Intracellular trafficking of particles inside endosomal vesicles is regulated by particle size

Journal of Controlled Release 260 (2017) 183–193

Contents lists available at ScienceDirect

Journal of Controlled Release journal homepage: www.elsevier.com/locate/jconrel

Intracellular trafficking of particles inside endosomal vesicles is regulated by particle size

MARK

Michihiko Aoyamaa, Yasuo Yoshiokaa,b,c,⁎, Yoshiyuki Araid, Haruna Hiraia, Rio Ishimotoa, Kazuya Naganoa, Kazuma Higashisakaa, Takeharu Nagaid,e, Yasuo Tsutsumia,f,⁎⁎ a

Laboratory of Toxicology and Safety Science, Graduate School of Pharmaceutical Sciences, Osaka University, 1-6 Yamadaoka, Suita, Osaka 565-0871, Japan Vaccine Creation Project, BIKEN Innovative Vaccine Research Alliance Laboratories, Research Institute for Microbial Diseases, Osaka University, 3-1 Yamadaoka, Suita, Osaka 565-0871, Japan c BIKEN Center for Innovative Vaccine Research and Development, The Research Foundation for Microbial Diseases of Osaka University, 3-1 Yamadaoka, Suita, Osaka 565-0871, Japan d Department of Biomolecular Science and Engineering, The Institute of Scientific and Industrial Research, Osaka University, 8-1 Mihogaoka, Ibaraki, Osaka 567-0047, Japan e Laboratory of Biomolecular Science and Engineering, Graduate School of Pharmaceutical Sciences, Osaka University, 1-6 Yamadaoka, Suita, Osaka 565-0871, Japan f The Center for Advanced Medical Engineering and Informatics, Osaka University, 1-6 Yamadaoka, Suita, Osaka 565-0871, Japan b

A R T I C L E I N F O

A B S T R A C T

Keywords: Endosomal vesicle Intracellular traffic Motility Particle size Particle tracking Silica nanoparticle

Little comparative information is available on the detailed intracellular dynamics (diffusion, active movement, and distribution mechanisms) of nanoparticles (≤100 nm) and sub-micron particles (> 100 nm). Here, we quantitatively examined the intracellular movements of different-sized particles and of the endosomal vesicles containing those particles. We showed that silica nanoparticles of various sizes (30 to 100 nm) had greater motility than sub-micron particles in A549 cells. Although particles of different sizes localized in the early endosomes, late endosomes, and lysosomes in different proportions, their motilities did not vary, regardless of the vesicles in which they were localized. However, surprisingly, endosomal vesicles containing silica nanoparticles moved faster than those containing sub-micron particles. These results suggest that nanoparticles included within endosomal vesicles do not suppress the motility of the vesicles, whereas sub-micron particles perturb endosomal vesicle transport. Our data support a new hypothesis that differences in particle size influence membrane trafficking of endosomal vesicles.

1. Introduction With the recent development of nanotechnology, nanoparticles have been used in a variety of fields, such as the food, cosmetics, and industries [1–3]. Recent studies in rodents have revealed that nanoparticles show greater tissue-penetration ability and internalization ability than conventional materials in various tissues (liver, spleen, and lung) [4,5]. Therefore, nanoparticles such as mesoporous silica nanoparticles, gold nanoparticles, and fullerene are expected to be particularly useful as novel drug-delivery carriers and contrast agents in the medical field [6–8]. In addition, the cellular uptake and intracellular localization of nanoparticles change depending on the properties of these particles (size, charge, chemical composition, and surface modification) [9–11]. For example, some reports show that after

nanoparticles enter the cell they are localized not only near the cell membrane but also at the perinuclear side of the cytoplasm and in organelles such as the nucleus; by carrying drugs to the perinuclear side of the cell they can thus improve therapeutic effects [12,13]. These data suggest that nanoparticles have unique and therapeutically attractive behaviors in vivo and in vitro that is barely accomplished by conventional materials; their specific in vivo kinetics and intracellular dynamics give them potential as novel non-viral drug-delivery carriers. However, although assessment of the detailed intracellular dynamics of these particles—including distribution speed, spatiotemporal localization, intracellular processing, and excretion—is important if we are to understand the mechanisms of nanoparticle-specific intracellular dynamics, these mechanisms are still poorly understood. Particle tracking by using real-time imaging is one of the most

⁎ Correspondence to: Y. Yoshioka, Vaccine Creation Project, BIKEN Innovative Vaccine Research Alliance Laboratories, Research Institute for Microbial Diseases, Osaka University, 3-1 Yamadaoka, Suita, Osaka 565-0871, Japan. ⁎⁎ Correspondence to: Y. Tsutsumi, Laboratory of Toxicology and Safety Science, Graduate School of Pharmaceutical Sciences, Osaka University, 1-6, Yamadaoka, Suita, Osaka 5650871, Japan. E-mail addresses: [email protected] (Y. Yoshioka), [email protected] (Y. Tsutsumi).

http://dx.doi.org/10.1016/j.jconrel.2017.06.007 Received 23 September 2016; Received in revised form 10 May 2017; Accepted 11 June 2017 Available online 13 June 2017 0168-3659/ © 2017 Elsevier B.V. All rights reserved.

Journal of Controlled Release 260 (2017) 183–193

M. Aoyama et al.

purchased from Invitrogen (Carlsbad, CA, USA). Plasmids encoding the green fluorescence protein (GFP)-fusion endosome/lysosome marker proteins GFP-EEA1 wt (#42307), GFP-rab7 WT (#12605), and LAMP1mGFP (GFP with an N-terminal palmitoylation to bind to the cell membrane; #34831) were purchased from Addgene (Cambridge, MA, USA). Nocodazole was purchased from Wako Pure Chemical Industries.

powerful methods of studying in detail the intracellular dynamics of particles such as protein aggregates, DNA aggregates, organelles, and viruses [14,15]. Particle tracking is expected to reveal those intracellular movements that are difficult to unravel by using conventional methods such as localization analysis in fixed cells [16]. However, only limited numbers of studies have reported the trajectories of nanoparticles and the mean square displacement (MSD) of their intracellular movements [17,18]. For this reason, the relationship between the properties and detailed intracellular dynamics of particles, including their motility (diffusion, velocity, and their mechanisms) has not yet been determined. Even the differences in intracellular motility between nanoparticles and sub-micron particles remain unclear. Here, we investigated the intracellular trajectories and motilities of silica particles with diameters of 30, 50, 70, 100, 300, and 1000 nm inside the cell by using real-time imaging. We showed that silica nanoparticles (with diameters ≤ 100 nm) moved faster than sub-micron particles (with diameters > 100 nm). The silica particles were included within endosomal vesicles, which were then transported along the microtubules. Although silica nanoparticles and sub-micron particles were localized in early endosomes (EE), late endosomes (LE), and lysosomes (Ly) in different proportions, the differences in particle motility did not result from these differences in preferential localization. Surprisingly, regardless of the type of endosomal vesicle, those containing silica nanoparticles moved faster than those containing submicron particles. These results suggest that silica nanoparticles do not suppress the motility of endosomal vesicles, whereas sub-micron particles have suppressive effects on endosomal vesicle transport. The difference in endosomal motility between silica nanoparticles and submicron particles likely results from a difference in perturbation effects on endosomal vesicle transport. Our results reveal part of the mechanism of nanoparticle-specific intracellular dynamics and support a new hypothesis that differences in particle size influence the membrane trafficking of endosomal vesicles. These findings should help unravel the mechanisms of nanoparticle-specific dynamics, along with part of the cellular physiology of membrane trafficking of endosomal vesicles.

2.3. Microscopy Fluorescence images (single-color live-cell images and immunofluorescence images) were obtained by using a custom-built objectivetype total internal reflection fluorescence (TIRF) microscope, based on an inverted microscope (Ti-E; Nikon Co., Tokyo, Japan), with a 100×/ 1.49 numerical aperture Apo TIRF oil-immersion objective lens (Nikon Co.), a 4× or 2.5 × intermediate variable magnification lens (VM lens C-4 ×, VM lens C-2.5 ×; Nikon Co.), and a scientific complementary metal oxide semiconductor camera (C11440-22CU, Hamamatsu Photonics, Shizuoka, Japan). Fluorescent probes were excited by a 488nm or 561-nm laser (Spectra-Physics, Tokyo, Japan). Live-cell images were observed through a 525-nm or 609-nm emission filter (FF01-525/ 45-25 and FF01-609/54-25, Semrock, Lake Forest, IL, USA). Dual-color fluorescence live-cell imaging was performed by using the same optical system as for single-color live-cell imaging and dual-view optics (A8509; Hamamatsu Photonics). Each emission was split into two parts at a wavelength of 550 nm by dichroic mirrors (DM550LP and DM550SP; Hamamatsu Photonics). 2.4. Live-cell imaging of movements of silica particles

Fluorescence-labeled silica particles were purchased from Micromod Partikeltechnologie (Rostock/Warnemünde, Germany). Silica particles that had diameters of 70, 300, and 1000 nm (nSP70, SP300, and SP1000; catalog numbers 40-00-701, 40-00-302, and 40-00103, respectively) and were labeled with orange fluorescence (Rhodamine B: excitation and emission wavelengths, 569 and 585 nm, respectively) were used. The silica particles were sonicated for 5 min and vortexed for 1 min before use.

