Nanoparticle core stability and surface functionalization drive the mTOR signaling pathway in hepatocellular cell lines

Specifically designed and functionalized nanoparticles hold great promise for biomedical applications. Yet, the applicability of nanoparticles is critically predetermined by their surface functionalization and biodegradability. Here we demonstrate that amino-functionalized polystyrene nanoparticles (PS-NH2), but not amino- or hydroxyl-functionalized silica particles, trigger cell death in hepatocellular carcinoma Huh7 cells. Importantly, biodegradability of nanoparticles plays a crucial role in regulation of essential cellular processes. Thus, biodegradable silica nanoparticles having the same shape, size and surface functionalization showed opposite cellular effects in comparison with similar polystyrene nanoparticles. At the molecular level, PS-NH2 obstruct and amino-functionalized silica nanoparticles (Si-NH2) activate the mTOR signalling in Huh7 and HepG2 cells. PS-NH2 induced time-dependent lysosomal destabilization associated with damage of the mitochondrial membrane. Solely in PS-NH2-treated cells, permeabilization of lysosomes preceded cell death. Contrary, Si-NH2 nanoparticles enhanced proliferation of HuH7 and HepG2 cells. Our findings demonstrate complex cellular responses to functionalized nanoparticles and suggest that nanoparticles can be used to control activation of mTOR signaling with subsequent influence on proliferation and viability of HuH7 cells. The data provide fundamental knowledge which could help in developing safe and efficient nano-therapeutics.

After endocytosis, most nanomaterials will eventually accumulate in acidic vesicular organelles, such as endosomes and lysosomes 2,12,13 . The hydrolytic enzymes in these organelles represent a hostile environment for endocytosed nanomaterials causing their degradation. Importantly, malignant and invasive cancer cells strongly depend on properly functioning acidic organelles. In transformed cells, lysosomal stability, trafficking and composition are frequently altered. Cancer cells display lysosome hypertrophy because of increased lysosomal hydrolases secretion which is important for tumor progression. Hypertrophy renders lysosomes fragile by increasing lysosomal membrane permeabilization (LMP) 14,15 . Therefore, targeting lysosomes to trigger lysosomal leakage may be utilized for cancer therapy. Such an approach could be associated with fewer side effects and higher therapeutic efficacy due to evasion of common resistance mechanisms 16 . Moreover, it has been shown that cationic amphiphilic drugs (CADs) selectively kill cancer cells via LMP 17 . Additionally, we and others have shown previously that amino-functionalized NPs can induce lysosomal swelling and result into cancer cell death 12,13,18,19 .
A key kinase controlling cell growth and proliferation under favorable environmental conditions is the mammalian target of rapamycin (mTOR). Membranes limiting acidic lysosomal compartments are important for the activation of mTOR 20,21 . mTOR as well as some of the targets of the mTOR kinase signaling are overexpressed or mutated in cancer, and it is regarded as a promising target for anticancer treatment 20,21 . It is worth noting here, that mTOR inhibitors display favorable pharmacological profiles and are well tolerated comparing to conventional anticancer therapy 22 .
Recent research demonstrated that various NPs modulate the activation of mTOR and even may result into cell cycle arrest in leukemia cells 19,[23][24][25] . More specifically, amino-functionalized NPs have been shown to inhibit mTOR activity and proliferation in three leukemia cell lines 19 . However, current knowledge of the physiological, pathophysiological effects of NPs on liver cells remain unclear. mTOR is frequently up-regulated in cancer including hepatocellular carcinoma (HCC) and its upregulation is associated with bad prognosis, poor tumor differentiation and earlier recurrence 26 . Therefore, in the present study we investigated NPs of different core composition functionalized either with amino or hydroxyl groups as a platform for targeting lysosomes and mTOR signaling in liver derived cell lines Huh7 and HepG2.
