Characterization of aSiNPs
Two sizes of aSiNPs (64 nm and 46 nm), namely Nano-Si64 and Nano-Si46, have been fully characterized in our previous study [28]. Images of transmission electron microscopy (TEM) displayed the spherical or ellipsoidal shape of the particles (Fig 2). The purity of both Nano-Si64 and Nano-Si46, detected by inductively coupled plasma atomic emission spectrometry (ICP-AES), was higher than 99.9%. Their hydrodynamic diameter and Zeta potential were also measured in distilled water and in 1640 culture medium at 0, 3, 6, 12, and 24 h previously, results indicated that two aSiNPs maintained fairly good monodispersity in both storage medium and experimental medium.
Multinucleation induced by aSiNPs in vivo and in vitro
After the mice were repeatedly intratrcheally instilled with 64 nm aSiNPs, the pathological analyses of liver and lung tissue were performed. As shown in Figure 3A, the structure of liver lobules was clearly visible, and the hepatocyte cords were arranged radially around the central vein in the control group. While in aSiNPs treated group, nuclear fragmentation, vacuolization and necrosis of hepatocytes were observed nearly the central vein. Additionally, the rate of multinucleated hepatocytes increased significantly in the aSiNPs treated group in the both proximal region and distal region of the central vein. Figure 3B displayed HE staining images of lung tissue, lymphocyte infiltration, pulmonary interstitial thickening and bronchial epithelial damage were observed in the aSiNPs treated group, and a few multinucleate cells were found in bronchial epithelium. The above results revealed that repeated intratrcheal instillation of aSiNPs not only damaged lung tissue, but also meat the particles could also penetrate into liver through blood circulation and cause liver injury as well as increase the proportion of multinucleated hepatocytes.
The multinucleation effect of aSiNPs was also detected in vitro as well. As shown in Figure 3C, both Nano-Si64 and Nano-Si46 caused the formation of multinucleated cells in the L-02 cell line, and the rate of multinucleated cells increased with the particle dose. Moreover, the multinucleation effect of smaller aSiNPs (Nano-Si46) was significantly stronger than that of 64 nm aSiNPs. So, 46 nm aSiNPs was used for subsequent vitro studies to explore the underlying mechanisms of how aSiNPs induced increased an multinucleation rate in hepatocytes.
Microfilaments agglomeration induced by aSiNPs
In the previous study, we have already confirmed that abnormal cytokinesis induced by aSiNPs should be responsible for the formation of multinucleated cells. In order to further investigate the potential effect of aSiNPs on the cytoskeleton and cytokinesis, we synthesized SiO2 nanoparticles with red fluorescence encapsulated in the core.
The effect of aSiNPs on the structure and cellular distribution of microfilaments were observed by Laser scanning confocal microscopy (LSCM). As manifested in Figure 4, L-02 cells could take in the red fluorescence-labeled aSiNPs after 24 h treatment, and for most of the particles distributed near the center of the cells, relatively small amount of aSiNPs adhered to or were located near the cell surface. Beyond that, microfilaments agglomeration and an obvious co-localization between aSiNPs and the agglomerated microfilaments were identified (Fig. 4B, 4C), which indicated that the particles could directly interact with the microfilaments in cytoplasm and resulted in significant changes in their structure and intracellular distribution. In general, microfilaments can be bundled or scattered in the cytoplasm, which have relationships with cell morphology maintenance, cell and organelle movement, cell division and so on. Therefore, microfilament agglomeration resulted by aSiNPs was likely to cause a series of adverse consequences, including abnormal mitosis.
Abnormal microtubules distribution and cytokinesis induced by aSiNPs
The influence of aSiNPs on microtubules was shown in Figure 5A. In the control group the microtubules mainly distributed around the nucleus, forming a network and extending radially to the periphery. In the aSiNPs treated group, the morphology and structure of microtubules were not changed significantly, but due to the occupation of agglomerated microfilaments, the distribution of microtubules in the perinuclear cytoplasmic region was notably affected.