A549 cells were seeded on a 35-mm glass-base dish (Iwaki, Shizuoka, Japan) at a density of 2.5 × 105 cells/2.5 mL/dish. The dishes were sonicated with 1/10 N potassium hydroxide solution (Wako Pure Chemical Industries), ultrapure water, 99.5% ethanol (Wako Pure Chemical Industries), and ultrapure water in turn for 15 min each and then sterilized by ultraviolet irradiation for 2 h before being seeded with the cells. Cells were incubated for 24 h at 37 °C in 5% CO2 before exposure to a silica particle suspension. Silica particle dispersions were prepared by diluting silica particle stock to a concentration of 25 μg/mL in culture medium just before addition to the cells. The cells were pulsed for 3 h at 37 °C with silica-particle-containing medium at 2.5 mL/dish. The particle-containing medium was then removed, and the cells were washed three times with phosphate-buffered saline (PBS). Fresh, particle-free medium was then added to the cells, which were chased for an additional 3 h. The medium was replaced with observation medium [10% FBS and 1% Ab in phenol-red-free DMEM (Wako Pure Chemical Industries)] just before observation. Cells were observed by using inclined illumination fluorescence microscopy. Images were captured at 100-ms intervals for 30 s.

2.2. Cell line, antibodies, plasmids, and reagents

2.5. Data analysis

A549 cells (human lung carcinoma) were purchased from the American Type Culture Collection (Manassas, VA, USA). A549 cells were maintained at 37 °C in 5% CO2 with culture medium [Dulbecco's Modified Eagle's Medium (DMEM; Wako Pure Chemical Industries, Osaka, Japan) supplemented with 10% fetal bovine serum (FBS; Biosera, Kansas City, MO, USA) and 1% antibiotic-antimycotic-mix stock solution (Ab; Gibco, Carlsbad, CA, USA)]. Mouse monoclonal antiearly endosome antigen 1 (EEA1) antibody (clone name: 14/EEA1) was purchased from BD Transduction Laboratories (BD Biosciences, San Jose, CA, USA). Rabbit polyclonal anti-rab7 antibody (clone name: H50) and mouse monoclonal anti-lysosome-associated membrane protein 1 (LAMP1) antibody (clone name: H4A3) were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). Alexa488-conjugated goat anti-mouse immunoglobulin G (IgG) secondary antibody and Alexa488-conjugated goat anti-rabbit IgG secondary antibody were

Live-cell images were processed by using ImageJ software (National Institutes of Health, Bethesda, MD, USA). The backgrounds of the images were subtracted by using ImageJ, and particle movements in the images were tracked by using the custom-written ImageJ plugin “PTA” (https://github.com/arayoshipta/projectPTAj) developed by Yoshiyuki Arai. The plugin determined the x–y position of each particle by twodimensional Gaussian fitting with the Levenberg-Marquardt method and performed particle tracking by using a nearest-neighbor algorithm. The x–y position information (i.e., the trajectory of movement) and the MSD against time [ρ(Δt)], which is a convenient quantitative measure of stochastic movement, were given by the plugin. We extracted those trajectories that we were able to track for 3 s (30 frames). To analyze the movement of the silica particles, we calculated the α-coefficients of the MSD curves until Δt = 2.5 s (25 frames) by fitting to a power law, log [ρ(Δt)] = α log [(Δt)] + ζ, where α is the α-coefficient and ζ is a

2. Materials and methods 2.1. Nanoparticles and sub-micron particles

184

Journal of Controlled Release 260 (2017) 183–193

M. Aoyama et al.

Fig. 1. Fluorescence images of silica particles. a to c, Typical fluorescence images of silica particles in A549 cells. A549 cells were treated with 25 μg/mL Rhodamine-B-labeled silica particles (nSP70, SP300, or SP1000). After a 3-h pulse with the particles and then a 3-h chase without the silica particles, bright-field and fluorescence images of cells (a), images of cells treated with nSP70 under 4 °C conditions (b), and images of cells treated with supernatant of Rhodamine-B-labeled nSP70 at 37 °C (c) were acquired by using inclined illumination fluorescence microscopy. White arrows in bright-field image of a indicate SP1000. d, e, TEM observations of nSP70-, SP300-, or SP1000-treated A549 cells. e shows magnifications of the areas in the white boxes in d. Arrows indicate silica particles. Single, or a few, silica particles were localized at a single point. Scale bars represent 10 μm in a to c, 5 μm in d, and 1 μm in e.

noise term. MSD data with coefficients of determination (R2) > 0.9 and 0 < α < 2 were classified into two patterns according to the α value and fitted by two equations. MSD data with 1.0 < α ≤ 2.0 were fitted by ρ2(Δt) = 2DΔt + ν2Δt2 + ζ, and MSD data with 0 < α ≤ 1.0 were fitted by ρ1(Δt) = 4DΔt + ζ, where D is the diffusion coefficient, ν is the (mean) velocity, and ζ is a noise term. Fitting and numerical calculations were analyzed with the program Origin 9 (LightStone, Tokyo, Japan). We also used x–y position information to calculate the spread of the trajectories in 3 s from the start point of tracking. We identified the x–y position farthest from the starting point in 3 s and calculated the distance of this position from the starting point.

2.7. Inhibition of microtubule-dependent active movements The microtubule-depolymerization reagent nocodazole was used to inhibit microtubule-dependent active movement. A549 cells pulsed with silica particles were prepared as above, and after a 3-h pulse the cells were chased for another 3 h with culture medium containing 1 μM nocodazole. The cells were then observed by using inclined illumination fluorescence microscopy, as above. All cells except controls were additionally treated with 1 μM nocodazole during live-cell imaging. 2.8. Immunostaining with endosomal markers A549 cells were seeded onto eight-well coverglass chambers (Iwaki) at 2 × 104 cells/200 μL/well. The coverglass chambers were prepared in the same way as for the live-cell imaging. Twenty-four hours after cell seeding, the A549 cells were pulsed for 3 h with silica particles at 25 μg/mL in culture medium and then chased for an additional 3 h at 37 °C. After the 3-h pulse and 3-h chase, the cells were fixed in 4% paraformaldehyde in neutral phosphate buffer (Wako Pure Chemical Industries) for 30 min at room temperature, and then washed three times with PBS. After fixation, the cells were permeabilized with 0.1% saponin from quillaja bark (Sigma-Aldrich, St Louis, MO, USA) in PBS for 5 min. After permeabilization, the cells were incubated for 30 min at room temperature with blocking buffer (2% bovine serum albumin in PBS). Then the blocking buffer was removed, and the cells were incubated with the appropriate dilution of primary antibody (anti-EEA1, anti-Rab7, and anti-LAMP1 diluted 1:100, 1:75, and 1:500, respectively, in blocking buffer) for 1 h at room temperature. The primary antibodies were then removed and the cells were washed three times (5 min each) with PBS. The cells were then incubated with 1:400 diluted Alexa488-labeled secondary antibody (in blocking buffer) for 1 h at room temperature. Finally, the secondary antibodies were removed and the cells were washed three times (5 min each) with PBS. The fixed cells were observed by epi-illumination with a 200-ms exposure by using the same custom-built microscopy setup as above, with a 2.5 × intermediate variable magnification lens. The microscopy images were segmented by using the ImageJ plugin “Squassh” [19], and the

2.6. Transmission electron microscopy (TEM) analysis of silica particles in A549 cells A549 cells were seeded onto a one-well chamber dish (Iwaki) at 2 × 105 cells/2 mL/dish. After a 24-h incubation at 37 °C, the cells were pulsed with 25 μg/mL silica particles in culture medium and then chased with fresh medium for an additional 3 h. After the 3-h pulse and 3-h chase, the cells were fixed in a 2% glutaraldehyde – 2% paraformaldehyde mixture for 30 min. The cells were washed with 0.1 mol/L phosphate buffer (pH 7.4) and post-fixed in 2% osmium tetraoxide (Merck Ltd., Tokyo, Japan). The fixed cells were dehydrated and embedded in EPON 812 resin (TAAB Laboratories Equipment, Berkshire, England). Ultrathin sections were stained with uranyl acetate (Electron Microscopy Science, Hatfield, PA, USA) and lead citrate EM (TAAB Laboratories Equipment) and observed by TEM (H-7650; Hitachi HighTechnologies Corporation, Tokyo, Japan). TEM analysis of the silica particles in the A549 cells was performed by the National Institutes of Biomedical Innovation, Health and Nutrition (Ibaraki, Osaka, Japan). The numbers of localized silica particles were analyzed from the TEM images. The number of particles localized in one vesicle (with a single membrane-like structure) or in one part of the cytosol (0.2 × 0.2 μm2; namely the area under the theoretical point spread function) was counted as the number localized in a single place.