In order to adress this question, we used polystyrene (PS) and silica NPs as model particles for our experiments. PS NPs possess many advantages, including straightforward synthesis in a wide range of sizes, relatively low costs, easy separation, and easy surface modification 27 . Furthermore, PS NPs were recognized as biologically inert, almost non-biodegradable and fully biocompatible [28][29][30][31] . These properties make PS NPs a perfect tool for studying biomedical consequences of various surface modifications 27 . Contrary to PS, silica NPs represent an excellent model of biodegradable biocompatible type of NPs suitable as a potential drug delivery system [32][33][34] . Therefore, we used those properties to study the effects of surface functionality on the NP cell interaction. A detailed understanding of the impact of the surface functionalization is a prerequisite for the rational design of nanomaterials targeting distinct cell types for diagnostic or therapeutic purposes. Additionally, we investigated the role of a protein corona in these effects, taking into account that the protein corona can give rise to undesirable adverse effects 35 .

Results
Characterization of BSA and RNase interaction with functionalized nanoparticles. Generally, when administrated intravenously NPs become coated with proteins and other biomolecules that form a so-called protein corona 36 . It is widely accepted that NPs are internalized into the cell and traffic along defined pathways such as the endo-lysosomal pathway. It has also been shown that specific proteins present in the original protein corona are retained on NPs until they accumulate in lysosomes 37 . Consequently, these retained proteins may play an important role in determining subsequent cellular processing 36,37 . Therefore, it is of great importance to determine how chemical surface modification of NPs and subsequent NP-protein interactions affect biological responses. In our study we utilized three types of NPs as a platform allowing to gain insight into biomedical relevant key determinants of surface functionalization of NPs. The physicochemical properties of the NPs were investigated by dynamic light scattering (summarized in Fig. 1A). All three types of particles had the same mean hydrodynamic diameter of about 30 nm. The particles were functionalized either with hydroxyl or amino groups as reflected by high negative and positive zeta potential, respectively (Fig. 1A). Similar physicochemical characteristics of the particles and the low polydispersity index (PDI) allowed us to analyze the role of the surface charge on the protein corona formation and subsequent pathophysiological responses of cancer cells (Fig. 2B). Previously, we and others showed that amino-functionalized polystyrene NPs (PS-NH 2 ) may result into lysosomal rupture associated with apoptotic cell death 12,18,19,38,39 . Therefore, in this study we addressed a question whether other types of amino-functionalized NP may lead to the same functional consequences. We selected amino-functionalized silica NPs (Si-NH 2 ) as an alternative to PS-NH 2 . Among inorganic-based materials, silica NPs have attracted much research attention for their potential application in nanomedicine, especially as suitable drug delivery system [32][33][34] . We did not utilize as an alternative for non-biodegradable NPs metallic inorganic particle (e.g. gold) due to the fact that in contrast to polystyrene gold NPs are not inert 40 . Additionally, gold or other metal NPs release ions from the NPs 41 . Indeed, intracellular gold ions are known to strongly inhibit the enzymatic activity of cells, leading to mitochondrial membrane depolarization and/or inactivation of mitochondrial enzymes 42 . Therefore, such NPs cannot represent a reliable model to study NP surface functionalization-related cellular effects.
As model proteins we selected serum albumin (a major soluble constituent of human blood plasma and thus a relevant target for studies of nanoparticle-protein interactions 36,43,44 ) and ribonuclease A (a protein the serum levels of which are elevated in pancreatic carcinoma patients 45 ). Our intention was to compare abundant protein in serum at normal conditions with the protein that is elevated in cancer pathology.