Abnormal cytokinesis induced by aSiNPs was displayed in Figure 5B. At the terminal stage of mitosis, the formed constriction could not separate the two daughter cells effectively, instead leaving a thin and long intercellular bridge between them. These incomplete cytokinesis cells might possibly recombine into one multinucleated cell and double the number of chromosomes, like the binucleated cells shown in Figure 5Ab and 5Ac. In addition, owing to the disappearance of the nuclear membrane during mitosis, aSiNPs entering the cells could directly interact with the chromosomes, causing chromosome breakage, and the formation of micronuclei outside the main nucleus in daughter cells (Fig. 5B white arrow heads). Thus, numerical aberration of chromosome followed by cytokinesis failure, as well as chromosome damage caused by aSiNPs would result in a further increase in chromosomal instability (CIN) of L-02 cells.
Down-regulation of the PI3K 110β/Aurora B pathway and cytokinesis regulatory proteins induced by aSiNPs
To better understand the mechanism of cytokinesis failure caused by aSiNPs, the protein expression of mitotic regulatory pathway PI3K/Aurora B signaling and cytokinesis regulators in L-02 cells were examined. Firstly, the protein expression levels of the catalytic subunit P85 and the regulatory subunits P110α and P110β of PI3K were detected. As shown in Figure 6, the expression of PI3K 110β was down regulated by aSiNPs significantly, but PI3K p85 and PI3K 110α were not affected by aSiNPs. Next, the phosphorylation level of Aurora B, a key downstream mitotic kinase, was also detected. The expression of Aurora B phosphorylation showed a downward trend after aSiNPs treatment. These results indicated that aSiNPs inhibited the regulatory activity of Aurora B in mitosis, mainly through the PI3K 110β subunit, possibly leading to the dysfunction of mitosis or cytokinesis.
The centralspindlin complex played a pivotal role in cytokinesis regulation, which was comprised of two subunits, MKLP1 and CYK-4. It was reported that Aurora B regulated the centralspindlin complex by phosphorylating the S708 site of MKLP1, thereby controlling the formation and bundling of the central spindle during anaphase and telophase of mitosis. Therefore, the protein content of MKLP1 and CYK4, as well as Ect2, Cep55, and CHMP2A, the key proteins that manipulated cell division downstream of centralspindlin were detected by western blot (Fig. 6A, 6B), and the activity of the Rho enzyme was assessed by G-LISA kit (Fig. 6C).
The results suggested that compared with the control, aSiNPs evidently decreased the expression of MKLP1 and CYK4. Moreover, Cep55 and CHMP2A which are downstream of MKLP1, and were related to central spindle fascicle formation, also decreased in a dose dependent way. A similar trend was found either in the Ect2 expression and Rho activity downstream of CYK4, which were concerned in the regulation of contractile ring and daughter cell shedding. The above observations demonstrated that aSiNPs could therefore reduce the content or activity of cytokines regulating proteins in L-02 cells, causing the disorder of telophase in cell division, and then contributing to the failure of cytokinesis.
Abnormal co-localization of Aurora B and Centralspindlin complex induced by aSiNPs
In anaphase and telophase of mitosis, S708 site phosphorylated by Aurora B was required for MKLP1 to localize to central spindle and cluster with CYK4. Hence, the co-localization of Aurora B and MKLP1 on the midbody was detected. Figure 7A confirmed that the normal co-localization of Aurora B and MKLP1 was altered by aSiNPs. The phenomenon of MKLP1 disappearance (Fig. 7Ab) and incorrect location of Aurora B and MKLP1 on midbody was spotted in aSiNPs treated group (Fig. 7Ac). Furthermore, the correct clustering of Centralspindlin complex subunits, MKLP1 and CYK4, on antiparallel microtubule and midbody was essential to the subsequent central spindle bundling and contractile ring formation. Then the co-localization of CYK4 and MKLP1 during cytokinesis was examined. As shown in Figure 7B, in the control group CYK4 and MKLP1 co-localized with each other very well on the midbody, while in the aSiNPs treated group lacked CYK4 (Fig. 7Bb) and had on error position of CYK4 and MKLP1 on the midbody (Fig. 7Bc) were found. Above all, aSiNPs not only suppressed the protein content of key cytokinesis regulators, but also disrupted the localization of Aurora B and Centralspindlin complex to the midbody in L-02 cells.