185

Journal of Controlled Release 260 (2017) 183–193

M. Aoyama et al.

Fig. 2. Intracellular movement of differentsized silica particles. Intracellular movements of silica particles were observed by live imaging at 100-ms intervals. The trajectories of silica particles were tracked, and the mean square disintracellular intraextraintraplacement (MSD), diffusion coefficient and cellular cellular cellular extracellular velocity of the particles were calculated from those trajectories trackable for > 3 s. a, b, Typical trajectories of silica particles in the c d e 10-2 10-2 cell were plotted (a). The trajectories of (%) nSP70 nSP70 5 120 20 nSP70 differed from those of SP300 and SP300 SP300 SP1000 (b; magnification of part of each super diffusion SP1000 SP1000 100 4 panel of a). Scale bar represents 10 μm in a (1.0<α<2.0) 15 80 and 2 μm in b. c, Percentages of trajectories sub-diffusion 3 (0.4≤α≤1.0) of silica particles that were categorized as 60 10 “confined”, “sub-diffusion”, and “super-dif2 confined 40 (α<0.4) fusion”, as defined in the text. Error bars 5 1 20 represent the standard error of the mean (SEM) from three individual experiments. d, 0 0 0 e, MSDs of the super-diffusion, sub-diffusion, 0 0.5 1 1.5 2 2.5 (sec) 0 0.5 1 1.5 2 2.5 (sec) nSP70 SP300 SP1000 and confined subgroups. Means of MSD versus time for the super-diffusion subgroup f g h ** ** ** (d) and for the confined and sub-diffusion * * * * * 1 ** ** 10 subgroups combined (e) were plotted from 1 three individual experiments. Data on -1 10 nSP70, SP300, and SP1000 used a total of -1 1 10 474, 218, and 93 trajectories, respectively. f 10-2 to h, Quantitative analysis of trajectories of -3 10 the intracellular movements of nSP70, 10-1 10-2 SP300, and SP1000. In f, spreads of trajec10-4 tories within 3 s were calculated from the 10-5 10-3 10-2 trajectories in d and e. Diffusion coefficients nSP70 SP300 SP1000 nSP70 SP300 SP1000 nSP70 SP300 SP1000 (g) and velocities (h) were calculated from the slope of the MSD plot of each trajectory. Quantitative data are represented as box plots. Each box encompasses the 25th to 75th percentiles; whiskers represent maximum and minimum values; horizontal line across each box is the median (50th percentile); and diamonds (♦) are geometric means. *P < 0.05, **P < 0.01.

nSP70

a

SP1000

SP300

b

nSP70

SP300

MSD (μm2)

SP1000

Velocity (μm/s)

Diffusion coefficient (μm²/s)

Spread of trajectory (μm)

MSD (μm2)

extracellular

2.10. Statistical analysis

proportion of particles co-localized with the endosomal markers was calculated. Particles that showed co-localization with endosomal markers over > 100 pixels (after segmentation, about 0.032 μm2) were defined as co-localized particles. We counted the total number of silica particles in the cell and the number of co-localized particles in the cell. We then calculated the co-localization rate [number of co-localized particles/number of total particles × 100 (%)].

Statistical analysis was conducted with Origin 9. Differences were compared by using Scheffé's method after analysis of variance (ANOVA). P < 0.05 was considered significant. 3. Results 3.1. Fluorescence images of silica particles in A549 cells

2.9. Dual-color fluorescence live-cell imaging of silica particles and endosome

We used silica particles as model nanoparticles because they are among the most common nanoparticles, and mesoporous silica nanoparticles are expected to be useful as drug-delivery carriers [7]. Furthermore, silica nanoparticles show higher dispersibility and lower ionization tendency than metal nanoparticles such as zinc dioxide and iron oxide [21]. Therefore, there is less concern that the size of silica particles will be changed by aggregation and degradation during the tracking of intracellular movements. For these reasons, we considered silica particles suitable for assessing the relationship between particle size and intracellular particle movement. The surface properties of the orange fluorescence (Rhodamine B)labeled silica particles (nSP70, SP300, and SP1000) that we used were unmodified (silanol group [eOH]). This dye was chosen because the fluorescence of Rhodamine B-labeled nSP70 did not photobleach within 10 s after fluorescence excitation with sufficient laser power to observe single nanoparticles; in contrast, fluorescein isothiocyanate (FITC)-labeled nSP70 showed rapid photobleaching (Supplementary Fig. S1a). The mean hydrodynamic particle diameters and standard deviation (s.d.) of nSP70, SP300, and SP1000 in PBS, as measured by dynamic light scattering as described previously [5], were 69.6 ± 0.4, 291.0 ± 3.7, and 1183.3 ± 36.8 nm, respectively. nSP70, SP300, and SP1000 were confirmed by TEM to be well-dispersed, smooth-surfaced spheres, as described previously [5]. The mean ( ± s.d.) numbers of

To visualize endosomal vesicles—namely, the EE, LE, and Ly—in live cells, A549 cells were seeded onto 35-mm glass-base dishes at 2 × 105 cells/2.5 mL per dish. After a 12-h incubation at 37 °C, the cells were transiently transfected with a plasmid encoding GFP-EEA1, GFP-Rab7, or mGFP-LAMP1 by using Lipofectamine 2000 (Invitrogen) in accordance with the manufacturer's protocol. Twenty-four hours after transfection, these GFP-fusion endosomal-marker-labeled cells were pulsed with 25 μg/mL silica particles in culture medium and then chased for an additional 3 h. Dual-color fluorescence images of the silica particles and endosomal vesicles were obtained by using dual-view optics for dual-color live-cell imaging. The particles within the endosomal vesicles were sorted by their trajectories; those that showed the same trajectories as the endosomal markers were defined as particles included within the endosomal vesicles. The trajectories of the particles were tracked automatically by using PTA (ImageJ). In six randomly chosen cells from three individual experiments (two cells were chosen in each individual experiment) the movements of vesicles that did not include silica particles were tracked manually by using MTrackJ (a plugin of ImageJ [20]). These trajectories were analyzed as above.

186

Journal of Controlled Release 260 (2017) 183–193

M. Aoyama et al.

Fig. 3. Microtubule-dependent movement of silica particles. A549 cells were incubated with nocodazole, a microtubule-polymerization inhibitor, for 3 h after a 3-h exposure to silica particles. The intracellular movements of the silica particles were observed by live imaging at 100-ms intervals under microtubule-inhibited conditions. a, Trajectories of silica particles under nocodazole treatment. nSP70 and SP300 treatments each show 100 trajectories, and SP1000-nocodazole(−) and SP1000-nocodazole(+) treatments show 61 and 56 trajectories, respectively. Scale bar represents 2 μm. b to d, Quantitative analysis of trajectories of intracellular movements of silica particles under nocodazole treatment. Spread of trajectories (b), diffusion coefficients (c) and velocities (d) were evaluated. Quantitative data were obtained from two individual experiments and are represented by box plots. Each box encompasses the 25th to 75th percentiles; whiskers represent maximum and minimum values; horizontal line across each box is the median (50th percentile); and diamonds (♦) are geometric means. The nSP70-nocodazole (−), nSP70-nocodazole (+), SP300nocodazole (−), SP300-nocodazole (+), SP1000-nocodazole (−), and SP1000-nocodazole (+) subtreatments had a total of 307, 298, 195, 270, 80, and 85 trajectories, respectively. **P < 0.01.

microscopy (Fig. 1a, Supplementary Fig. S2a)—even at 25 μg/mL, a dose at which it was difficult to observe fluorescence in the cell by using confocal microscopy (Supplementary Fig. S3a, b). Fluorescent dots of SP1000, which were detectable by confocal imaging (Supplementary Fig. S3a, b), were also observed clearly under inclined illumination fluorescence microscopy at the same dose (Fig. 1a). These fluorescent dots could not be observed after treatment with silica particles under low-temperature (4 °C) conditions, which inhibited the uptake of particulate matter into the cells (Fig. 1b, Supplementary Fig. S2b). Cells that were treated with the supernatant of Rhodamine-B-labeled silica nanoparticle solution at 37 °C to assess the influence of free molecules from the silica nanoparticles showed no accumulation of fluorescence (Fig. 1c, Supplementary Fig. S2c). Rhodamine B solution with a fluorescence intensity equal to that of Rhodamine-B-labeled nSP70 suspension stained tubule-like compartments at 37 °C (Supplementary Fig. S3c). We checked the localization of Rhodamine B on the mitochondria (Supplementary Fig. S3d), because some kinds of Rhodamine are known to accumulate in the mitochondria [23,24]. However, we found no co-localization with the mitochondrial marker (cytochrome c oxidase subunit 4 isoform 1; CoxIV). Therefore, we did not clearly identify the organelles with tubule-like compartments, but they were not mitochondria. The data suggest that the fluorescent dots in the silicaparticle-treated cells were intracellular silica particles, not silica particles floating in the medium or free fluorescent molecules from the silica particles. Furthermore, when cells exposed to nSP70, SP300, or SP1000 at the same dose (25 μg/mL) were observed by TEM, only 1 to 3 particles were found in part of the cytosol (≤ about 0.2 × 0.2 μm2, i.e. in the area under the theoretical point spread function) or in a single vesicle (such as the endosome), regardless of particle size (Fig. 1d, e and Supplementary Fig. S4a). We counted the mean numbers of particles localized in a single place (i.e. in part of the cytosol or in a single vesicle). Counting of the mean numbers of nSP70 or SP300 in part of the cytosol or in a single vesicle on these TEM images revealed little more than one particle (nSP70, 1.3 ± 0.6 particles; nSP300, 1.2 ± 0.8 particles; n = 45 and 32 places, respectively). When we overdosed the cells with nSP70 (100 μg/mL) for 24 h, 6.7 ± 5.7 particles (n = 31 places) were detected in one place on the TEM images (Supplementary

