating the molecular mechanism of Aβ42 aggregation-dependent cell death through F-actin accumulation.

Materials and Methods

Cell culture

Rat adrenal pheochromocytoma, PC12 cells, were obtained from the JCRB Cell Bank (Japan). Mouse neuroblastoma × rat glioma, NG108-15 cells, were a kind gift by Prof. Hiroyuki Nakagawa (Fukuoka University). PC12 cells were maintained in RPMI (Gibco/Life Technologies, USA) supplemented with 5% fetal bovine serum (FBS) (Gibco/Life Technologies) and 10% horse serum (HS) (Gibco/Life Technologies) or DMEM (Wako, Japan) supplemented with 10% FBS. NG108-15 cells were maintained in DMEM supplemented with 10% FBS. All culture media were supplemented with 100 U/mL penicillin and 100 μg/mL streptomycin (Wako). Cells were cultured at 37 °C in humidified air containing 5% CO2.

Reagents

Poly-D-Lysine was purchased from Sigma-Aldrich (USA). Human Aβ42 (4349-v; Peptide Institute, Japan) and Cys-conjugated Aβ40 (23519; Anaspec, USA) were purchased commercially. Nerve growth factor was purchased from Cosmo Bio (Japan). Alexa 647-phalloidin and Hoechst 33342 were purchased from Invitrogen (USA). CellBrite were purchased from Biotium (USA).

Preparation of QDAβ nanoprobes

QDAβ nanoprobe was prepared using QD-PEG-NH2 (Qdot 655 ITK Amino (PEG) Quantum dot; Q21521MP or Qdot 605 ITK Amino (PEG) Quantum dot; Q21501MP, Thermo Fisher Scientific, USA) according to our previous reports28,29,30. The QDAβ nanoprobe was prepared as follows: 10 µM QD-PEG-NH2 was first reacted with 1 mM sulfo-EMCS (22307; Pierce/Thermo Fisher Scientific) in PBS for 1 h at room temperature. After quenching and eliminating unreacted sulfo-EMCS, the QD-PEG-NH2-bound sulfo-EMCS was reacted with 74 μM Cys-conjugated Aβ40 for 1 h at room temperature. The concentration of QDAβ was determined by comparing absorbance at 350 nm to unlabeled QD-PEG-NH2.

MSHTS system

The SD values in existence of cytochalasin D were determined by a modified MSHTS system, as was described in our previous reports28,29,30. More specifically, various concentrations of cytochalasin D, 25 μM Aβ42 and 25 nM QDAβ in PBS were incubated in a 1536-well plate (782096, Greiner Bio-One, Austria) at 37 °C for 24 h. The QDAβ-Aβ42 aggregates that formed in each well were observed by an inverted fluorescence microscope (TE2000, Nikon) equipped with a color CCD camera (DP72, Olympus) and an objective lens (Plan Fluor 4×/0.13 PhL DL, Nikon). SD values of fluorescence intensities of 40,000 pixels (200 × 200 pixels) around the central region of each well were measured by Image J software (NIH, USA).

MTT assay

PC12 cells were plated at 2.0 × 104 cells in poly-D-lysine-coated 96-well plates (IWAKI, Japan) and incubated for 24 h. After incubation, cells were treated with 4.5 ng/ml NGF and further incubated for 24 h. Then, cells were treated with QD, QDAβ, and Aβ42 and cells were incubated for a further 24 h. Then, cells were treated with 1.2 mM MTT and incubated for 4 h. After incubation, the supernatant of each well was removed and then crystals of formazan were dissolved in 10% SDS/0.01 M HCl solution. After overnight incubation, absorbance intensity (570 nm) was measured in a 96-well plate. Cell viability was assessed as a percentage relative to DMSO-treated cells.

Transfection

pEGFP was a kind gift from Assoc. Prof. Masayuki Takahashi (Hokkaido University). Cells were transfected with plasmid DNA using Superfect Transfection Reagent (Qiagen, Germany) in DMEM according to the manufacturer’s protocol. Transfected cells were replated onto a glass-bottom 96-well plate (Corning, USA) precoated with 0.1 mg/mL poly-D-lysine for live cell imaging.

time-lapse observation

For single cell motility observation, time-lapse images were captured with an inverted microscope (Ti-E; Nikon, Japan) equipped with a color CMOS camera (DS-Ri2; Nikon) and an objective lens (PlanApo λ 20×/0.75 NA; Nikon). During observation, the cells were maintained in DMEM/F12 (1:1) (Gibco/Life Technologies) supplemented with 10% FBS and warmed in a chamber set at 37 °C. Images were captured every 10 min and analyzed using NIS-Elements C software (Nikon). For WHA, PC12 cells were plated at 10 × 104 cells in a glass-bottom 96-well plate precoated with 0.1 mg/mL poly-D-lysine. Cells were incubated for 24 h. After incubation, cells were treated with 45 ng/ml NGF and further incubated for 24 h. Then, after removal of culture medium, cell monolayer was scratched by toothpick and defined as 0 h. Then, the wells were filled by DMEM/F12 (1:1) supplemented with 10% FBS and 45 ng/ml NGF. This new medium included 1% DMSO or 25 μM Aβ42 and 30 nM QDAβ. Time lapse observations were performed in the same way as above described. Fluorescence images were captured as 8-bit grayscale images. All images were adjusted by a macro program so that average intensity became 50% according to our previous report method29. The mean gray value per pixel of Aβ42 aggregates was measured by the plot profile, the function of Image J software.

Confocal laser scanning microscopy

Z-stacks images were captured using an inverted microscope (Ti-E) and a confocal laser microscope system (C2 Plus; Nikon) equipped with an objective lens (PlanApo λ 20×/0.75 NA; Nikon, Plan Apo λ 100×/1.45 NA Oil; Nikon). Cells were maintained in RPMI supplemented with 5% FBS and 10% HS and warmed in a chamber set at 37 °C (INUBTF-WSKM-B13I; Tokai Hit) during observation. Images were captured and analyzed using NIS-Elements C software. For 3D time-lapse imaging, cells were maintained in DMEM/F12 (1:1) supplemented with 10% FBS and 4.5 ng/ml NGF, and warmed in a chamber set at 37 °C during observation. EGFP and QD were excited by a laser at 488 and 561 nm, respectively. Z-stack images were captured every 15 min and were oversampled by taking 1000 nm z-steps between acquired images and were analyzed using NIS-Elements C software. For observation of Aβ42 aggregation by the MSHTS system, 25 μM Aβ42 in and 25 nM QDAβ, PBS containing 5% EtOH and 3% DMSO were incubated in a 1536-well plate (782096; Greiner) warmed in a chamber set at 37 °C. Z-stack images were captured every 3 min and were oversampled by taking 2500 nm z-steps between acquired images and were analyzed using NIS-Elements C software.