PI3K activator IGF reduced the impaction of aSiNPs on Aurora B phosphorylation and centralspindlin cluster
PI3K activator IGF was introduced, in order to confirm whether the influence of aSiNPs on Centralspindlin and subsequent cytokinesis was through the inhibition of PI3K/Aurora B signaling. As indicated by pre-experiment, 2 ng/mL IGF pretreated L-02 cells for 1 h could activate PI3K/Akt pathway obviously. Thus, the experimental groups were set as: control group, 2 ng/mL IGF pretreated group, 20 μg/mL aSiNPs treated group, and 20 μg/mL aSiNPs plus 2 ng/mL IGF pre-treated group. Result of western blot in Figure 8A indicated the 2 ng/mL IGF could activate PI3K/Akt pathway and alleviate the decrease of Aurora B phosphorylation caused by aSiNPs. Co-immunoprecipitation of MKLP1 and CYK4 was further conducted to examine the relative ratio of CYK4 bound to MKLP1, which could reflect the level of centralspindlin complex formation. As exhibited in Figure 8B, aSiNPs depressed the relative ratio of CYK4 bound to MKLP1 significantly, but this phenomenon was mitigated by the pretreatment of IGF. Meanwhile, the increased rate of multinucleated cells induced by aSiNPs was also reduced owing to IGF (Fig. 8C). Thus, result of this part suggested that aSiNPs mainly inhibited Aurora B activity through PI3K signaling, thereby affecting its regulation of centralspindlin clustering, resulting in centralspindlin complex dysfunction and multinucleated cells formation.
ROS inhibitor NAC reduced the multinucleation effect of aSiNPs
Oxidative stress has been identified as an important toxic mode of aSiNPs. After exposing L-02 cells to different concentrations of aSiNPs for 24h, the intracellular ROS level was detected. Compared with control, ROS level in 20 and 50 μg/ml aSiNPs treated group was significantly alleviated (Fig. 9A).
To further explore the relationship between excessive intracellular ROS and the multinucleation effect of aSiNPs, ROS inhibitor NAC was used to suppress the aSiNPs-induced oxidative damage. The result of pre-experiment confirmed that 5 mM NAC pretreated L-02 cells for 2 h could effectively inhibit the excessive generation of ROS. Thus, the experimental groups were set as: control group, 5 mM NAC pretreated group, 20 μg/mL aSiNPs treated group, and 20 μg/mL aSiNPs plus 5 mM NAC pretreated group. As shown in Figure 9B and 9C, NAC reduced the level of intracellular ROS and the rate of apparent multinucleated cells resulted from aSiNPs, suggesting that aSiNPs might lead to multinucleatation through excessive ROS and oxidative damage.
NAC reduced the impaction of aSiNPs on PI3K 110β/Aurora B pathway and cytokinesis regulatory proteins
To further investigate the specific mechanism by which aSiNPs induce abnormal cytokinesis via oxidative stress, the influence of NAC on the down-regulation of PI3K/Aurora B signaling and related cytokinesis regulatory proteins induced by aSiNPs was investigated. As revealed in Figure 10A and 10B, NAC increased the expression of PI3K 110β and the phosphorylation of Aurora B obviously in the aSiNPs treated group. Similarly, the down regulation of protein content of cytokinesis regulators, including MKLP1, CYK4, Ect2, Cep55, and CHMP2A, as well as Rho activity resulted from aSiNPs were attenuated by NAC involvement. Meanwhile, the effect of NAC on aSiNPs induced abnormal clusters of centralspindlin was assessed by co-immunoprecipitation assay. Result in Figure 10C manifested that NAC improved CYK4 bound to MKLP1 obviously in the aSiNPs treated group. This obtained data indicated that excessive ROS and oxidative damage should be responsible for aSiNPs induced down regulation of the PI3K 110β/Aurora B pathway as well as it's decrease in the content and function of cytokinesis regulatory proteins.