fluorophores in each individual nSP70, SP300, and SP1000 particle were measured from the fluorescence images (Supplementary Fig. S1b). The mean fluorescence intensities of Rhodamine B, nSP70, SP300, and SP1000 were 1.7 ± 0.2, 12.2 ± 3.9, 115.0 ± 56.1, and 3985.3 ± 2132.7 (×104, a.u.), respectively. Thus, nSP70 had about 7 ± 2 Rhodamine B-fluorophores; SP300 had 67 ± 33, and SP1000 had 2329 ± 1246. In addition, to assess the abundance of free fluorescent molecules from the silica particles, we measured the fluorescence intensity of the supernatant after centrifuging a suspension of silica particles of each size labeled with Rhodamine B, or Rhodamine B solution alone (Supplementary Fig. S1c). There was barely any fluorescence in the supernatant of each type of silica particle after centrifugation, whereas the fluorescence intensity of the Rhodamine B solution did not change after centrifugation. This result suggested that virtually no free fluorescent molecules were present in the silica particle suspensions. Real-time imaging with sufficiently high temporal resolution is required to monitor in detail the movements of intracellular particles. However, the smallness of nanoparticles limits the numbers of fluorescent molecules in those particles, making it difficult to track their movements. Therefore, to detect single particles with small numbers of fluorescent molecules in a cell, we need to observe fluorescence with a high signal-to-noise ratio. Hence, to investigate the intracellular motility of silica particles, we acquired real-time images of nSP70-, SP300-, or SP1000-treated cells by using inclined illumination. Inclined illumination provides a higher signal-to-noise ratio than does epi-illumination at the intracellular focal plane, including in the area away from the cell membrane [22]; this area is difficult to observe by using TIRF microscopy, which enables selective visualization of the cell membrane. The optical layer thickness under inclined illumination fluorescence microscopy, as calculated from the fluorescence intensity of z-stack images in accordance with the method used in a previous study [22], was 9.3 μm (Supplementary Fig. S1d); this thickness was almost the same as in the previous study. After pulsing of A549 (human lung carcinoma) cells with silica particles for 3 h and chasing without silica particles for an additional 3 h, fluorescent dots representing nSP70 were found in the cells by using inclined illumination fluorescence 187

Journal of Controlled Release 260 (2017) 183–193

M. Aoyama et al.

Fig. 4. Silica particle localization in endosomes. A549 cells were incubated with silica particles for 3 h and then chased for an additional 3 h without silica particles. The cells were fixed and permeabilized and then immunostained with anti-EEA1, anti-Rab7, or anti-LAMP1 antibody. a, b, Fluorescence images of silica particles (Rhodamine B, red) and endosomal markers (Alexa-488; green). White dotted lines indicated cell contours (a). Some of the silica particles co-localized with each type of endosomal marker (b, magnification of part of each panel in a). Arrows indicate co-localization of silica particles and endosomal vesicles in b. Scale bars represent 10 μm in a and 2 μm in b. c, Percentage co-localizations of silica particles and each type of endosomal vesicle were analyzed from fluorescence images of 17 or 18 cells from two individual experiments. Error bars represent standard deviation of the mean. †P < 0.05 vs. nSP70 and *P < 0.05, **P < 0.01.

excluded trajectories with α > 2 or R2 < 0.9 and then defined the remaining trajectories as three types of model of intracellular movement according to the α value, with two digits (confined, α < 0.4; subdiffusion, 0.4 ≤ α ≤ 1.0; and super-diffusion, 1.0 < α ≤ 2.0) in accordance with previous examinations of the motility of microspheres in reconstitution experiments or in endosomes in the cell [26,27]. When we applied these definitions, the proportions of super-diffusion, subdiffusion, and confined particles showed no significant changes among nSP70, SP300, and SP1000 (Fig. 2c). In the super-diffusion (activetransport-like movement, 1.0 < α ≤ 2.0) group the MSD curve developed a parabolic shape with time (Fig. 2d); this shape is similar to that of microtubule-dependent active movements in the cell [28]. In contrast, the plot of MSD versus time with an α value of ≤ 1.0 approximated a linear equation [26]. Indeed, the MSD curves of the subdiffusion and confined groups, which had α values of ≤1.0, developed a linear shape with time (Fig. 2e); this is similar to the shape of the diffusion of molecules or particles in the cell [26,28]. These data suggest that particles of each size had two types of movement—active movement and diffusion—and that particle size did not affect the proportions of particles in active movement or diffusion. To quantitatively assess particle movement, we measured the spread of the trajectories in 3 s from the start point of trajectory tracking (Fig. 2f, Supplementary Fig. S6a). Trajectory spread diminished significantly with increasing particle size. These results suggest that intracellular movement of the particles was determined by particle size. Next, we calculated the diffusion coefficient (a parameter of thermodynamic and passive simple diffusion movement) and velocity (a parameter of motor-protein-dependent active movement) from the MSD (Fig. 2g, h, Supplementary Fig. S6b, c). The diffusion coefficient and the particle velocity decreased significantly with increasing particle size; nSP70 showed significantly greater motility than did the sub-micron particles (SP300 and SP1000). It is possible that many of the SP1000 did not show active movement in the cell, because they had very slow velocities. Furthermore, we evaluated the intracellular movements of some different-sized silica nanoparticles (30 nm, 50 nm, and 100 nm; Supplementary Fig. S7). nSP30, nSP50, nSP70, and nSP100 had almost the same intracellular motility (i.e. spread of trajectories, diffusion coefficients, and velocities), whereas SP300 had significantly slower motility than these silica nanoparticles. In addition, we assessed the

Fig. S4b). Moreover, under inclined illumination fluorescence microscopy we measured the fluorescence intensities of nSP70 localized on the glass-base dish and those of the fluorescent dots in nSP70-treated cells (pulsed with 25 μg/mL nSP70 for 3 h and chased without silica particles for an additional 3 h) and in cells treated with an overdose of nSP70 for 24 h (Supplementary Fig. S5). The fluorescence intensities of nSP70 localized on the glass-base dishes and of the fluorescent dots in the 25 μg/mL nSP70-treated cells were almost the same (areas under the fluorescence intensity curve, 490.2 ± 142.4 and 657.0 ± 319.6; n = 15, respectively; Supplementary Fig. S5). In contrast, from a single fluorescent dot in cells overdosed with nSP70 we found > 10 times the fluorescence intensity (area under the curve 7983.0 ± 5981.9; n = 15) from a single dot in 25 μg/mL nSP70-treated cells (Supplementary Fig. S5). These results suggest that the fluorescent dots we observed were nearly all single particles, not aggregations of many particles. 3.2. Intracellular movement of different-sized silica particles Next, we attempted to track the intracellular movements of the particles by analyzing the position of each fluorescent dot as a single particle. The x–y positions of the silica particles were determined by two-dimensional Gaussian fitting of the fluorescence intensity profile of each dot. We first examined the intracellular motility of different-sized silica particles. Various types of intracellular movement of nSP70 were detected during the imaging (Supplementary Movie 1), whereas the sub-micron-sized particles (SP300, SP1000) moved little (Supplementary Movies 2, 3). Two-dimensional analysis of the particle tracking revealed that some nSP70 showed directional movements similar to kinesin-dependent movements of peroxisome [25], whereas many particles of sub-micron size moved relatively slowly (Fig. 2a, b). Next, for more detailed characterization of these intracellular movements, we extracted the movement trajectories of those silica nanoparticles that had been detected continuously for 3 s and analyzed the MSD and the distribution of the alpha-coefficient (α) from these trajectories. The α-coefficients of the MSD curves were calculated by fitting to a power law, log [ρ(Δt)] = α log [(Δt)] + ζ, where α is the αcoefficient and ζ is a noise term. It is known that α-coefficients reflect the intracellular movements of molecules or particles [26]. We 188

