DLS only provides a Z average size based on spherical

DLS only provides a Z-average size based on spherical model, hence the size of nanorods was determined as 300 nm by DLS measurement. However, the size of length is nearly 400 nm while width is about 100 nm, which was clearly observed by transmission electron microscopy (TEM) (D). Meanwhile, the TEM images (C and D) also showed nanospheres and nanorods were quite uniform in size distribution. Though the size of nanorods was significant different with that of nanospheres, the nanorods was obtained by stretching nanospheres. In fact, the differences between nanorods and nanospheres are only caused by the change of shape. Thus, these nanoparticles are the optimal objects to evaluate the shape effect. In order to quantify and observe the intact nanoparticles, both nanoparticles were labeled with DiO and DiI which were able to result in fluorescence resonance energy transfer (FRET) effect upon loading into nanoparticles simultaneously. B showed there was no peak at 565 nm (DiI emission wavelength) in mixed solution of DiI and DiO upon excitation at 484 nm (DiO excitation wavelength), while the peak at 565 nm occurred when DiI and DiO were loaded into nanoparticles together, which demonstrated the occurrence of FRET effect. The behaviors of intact nanoparticles are able to be tracked by FRET probes because the FRET fluorescence can be detected only when the nanoparticles maintain intact []. Once the nanoparticles are destroyed, the FRET effect immediately disappears because the distance of two probes exceeds far beyond 10 nm. Therefore, this study employed the FRET probes (DiI and DiO) to label nanospheres and nanorods for elucidating the intracellular behaviors accurately. The fluorescent intensity of nanospheres and nanorods is similar after loading the FRET probes, which are 1.12 × 10 and 1.08 × 10 [p/s]/[μW/cm] by IVIS spectrum respectively. Caco-2 cpmt have been extensively used as an model of intestinal epithelia due to similar properties. In addition, the intestinal epithelia are covered by mucus layers protecting against bacterial infection []. The mucus is secreted by goblet cells which are substituted by HT29-MTX []. Therefore, single Caco-2 cell model and co-culture model of Caco-2 and HT29-MTX were developed to evaluate the uptake, transport and exocytosis of different shaped nanoparticles. A shows nanorods were taken up by both Caco-2 cells and co-culture models in greater amount than nanospheres. However, the uptake of nanoparticles in co-culture model decreases significantly compared to single Caco-2 cells, which could be ascribed to the impediment of mucus layer. After 2 h of uptake, nanoparticles were removed and the cells were washed by cold D-hank’s buffer for three times. And then, the fresh HBSS was added in cells for exocytosis of nanoparticles. Nanoparticles were detected in fresh HBSS and the amount increased with the extension of time. B shows the exocytosis rate in Caco-2 cells before 30 min was very fast and slowed down after 1 h, and the exocytosis rate of nanospheres was much faster than that of nanorods. There were over 40% of nanospheres exocytosed from Caco-2 cells within 4 h. The similar trends were observed in co-culture models (Figs. S1 and S2 in Supporting information), but with smaller amount of exocytosed nanoparticles in total compared with single Caco-2 cells. Nanoparticles are taken up by endocytosis mediated transport proteins, such as caveolin or clathrin, which is non-specific []. Nanorods contact with cells in multi-orientation. If they contact with cells horizontally, nanorods would exert higher endocytosis effect [,]. However, the contact area of nanospheres with cells is limited and unchanged in various directions. Most nanoparticles internalized into cells may suffer from degradation in lysosomes first and then escape into cytoplasma []. However, polystyrene is a non-biodegradable polymer which is rarely degraded in lysosomes [], hence the polystyrene nanoparticles should exist as intact particles within the cells. That is to say, the fluorescent signals detected in cells were able to represent the intact nanoparticles as shown in (C0∼C3 and D0∼D3). 2 h after exocytosis, the fluorescent intensity of nanospheres (C3) in cells was far lower than that of nanorods (D3). Most papers considered that the exocytosis of nanoparticles was mediated by microtubule [,]. Nanospheres are easier to enter into microtubule than nanorods which were hindered by their specific shape. The rod-like shape led to the long-term retention of particles in cells, which could produce additional toxicity or be used to treat diseases inside cells.