Dispersion of silicon dioxide NPs was titrated with fluorescently labeled proteins albumin (BSA) or ribonuclease A (RNase). Their adsorption on the NP surface was monitored using fluorescence correlation spectroscopy SCIEntIfIC REPORTs | 7: 16049 | DOI:10.1038/s41598-017-16447-6 (FCS), which allowed determination of particle size and concentration 46 . To modify the electric charge of the NPs, they were functionalized either with either hydroxyl or amino groups (Fig. 1A). Examples of the obtained autocorrelation curves are shown in Fig. 1C. Their amplitudes were normalized for easier comparison. After mixing proteins with NPs the curves shifted to the right, which is a result of the slow-down of the diffusion of protein molecules upon binding to the NPs. While titrating the NPs with fluorescent proteins the mean number of the diffusion entities increased (Fig. 1D). These entities included protein-coated NPs, the number of which was constant, as well as free protein molecules, the number of which increased during the titration. While the number of particles gave us a control over the protein concentration, a better marker of the adsorption was a count per particle -the mean brightness of the diffusing particles (i.e. fluorescence intensity divided by the number of particles) (Fig. 1E). At low protein concentrations most of the protein molecules were adsorbed and diffused together with the NPs, therefore the mean particle brightness was relatively high. At higher protein concentrations the adsorption sites saturated and the fraction of free protein molecules increased. Since the brightness of a free protein was lower than the brightness of the NP with many adsorbed proteins, the mean particle brightness decreased. Above 20 nM of protein count per particle leveled out. The mean diffusion times of the free proteins (0.25 ms for RNase and 0.47 ms for BSA) depicted as triangles in Fig. 1F were measured in a separate experiments and fixed during the fitting of the data from the NP titration. The second diffusion time corresponded to the diffusion of the NPs with adsorbed proteins. It increased slightly with increasing protein concentration, which reflected slight tendency of the coated NPs to aggregate. Taking together, these data indicate that saturation in protein corona formation was reached at physiological concentrations of both proteins in human blood plasma 36,[43][44][45] . Formation of protein corona was very rapid in time reaching an equilibrium within one hour. These results provided a platform for protocol optimization for experiments with cell treatment.
Amino-functionalized polystyrene NPs inhibit whereas amino-functionalized silica NPs enhance proliferation of hepatocellular carcinoma cells. In our previous works we analyzed the mechanisms of functionalized NP uptake by different cell lines as well as cytotoxic potential of amino-functionalized NPs 13,38 . Moreover, we and others have shown that positively charged NPs may passively target tumor cells even in vivo 8,13,47 . Numerous studies have shown that functionalization of NPs with amino groups improves passive tumor targeting via EPR effect 19,[48][49][50][51] . Therefore, based on these data in our experiments, we studied interaction of amino-functionalized NPs with cells as a tentative model for passive tumor targeting. PS and silica NPs served solely as model particles providing a convenient nanosized platform to study biomedical consequences of surface modifications and functionalization.
NPs surface-functionalized with amino groups can induce lysosomal swelling 12,18 . We and others have shown that unsaturated amino groups on the surface of the NP are capable of sequestering protons in the lysosomes leading to activation of a proton pump v-ATPase and retention of water. This so-called "proton sponge effect" , amino-functionalized (Si-NH 2 ) silica, or amino-functionalized (PS-NH 2 ) PS NPs. Cell viability was assessed by the WST-1 assay. The data were normalized to control values (no particle exposure) and expressed as mean ± SEM, n = 3 each. (B) Analysis of cytotoxicity in Huh7 cultured with Si-OH, Si-NH 2 , or PS-NH 2 NPs bearing RNase as hard protein corona. Cells were cultured in the presence or absence of Si-OH or Si-NH 2 NPs (all 100 µg/ml) pre-incubated with increasing concentrations of RNase for 1 h. Cell viability was assessed by the WST-1 assay. The data were normalized to control values (no particle exposure) and expressed as mean ± SEM, n = 3 each. (C) Analysis of cytotoxicity in Huh7 cultured with Si-OH or Si-NH 2 , NPs bearing BSA as hard protein corona. Cells were cultured in the presence or absence of Si-OH or Si-NH 2 NPs (all 100 µg/ml) pre-incubated with increasing concentrations of BSA for 1 h. Cell viability was assessed by the WST-1 assay. The data were normalized to control values (no particle exposure) and expressed as mean ± SEM, n = 3 each. (D) Comparison of proliferative activity of Huh7 cultured with Si-OH, Si-NH 2 , or PS-NH 2 NPs bearing BSA or RNase (100 µM both) as hard protein corona or bare NPs. Cell were treated as in (A-C). The data were normalized to control values (no particle exposure) and expressed as mean ± SEM, n = 3 each.
Consistent with previous findings 12,18,19,38,39 , PS-NH 2 but not Si-OH induced cell death of Huh7 cancer cells ( Fig. 2A). Unexpectedly, Si-NH 2 significantly enhanced Huh7 proliferation rate ( Fig. 2A). Furthermore, Huh7 cell treatment with Si-NH 2 bearing hard protein corona (formed either by BSA or RNase) resulted in higher proliferation rate in comparison with bare Si-NH 2 treatment ( Fig. 2B-D). Importantly, cell treatment with the same concentrations used for protein corona formation of either BSA or RNase did not affect proliferation ( Fig. 2B-D). Similar proliferation rate was observed in cells treated with either bare or bearing hard protein corona Si-OH NPs ( Fig. 2A-D).