Journal of Controlled Release 260 (2017) 183–193

M. Aoyama et al.

Fig. 5. Dual-color live-cell imaging of silica particles and various endosomes. A549 cells expressing GFP-fusion endosomal markers (EEA1, Rab7, or LAMP1) were incubated with silica particles. After a 3-h pulse and 3-h chase, the movements of the silica particles and endosomal markers were tracked at the same time by using dual-color live-cell imaging. The motilities of silica particles within early endosomes (EE), late endosomes (LE), or lysosomes (Ly), and of endosomal vesicles that did not contain silica particles, were evaluated qualitatively from the trajectories. a to c, Quantitative analysis of trajectories of intracellular movements of silica particles within EE, LE, and Ly. Spread of trajectories (a), diffusion coefficients (b), and velocities (c) were evaluated from three individual experiments (nSP70 and SP300 in EE had a total of 41 and 27 trajectories; nSP70 and SP300 in LE had a total of 210 and 76 trajectories; and nSP70 and SP300 in Ly had a total of 214 and 65 trajectories, respectively). d to f, Quantitative analysis of trajectories of EE, LE, or Ly that did not contain silica particles, in nine randomly chosen cells from three individual experiments (three cells per each experiment). Spread of trajectories (d), diffusion coefficients (e), and velocities (f) were evaluated. EE in the untreated and nSP70- or SP300-treated groups had a total of 116, 93, and 74 trajectories; LE in the untreated and nSP70- or SP300-treated groups had a total of 230, 200, and 190 trajectories; and Ly in the untreated and nSP70- or SP300-treated groups had a total of 291, 243, and 266 trajectories, respectively. All data are represented by box plots. Each box encompasses the 25th to 75th percentiles; whiskers represent maximum and minimum values; horizontal line across each box is the median (50th percentile); and diamonds (♦) are geometric means. *P < 0.05, **P < 0.01.

internalizes to the cell through the endocytic pathway [29,30]. We confirmed here that uptake of silica particles of all three sizes could be inhibited by low temperature (4 °C), which inhibited endocytosis (Supplementary Fig. S9). Polystyrene nanoparticles, nanosized polyplex, silica nanoparticles, and other nanoparticles become localized in endosomal vesicles—namely, the EE, LE, and Ly—after endocytosis [31–33]. It is therefore likely that the nanoparticles are transported by these vesicles in the cell. We tried to confirm the role of vesicle transport of the endocytic pathway in particle active movement. Because the EE, LE, and Ly are moved along the microtubules by the motor proteins kinesin and dynein [34] and microtubules function in the transcytosis of polystyrene nanoparticles [35], we first assessed whether silica

intracellular movements of surface-modified silica nanoparticles, namely nSP70 with carboxyl surface functional groups (nSP70-C) (Supplementary Fig. S8). nSP70-C had almost the same motility as nSP70, suggesting that surface properties were not important in the active movement of intracellular particles. The size-dependent declines in the diffusion coefficient and velocity suggest that particle size determined the amount of intracellular motility (i.e., the diffusion coefficient and velocity) in the cells. 3.3. Microtubule-dependent movement of silica particles It is well known that particulate matter, including nanoparticles, 189

Journal of Controlled Release 260 (2017) 183–193

M. Aoyama et al.

with Rab7 (marker protein of LE; 21% ± 13%) and LAMP1 (marker protein of Ly; 44% ± 17%) than with EEA1 (marker protein of EE, 4.9% ± 3.5%). SP300 showed either markedly or significantly greater co-localization with Rab7 or LAMP1 (25% ± 22% and 31% ± 17%, respectively) than with EEA1 (13% ± 11%). In contrast, SP1000 showed markedly greater co-localization with Rab7 (42% ± 34%) than with EEA1 or LAMP1 (23% ± 29% and 24% ± 31%, respectively). In addition, SP1000 showed significantly greater co-localization with EEA1 and Rab7 than did nSP70, whereas the co-localization of SP1000 with LAMP1 was significantly lower than that of nSP70 (Fig. 4c). Although some LAMP1 is known to be localized in LE [38], our results—especially for SP1000—showing greater co-localization with Rab7 than with LAMP1 suggest that the proportions of particles localized in EE, LE, and Ly changed with particle size: small particles were localized mainly in Ly, whereas larger particles were localized mainly in EE and LE.

particle active movement depended on microtubule function. We confirmed that treatment with1 μM of the microtubule polymerization inhibitor nocodazole for 3 h made the microtubule network unstable; the disrupted network recovered 30 min after removal of the nocodazole (Supplementary Fig. S10). Therefore, to assess the contribution of microtubule-dependent vesicle transport to particle movement, we observed the movements of the particles under nocodazole treatment and tracked their trajectories (Fig. 3a, b). Inhibition of microtubuledependent transport decreased the spread of the trajectories of nSP70 and SP300 significantly compared with those in the nocodazole (−) treatments. In contrast, there was no change in the spread of SP1000 under nocodazole treatment compared with that in the nocodazole (−) treatment. Furthermore, inhibition of microtubules led to some decrease in the percentages of nSP70 and SP300 with “super-diffusion”, whereas in the case of SP1000 this percentage did not change (Supplementary Fig. S11). We therefore calculated the diffusion coefficients and velocities of silica particles under nocodazole treatment (Fig. 3c, d). The diffusion coefficients and velocities of nSP70 and SP300 in the presence of nocodazole were significantly lower than those in the nocodazole (−) groups. In contrast, like the trajectory spread, the diffusion coefficient and velocity of SP1000 showed no decrease under nocodazole treatment. However, we were concerned that the results for the active movement of SP1000 were under the limit of the tracking resolution, regardless of whether nocodazole treatment was given. We therefore tracked the movements of SP1000 attached to the glass surface outside the cells (n = 10). Four particles could not be fitted (MSD curves of these particles gave R2 < 0.9 upon fitting by a power law [log [ρ(Δt)] = α log [(Δt)] + ζ]); only one particle showed active movement (2.0 × 10− 3 μm/s; slower than all the active movements of SP1000, regardless of whether nocodazole treatment was given), one particle showed sub-diffusion movement, and the other four particles showed confined movement (α < 0.4). Therefore, we considered that the independent active movements of SP1000 microtubules were not false negative results induced by limitations of the tracking resolution. These results suggest that nSP70 and SP300 moved along the microtubules, whereas the movement of SP1000 did not rely on fast microtubule-dependent transportation. The intracellular movements of nSP70 and SP300 were thus completely different from that of SP1000 and were size dependent.

3.5. Dual live-cell imaging of silica particles and various endosomes The above results indicated that nSP70 and SP300 moved differently from SP1000 in the cell: nSP70 and SP300 had microtubule-dependent fast active movements, whereas SP1000 showed very little movement, which was independent of the microtubules. Therefore, next, we assessed in detail the intracellular movements of nSP70 and SP300 localized in the endosomal vesicles in order to unravel the mechanisms behind the difference in the amounts of motility between nSP70 and SP300. To investigate whether the amounts of motility of nSP70 and SP300 localized in EE, LE, or Ly differed, we used dual-color fluorescence live imaging of silica particles and a GFP-fusion endosome marker (EEA1, Rab7, or LAMP1). These GFP-fusion endosome markers were co-localized with immunostaining of the same endosome markers by antibodies (Supplementary Fig. S15). After the expression of each GFP-fusion endosome marker, the movements of both the silica particles and the endosome markers were observed by using dual-color imaging. We extracted the movements of those silica particles that colocalized and moved with the endosome markers and defined these particles as being localized in EE, LE, or Ly. The percentage co-localizations of nSP70 or SP300 and GFP-fusion endosomal markers, as calculated from the dual-color imaging, were almost the same as in Fig. 4c (Supplementary Fig. S16). We then assessed the spread of trajectories as an indicator of motility (Fig. 5a; Supplementary Fig. S17a). The spread of the trajectories of SP300 localized in LE or Ly was significantly narrower than those of nSP70 in LE or Ly. These data suggest that particle size determined the movements of the endosomal vesicles—especially of LE and Ly that contained particles. To reveal whether the difference in the amount of motility was due to simple diffusion or active movement, we calculated the diffusion coefficients and velocities (Fig. 5b, c; Supplementary Fig. S17b, c). In LE and Ly, the diffusion coefficient and velocity of SP300 were significantly smaller than those of nSP70. Although the diffusion coefficient and velocity of SP300 localized in EE did not differ significantly from those of nSP70, the same declining trend in motility of SP300 as in LE and Ly was observed. In addition, with the exception of our finding that the spread of trajectories of nSP70 localized in EE was significantly narrower than that of nSP70 localized in Ly, the spreads of trajectories, diffusion coefficients, and velocities of either nSP70 or SP300 localized in each kind of endosomal vesicle did not differ significantly from those of particles of the same size localized in other kinds of endosomal vesicle. These data suggest that, overall, nSP70 and nSP300 had differing effects on the motility of the endosomal vesicles in which they were located, but there were no differences in motility among the different types of endosome in which each type of particle was localized. We were concerned that exposure to silica particles—especially SP300—might inhibit the motility of all endosomal vesicles, regardless of whether they contained particles. Therefore, in cells exposed to silica particles, we assessed the changes in motility of each type of endosomal

3.4. Silica particle localization in endosomal vesicles The three types of endosomal vesicles (EE, LE, and Ly) show different motilities; for example, LE moves faster than EE along the microtubules [36]. Consequently, the differences in intracellular movement among nSP70, SP300, and SP1000 might have been caused by differences in the types of endosomal vesicles in which they were localized. We therefore assessed the localization of the silica particles by using a co-localization experiment with EEA1, Rab7, and LAMP1 as EE, LE, and Ly markers, respectively (Fig. 4a, b). Some nSP70, SP300, and SP1000 were localized in the EEA1-, Rab7-, and LAMP1-positive endosomal vesicles. Next, to quantitatively evaluate the co-localization of silica particles and endosomes, we segmented the images and calculated the proportions of co-localized particles (Fig. 4c, Supplementary Figs. S12, S13). We confirmed that transferrin, as a positive control for EE, showed significantly greater co-localization with EEA1 than with Rab7 or LAMP1 (Supplementary Fig. S14). Totals of at least 60% of each of nSP70, SP300, and SP1000 in the cell were localized in EEA1-, Rab7-, and LAMP1-positive endosomal vesicles (Fig. 4c). These results suggested that about 40% of nSP70 and 30% or less of SP300 were localized in EEA1-, Rab7-, and LAMP1-negative components. We considered that these particles were localized in EEA1-, Rab7-, and LAMP1negative endosomal vesicles (such as recycling endosomes), the cytosol, and other organelles (such as the nucleus), because we previously showed that some nSP70 were localized in the cytosol and nucleus [37]. On the other hand, nSP70 showed significantly greater co-localization 190

Journal of Controlled Release 260 (2017) 183–193

M. Aoyama et al.

different proportions, and that most of the nSP70 and SP300 were localized in Ly, whereas most of the SP1000 were localized in EE and LE (Fig. 4). This result is consistent with those of previous studies suggesting that smaller particles or dextrans show greater co-localization with lysosomes [9,42,43]. EE and LE are transported in a microtubuledependent way [36], whereas SP1000, which localize mainly in EE and LE, do not show microtubule-dependent fast movement. Therefore, we now hypothesize that SP1000-containing endosomal vesicles may not be able to move fast and microtubule dependently because of the large size of the SP1000; moreover, the different endosomal localization of SP1000 compared with those of nSP70 or SP300 appears to be independent of the movements of SP1000. We used plasmids encoding GFP-fusion endosomal markers to assess the movements of nSP70 and SP300 localized in EE, LE, or Ly to look for relationships between particle movement and type of endosome. Therefore, it was possible that our results differed from those under plasmid-untreated conditions because overexpression of Rab7 suppresses the motility of LE [41]. However, we found that the motility of nSP70 or SP300 localized in one kind of endosomal vesicle was almost the same as those of the same-sized particles localized in other kinds of endosomal vesicles (Fig. 5a, b, c). These data suggest that proportional differences in the localization of nSP70 and SP300 were not the immediate cause of the differences in motility between nSP70 and SP300. Our data are consistent with those of Deville et al. [44], who suggested that the velocities of 115-nm polystyrene particles localized in EE, LE, and Ly were almost the same. On the other hand, Lai et al. [45] suggested that 25-nm and 42-nm polymer nanoparticles showed different motilities owing to preferential localization to different types of vesicles. In summary, the intracellular movements of nanoparticles localized to different kinds of vesicles remain controversial; further study is needed to determine the relationship between particle movement and localization. We found here that exposure to nSP70 or SP300 decreased the motility of endosomal vesicles that were empty of particles (Fig. 5d, e, f). Some chemicals, such as U18666A, which induces cholesterol accumulation, suppress the motility of endosomal vesicles via Rab proteins such as Rab7 through the inhibition of kinesin [46]. Therefore, exposure to particulate matter might induce changes in cholesterol levels in the cell and suppress the motility of endosomal vesicles, even when no particles are included within the vesicles. Exposure to carbonblack nanoparticles downregulates cholesterol-efflux-related genes such as abcg1 and abca1 in the mouse lung [47]. This result suggests that particulate matter induces cholesterol accumulation and thus disturbs membrane trafficking in the cell. On the other hand, because the suppressive effects induced on empty vesicles by exposure to nSP70 and SP300 were the same, the mechanisms behind the different motilities of intracellular nSP70 and SP300 are likely independent of the influence of particle exposure on all endosomal vesicles in the cell. Furthermore, we demonstrated that nSP70 localized in the endosomal vesicles did not suppress the motility of those vesicles, because the velocity of vesicles that contained nSP70 was almost the same as that of particle-free endosomal vesicles in nSP70-treated cells (Fig. 5c, f), whereas nSP300 localized in the vesicles had further motility-suppression effects (Fig. 5a, b, c). These results showing that particle size affected the movement of endosomal vesicles were independent of the influence of expression of GFP-fusion endosomal markers, because both nSP70 and SP300 were localized in cells expressing these markers. On the other hand, in terms of particle-size-dependent perturbation, it is unclear why nSP70 within the endosomal vesicles did not affect vesicle motility, whereas SP300 did. We speculate that particle-size-dependent inhibition of endosomal vesicle transport is induced by a change in size of the endosomal vesicles containing sub-micron particles, because it has been reported that the size of endosomal vesicles changes with changes in the kinds of particles (polylysine polyplex or polyamidoamine polyplex) localized within them [48]. In theory, if all conditions (e.g., temperature and viscosity of the solvent) except the

vesicle that did not contain any silica particles (Fig. 5d, e, f). The spread of trajectory of nSP70- or SP300-treated Ly that did not contain particles was significantly lower than that of untreated Ly. The spreads of trajectories of particle-free nSP70- or SP300-treated EE and LE were non-significantly lower than those of untreated EE and LE. Furthermore, the diffusion coefficients of particle-free nSP70- or SP300-treated EE and of particle-free SP300-treated Ly were significantly lower than those of untreated EE and of untreated Ly, respectively, and similar but not significant declines were observed in the case of LE. In addition, particle-free Ly that had been treated with nSP70 or SP300 had a significantly lower velocity than untreated Ly, and similar but non-significant declining trends were observed in the case of particle-free EE or LE that had been treated with nSP70 or SP300. These data suggest that, in terms of both diffusion and active movement, exposure of the cell to silica particles even inhibited the motility of endosomal vesicles that did not include these particles. However, the amount of motility of treated endosomal vesicles that did not include particles did not differ significantly between nSP70- and SP300-treated cells. Taken together, these results suggest that, although exposure to silica particles suppressed the intracellular movement of endosomal vesicles across the board, nSP70 did not further suppress the intracellular movement of endosomal vesicles by localizing within those vesicles. In contrast, in addition to this across-the-board suppression, the motility of endosomal vesicles that included SP300 was further suppressed by the localization of SP300 to these vesicles. We consider that the difference in intracellular motility between nSP70 and SP300 resulted from a difference in perturbation effects on endosomal vesicle transport. 4. Discussion Here, we assessed the intracellular movement of silica nanoparticles and sub-micron particles at the single-particle level in the living cell by using inclined illumination fluorescence microscopy. We first revealed that silica nanoparticles (nSP70) showed greater diffusion and faster active movement than silica sub-micron particles (SP300 and SP1000) (Fig. 2f, g, h). nSP70 and SP300 showed microtubule-dependent fast active movements similar to those of the endosomal vesicles, whereas SP1000 showed microtubule-independent, slow movement (Fig. 3). More than half of the particles in the cell were localized in endosomal vesicles (Fig. 4c). We therefore focused on the movements of endosomal vesicles that included particles and tried to assess the mechanisms behind the different motilities of the different-sized silica particles. We found that particle size, but not particle surface properties, was an important determinant of the intracellular movements of silica particles. It is well known that surface modifications are important for the intracellular dynamics of these particles—especially cellular uptake—associated with changes of protein corona on the nanoparticle surface. Our finding of a lack of difference in intracellular movements of nSP70-C compared with nSP70 appeared to conflict with these data (Supplementary Fig. S8). On the other hand, the protein corona separates gradually from the nanoparticle surface under low pH conditions in endosomal vesicles [39,40]. Therefore, dissociation of the protein corona after nSP70 had become internalized in the cells may have been the reason why particle surface modifications did not affect the intracellular movements of these particles in our study. However, it is possible that the protein corona is reconstituted in the cytosol, and further studies of surface-modified groups are needed. Endosomal vesicles move along the microtubules, and the velocities of unidirectional movement of EE, LE, and Ly differ (0.27, 0.49, and 0.45 μm/s, respectively) [36,41]. In addition, particles and dextrans of different sizes localize in different proportions in different kinds of endosomal vesicles [9,42]. Therefore, we considered that the different motilities of nSP70 and SP300 likely resulted from relative differences in the localizations of nSP70 and SP300 in EE, LE, and Ly. Indeed, we found that nSP70, SP300, and SP1000 localized in EE, LE, and Ly in 191

Journal of Controlled Release 260 (2017) 183–193

M. Aoyama et al.

the Ministry of Health, Labour and Welfare, Japan (No. H25-kagakuippan-005 to Y.T.); and by The Uehara Memorial Foundation (to Y.Y.).

size of the endosomal vesicles are the same, then the diffusion coefficient will be in inverse proportion to the size of the endosomal vesicle according to the Stokes-Einstein equation [49,50]. In addition, a recent study of the intracellular motility of peroxisomes has suggested that the velocity of kinesin-1-dependent active movement decreases as the size of the peroxisome increases [51]. This finding suggests that the size of the cargo vesicles (e.g., peroxisomes or endosomes) transported by kinesin or dynein controls the velocity of the cargo vesicles. Although the relationship between cargo size and velocity of active movement in the cell is still controversial [52,53], taking into account both our results and those of Efremov et al. [51], we consider that the mechanism behind the different motilities of nSP70 and SP300 localized in the endosomal vesicles might be an increase in size of the endosomal vesicles containing SP300 but not of those containing nSP70. Therefore, to reveal the mechanism in future, we need to assess the abundance of connecting motor proteins and the sizes of the endosomal vesicles containing nSP70 and SP300. Here, we did not demonstrate whether it was particle size or particle weight that was important in the difference in intracellular dynamics between nSP70 and SP300 (nSP70 is smaller and lighter than nSP300: nSP300 is about 4.3 times the diameter of nSP70 and about 77.8 times heavier). We therefore also need to assess the contribution of particle size and weight to the intracellular movements of particulate matter. It has been reported that different kinds of molecules, such as those of receptors, nutrients, and foreign materials, are sorted by EEs [54,55]. However, recently, Kasmapour et al. [56] suggested that Rab34 regulates molecular-size-selective transportation of LE or Ly into phagosomes. Our data and theirs suggest that the intracellular dynamics of endosomal vesicles are controlled not only by molecular sequences (such as through ubiquitylation [57]) but also by the size or weight of the included molecules or particles. Therefore, an understanding of the relationship between particle properties and intracellular dynamics may help us to unravel the mechanisms of membrane trafficking of endosomal vesicles.

References [1] M. Das, N. Saxena, P.D. Dwivedi, Emerging trends of nanoparticles application in food technology: safety paradigms, Nanotoxicology 3 (2009) 10–18. [2] D.M. Bowman, G. van Calster, S. Friedrichs, Nanomaterials and regulation of cosmetics, Nat. Nanotechnol. 5 (2010) 92. [3] J. Lee, S. Mahendra, P.J. Alvarez, Nanomaterials in the construction industry: a review of their applications and environmental health and safety considerations, ACS Nano 4 (2010) 3580–3590. [4] R. Kumar, I. Roy, T.Y. Ohulchanskky, L.A. Vathy, E.J. Bergey, M. Sajjad, P.N. Prasad, In vivo biodistribution and clearance studies using multimodal organically modified silica nanoparticles, ACS Nano 4 (2010) 699–708. [5] K. Yamashita, Y. Yoshioka, K. Higashisaka, K. Mimura, Y. Morishita, M. Nozaki, T. Yoshida, T. Ogura, H. Nabeshi, K. Nagano, Y. Abe, H. Kamada, Y. Monobe, T. Imazawa, H. Aoshima, K. Shishido, Y. Kawai, T. Mayumi, S. Tsunoda, N. Itoh, T. Yoshikawa, I. Yanagihara, S. Saito, Y. Tsutsumi, Silica and titanium dioxide nanoparticles cause pregnancy complications in mice, Nat. Nanotechnol. 6 (2011) 321–328. [6] J. Liu, M. Yu, C. Zhou, S. Yang, X. Ning, J. Zheng, Passive tumor targeting of renalclearable luminescent gold nanoparticles: long tumor retention and fast normal tissue clearance, J. Am. Chem. Soc. 135 (2013) 4978–4981. [7] X. Liu, A. Situ, Y. Kang, K.R. Villabroza, Y. Liao, C.H. Chang, T. Donahue, A.E. Nel, H. Meng, Irinotecan delivery by lipid-coated mesoporous silica nanoparticles shows improved efficacy and safety over liposomes for pancreatic cancer, ACS Nano 10 (2016) 2702–2715. [8] J. Shi, B. Wang, L. Wang, T. Lu, Y. Fu, H. Zhang, Z. Zhang, Fullerene (C60)-based tumor-targeting nanoparticles with “off-on” state for enhanced treatment of cancer, J. Control. Release 235 (2016) 245–258. [9] M. Al-Rawi, S. Diabate, C. Weiss, Uptake and intracellular localization of submicron and nano-sized SiO2 particles in HeLa cells, Arch. Toxicol. 85 (2011) 813–826. [10] H. Cabral, Y. Matsumoto, K. Mizuno, Q. Chen, M. Murakami, M. Kimura, Y. Terada, M.R. Kano, K. Miyazono, M. Uesaka, N. Nishiyama, K. Kataoka, Accumulation of sub-100 nm polymeric micelles in poorly permeable tumours depends on size, Nat. Nanotechnol. 6 (2011) 815–823. [11] A. Albanese, P.S. Tang, W.C. Chan, The effect of nanoparticle size, shape, and surface chemistry on biological systems, Annu. Rev. Biomed. Eng. 14 (2012) 1–16. [12] M. Murakami, H. Cabral, Y. Matsumoto, S. Wu, M.R. Kano, T. Yamori, N. Nishiyama, K. Kataoka, Improving drug potency and efficacy by nanocarriermediated subcellular targeting, Sci. Transl. Med. 3 (2011) 64ra62. [13] S. Huo, S. Jin, X. Ma, X. Xue, K. Yang, A. Kumar, P.C. Wang, J. Zhang, Z. Hu, X.J. Liang, Ultrasmall gold nanoparticles as carriers for nucleus-based gene therapy due to size-dependent nuclear entry, ACS Nano 8 (2014) 5852–5862. [14] K.I. Joo, Y. Fang, Y. Liu, L. Xiao, Z. Gu, A. Tai, C.L. Lee, Y. Tang, P. Wang, Enhanced real-time monitoring of adeno-associated virus trafficking by virus-quantum dot conjugates, ACS Nano 5 (2011) 3523–3535. [15] C. Rosazza, A. Buntz, T. Riess, D. Woll, A. Zumbusch, M.P. Rols, Intracellular tracking of single-plasmid DNA particles after delivery by electroporation, Mol. Ther. 21 (2013) 2217–2226. [16] C. Manzo, M.F. Garcia-Parajo, A review of progress in single particle tracking: from methods to biophysical insights, Rep. Prog. Phys. 78 (2015) 124601. [17] R. Bausinger, K. von Gersdorff, K. Braeckmans, M. Ogris, E. Wagner, C. Brauchle, A. Zumbusch, The transport of nanosized gene carriers unraveled by live-cell imaging, Angew. Chem. Int. Ed. Eng. 45 (2006) 1568–1572. [18] E. Zagato, K. Forier, T. Martens, K. Neyts, J. Demeester, S. De Smedt, K. Remaut, K. Braeckmans, Single-particle tracking for studying nanomaterial dynamics: applications and fundamentals in drug delivery, Nanomedicine (London) 9 (2014) 913–927. [19] A. Rizk, G. Paul, P. Incardona, M. Bugarski, M. Mansouri, A. Niemann, U. Ziegler, P. Berger, I.F. Sbalzarini, Segmentation and quantification of subcellular structures in fluorescence microscopy images using Squassh, Nat. Protoc. 9 (2014) 586–596. [20] E. Meijering, O. Dzyubachyk, I. Smal, Methods for cell and particle tracking, Methods Enzymol. 504 (2012) 183–200. [21] T.J. Brunner, P. Wick, P. Manser, P. Spohn, R.N. Grass, L.K. Limbach, A. Bruinink, W.J. Stark, In vitro cytotoxicity of oxide nanoparticles: comparison to asbestos, silica, and the effect of particle solubility, Environ. Sci. Technol. 40 (2006) 4374–4381. [22] M. Tokunaga, N. Imamoto, K. Sakata-Sogawa, Highly inclined thin illumination enables clear single-molecule imaging in cells, Nat. Methods 5 (2008) 159–161. [23] L.V. Johnson, M.L. Walsh, L.B. Chen, Localization of mitochondria in living cells with rhodamine 123, Proc. Natl. Acad. Sci. U. S. A. 77 (1980) 990–994. [24] R.C. Scaduto Jr., L.W. Grotyohann, Measurement of mitochondrial membrane potential using fluorescent rhodamine derivatives, Biophys. J. 76 (1999) 469–477. [25] P. van Bergeijk, M. Adrian, C.C. Hoogenraad, L.C. Kapitein, Optogenetic control of organelle transport and positioning, Nature 518 (2015) 111–114. [26] B.M. Regner, D. Vucinic, C. Domnisoru, T.M. Bartol, M.W. Hetzer, D.M. Tartakovsky, T.J. Sejnowski, Anomalous diffusion of single particles in cytoplasm, Biophys. J. 104 (2013) 1652–1660. [27] A.L. Zajac, Y.E. Goldman, E.L. Holzbaur, E.M. Ostap, Local cytoskeletal and organelle interactions impact molecular-motor- driven early endosomal trafficking, Curr. Biol. 23 (2013) 1173–1180. [28] Y. Okada, N. Hirokawa, A processive single-headed motor: kinesin superfamily

5. Conclusions We showed here that endosomal vesicles containing 70-nm silica nanoparticle moved faster than those containing 300-nm sub-micron silica particle The differences in particle motility in the cells did not result from differences in preferential localization to endosomal vesicles; instead, they likely resulted from differences in perturbation effects on endosomal vesicle transport. Our results have revealed part of the mechanism of nanoparticle-specific intracellular dynamics and support a new hypothesis that differences in particle size or weight are responsible for differences in the membrane trafficking of endosomal vesicles. We believe that further study of the intracellular dynamics of nanoparticles and sub-micron particles would promote a better understanding of the mechanisms of nanoparticle-specific intracellular dynamics and of the physiology of membrane trafficking of endosomal vesicles. Such information would help to develop nanoparticles or drugdelivery carriers that could control intracellular dynamics. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jconrel.2017.06.007. Competing financial interests Y.Y. is employed by The Research Foundation for Microbial Diseases at Osaka University. All other authors declare no competing financial interests. Acknowledgments This study was supported by JSPS KAKENHI (No. JP25136712 to Y.Y., No. JP26-5340 to M.A., No. JP26242055 to Y.T. and No. JP15K12540 to Y.T.); by a Health Labour Sciences Research Grant from 192

Journal of Controlled Release 260 (2017) 183–193

M. Aoyama et al.

endocytosis in macrophages, J. Cell Biol. 129 (1995) 989–998. [43] E. Frohlich, C. Meindl, E. Roblegg, B. Ebner, M. Absenger, T.R. Pieber, Action of polystyrene nanoparticles of different sizes on lysosomal function and integrity, Part. Fibre Toxicol. 9 (2012) 26. [44] S. Deville, R. Penjweini, N. Smisdom, K. Notelaers, I. Nelissen, J. Hooyberghs, M. Ameloot, Intracellular dynamics and fate of polystyrene nanoparticles in A549 lung epithelial cells monitored by image (cross-) correlation spectroscopy and single particle tracking, Biochim. Biophys. Acta 1853 (2015) 2411–2419. [45] S.K. Lai, K. Hida, S.T. Man, C. Chen, C. Machamer, T.A. Schroer, J. Hanes, Privileged delivery of polymer nanoparticles to the perinuclear region of live cells via a non-clathrin, non-degradative pathway, Biomaterials 28 (2007) 2876–2884. [46] H. Chen, J. Yang, P.S. Low, J.X. Cheng, Cholesterol level regulates endosome motility via Rab proteins, Biophys. J. 94 (2008) 1508–1520. [47] J.A. Bourdon, S. Halappanavar, A.T. Saber, N.R. Jacobsen, A. Williams, H. Wallin, U. Vogel, C.L. Yauk, Hepatic and pulmonary toxicogenomic profiles in mice intratracheally instilled with carbon black nanoparticles reveal pulmonary inflammation, acute phase response, and alterations in lipid homeostasis, Toxicol. Sci. 127 (2012) 474–484. [48] N.D. Sonawane, F.C. Szoka Jr., A.S. Verkman, Chloride accumulation and swelling in endosomes enhances DNA transfer by polyamine-DNA polyplexes, J. Biol. Chem. 278 (2003) 44826–44831. [49] A. Einstain, Über die von der molekularkinetischen Theorie der Wärme geforderte Bewegung von in ruhenden Flüssigkeiten suspendierten Teilchen, Ann. Phys. 322 (1905) 549–560. [50] M.V. Smoluchowski, Zur kinetischen Theorie der Brownschen Molekularbewegung und der Suspensionen, Ann. Phys. 326 (1906) 756–780. [51] A.K. Efremov, A. Radhakrishnan, D.S. Tsao, C.S. Bookwalter, K.M. Trybus, M.R. Diehl, Delineating cooperative responses of processive motors in living cells, Proc. Natl. Acad. Sci. U. S. A. 111 (2014) E334–E343. [52] R.P. Erickson, Z. Jia, S.P. Gross, C.C. Yu, How molecular motors are arranged on a cargo is important for vesicular transport, PLoS Comput. Biol. 7 (2011) e1002032. [53] S.R. Nelson, K.M. Trybus, D.M. Warshaw, Motor coupling through lipid membranes enhances transport velocities for ensembles of myosin Va, Proc. Natl. Acad. Sci. U. S. A. 111 (2014) E3986–E3995. [54] J. Gruenberg, The endocytic pathway: a mosaic of domains, Nat. Rev. Mol. Cell Biol. 2 (2001) 721–730. [55] F.R. Maxfield, T.E. McGraw, Endocytic recycling, Nat. Rev. Mol. Cell Biol. 5 (2004) 121–132. [56] B. Kasmapour, A. Gronow, C.K. Bleck, W. Hong, M.G. Gutierrez, Size-dependent mechanism of cargo sorting during lysosome-phagosome fusion is controlled by Rab34, Proc. Natl. Acad. Sci. U. S. A. 109 (2012) 20485–20490. [57] K. Haglund, I. Dikic, The role of ubiquitylation in receptor endocytosis and endosomal sorting, J. Cell Sci. 125 (2012) 265–275.

protein KIF1A, Science 283 (1999) 1152–1157. [29] F. Zhao, Y. Zhao, Y. Liu, X. Chang, C. Chen, Y. Zhao, Cellular uptake, intracellular trafficking, and cytotoxicity of nanomaterials, Small 7 (2011) 1322–1337. [30] D. Vercauteren, J. Rejman, T.F. Martens, J. Demeester, S.C. De Smedt, K. Braeckmans, On the cellular processing of non-viral nanomedicines for nucleic acid delivery: mechanisms and methods, J. Control. Release 161 (2012) 566–581. [31] D. Vercauteren, H. Deschout, K. Remaut, J.F. Engbersen, A.T. Jones, J. Demeester, S.C. De Smedt, K. Braeckmans, Dynamic colocalization microscopy to characterize intracellular trafficking of nanomedicines, ACS Nano 5 (2011) 7874–7884. [32] P. Sandin, L.W. Fitzpatrick, J.C. Simpson, K.A. Dawson, High-speed imaging of Rab family small GTPases reveals rare events in nanoparticle trafficking in living cells, ACS Nano 6 (2012) 1513–1521. [33] A. Lesniak, F. Fenaroli, M.P. Monopoli, C. Aberg, K.A. Dawson, A. Salvati, Effects of the presence or absence of a protein corona on silica nanoparticle uptake and impact on cells, ACS Nano 6 (2012) 5845–5857. [34] S. Balint, I. Verdeny Vilanova, A. Sandoval Alvarez, M. Lakadamyali, Correlative live-cell and superresolution microscopy reveals cargo transport dynamics at microtubule intersections, Proc. Natl. Acad. Sci. U. S. A. 110 (2013) 3375–3380. [35] A.M. Bannunah, D. Vllasaliu, J. Lord, S. Stolnik, Mechanisms of nanoparticle internalization and transport across an intestinal epithelial cell model: effect of size and surface charge, Mol. Pharm. 11 (2014) 4363–4373. [36] S. Loubery, C. Wilhelm, I. Hurbain, S. Neveu, D. Louvard, E. Coudrier, Different microtubule motors move early and late endocytic compartments, Traffic 9 (2008) 492–509. [37] H. Nabeshi, T. Yoshikawa, K. Matsuyama, Y. Nakazato, K. Matsuo, A. Arimori, M. Isobe, S. Tochigi, S. Kondoh, T. Hirai, T. Akase, T. Yamashita, K. Yamashita, T. Yoshida, K. Nagano, Y. Abe, Y. Yoshioka, H. Kamada, T. Imazawa, N. Itoh, S. Nakagawa, T. Mayumi, S. Tsunoda, Y. Tsutsumi, Systemic distribution, nuclear entry and cytotoxicity of amorphous nanosilica following topical application, Biomaterials 32 (2011) 2713–2724. [38] K. Dunster, B.H. Toh, J.W. Sentry, Early endosomes, late endosomes, and lysosomes display distinct partitioning strategies of inheritance with similarities to Golgi-derived membranes, Eur. J. Cell Biol. 81 (2002) 117–124. [39] F. Wang, L. Yu, M.P. Monopoli, P. Sandin, E. Mahon, A. Salvati, K.A. Dawson, The biomolecular corona is retained during nanoparticle uptake and protects the cells from the damage induced by cationic nanoparticles until degraded in the lysosomes, Nanomedicine 9 (2013) 1159–1168. [40] F. Bertoli, D. Garry, M.P. Monopoli, A. Salvati, K.A. Dawson, The intracellular Destiny of the protein corona: a study on its cellular internalization and evolution, ACS Nano 10 (2016) 10471–10479. [41] C. Lebrand, M. Corti, H. Goodson, P. Cosson, V. Cavalli, N. Mayran, J. Faure, J. Gruenberg, Late endosome motility depends on lipids via the small GTPase Rab7, EMBO J. 21 (2002) 1289–1300. [42] E.P. Berthiaume, C. Medina, J.A. Swanson, Molecular size-fractionation during

193