It is worth noting here that the adverse effects triggered by NPs were shown to be significantly ameliorated upon formation of the protein corona 37,52,53 . Consistently with these results, we found that hard protein corona had protective effects against acute toxicity induced by PS-NH 2 NPs ( Fig. 2D and Fig. S1 in Supporting Information).

Conformational changes of adsorbed BSA and RNase proteins on functionalized nanoparticles.
It has been shown that changes in protein structure and function occur as a result of their interaction with the NP surface 12 . Thus, we hypothesized that such potential changes in protein structure upon interaction with NP may be responsible for either stimulation or inhibition of Huh7 cells proliferation. Moreover, it has been showed that activity of RNase A is linked with its ability to modulate cell proliferation 54 .
Firstly, we checked whether proteins change their conformation after adsorption to the NP surface. In order to do this, we labelled BSA and RNase with the bromoacetamidomethylproxyl spin label (SL) that reacts selectively with His residues. As a consequence, in both cases, selective spin labeling of His residues provides a tool for assessing the interaction between the NP surface and specific patches on the protein surface 55 . EPR spectra of BSA and RNase adsorbed on either Si-OH or Si-NH 2 NPs suggest an important reduction in the mobility of the spin label of both proteins on both types of NPs (Fig. 3A,B). This means that, in both cases for both proteins, the SL-containing protein domain is directly involved in the interaction with the NP surface. Furthermore, there was a substantial shift of EPR spectra of both proteins adsorbed on either Si-OH or Si-NH 2 NPs (Fig. 3A,B). Taking together, spin labelled BSA and RNase adsorbed on either Si-OH or Si-NH 2 NPs show distinct spectral patterns arising from distinct local environment of the SL. In both cases, some rearrangements of the adsorbed molecules are likely to occur as well on both types of nanoparticle.
Further, we checked whether such conformational rearrangements of RNase adsorbed on the NPs could lead to changes in RNase activity. Indeed, adsorption of RNase on either Si-OH or Si-NH 2 NPs resulted in significant activity changes as measured using specific fluorescent substrate that emits fluorescence upon cleavage by RNase (Fig. 3C,D). Moreover, RNase adsorption on Si-OH NPs significantly increased enzymatic activity (Fig. 3D). In contrast, RNase adsorbed on Si-NH 2 NPs showed decreased enzymatic activity (Fig. 3D). This observation is not surprising since it is known that RNase changes its activity differently during adsorption onto either "positively" or "negatively" charged surfaces 56 . This happens due to the fact that RNase molecules orient differently on "positively" or "negatively" charged surfaces 56 . However, even though RNase adsorption on Si-OH and Si-NH 2 NPs changed the enzymatic activity significantly (Fig. 3D), the changes were too low to have any noticeable impact on cellular proliferation. It has been shown that RNase enzymatic activity should change order of magnitude to affect cellular proliferation 57,58 . Furthermore, such changes in enzymatic activity are accompanied with formation of RNase A aggregates (dimers, trimers, tetramers, pentamers and hexamers) 57,58 . RNase adsorption on either Si-OH or Si-NH 2 NPs did not increase any aggregates formation (Fig. S2 in Supporting Information). Therefore, we focused our research on other possible molecular mechanisms that could explain differential effects of Si-NH 2 and PS-NH 2 NPs on cellular viability.

PS-NH 2 but not Si-NH 2 nanoparticles induce lysosomal permeabilization.
To understand the mechanism of the inhibition of cell proliferation, we, firstly, have analyzed the subcellular localization of NPs bearing protein corona by confocal microscopy. Both proteins (BSA and RNase) as well as particles with hard protein corona (formed by either BSA or RNase) colocalized with acidic vesicular organelles which we visualized using LysoTracker, a dye targeted to compartments with low internal pH such as lysosomes (Fig. 4). Further, we analyzed uptake efficiency of both proteins and particles with hard protein corona (formed by either BSA or RNase) by Huh7 cells. Both types of proteins and NPs with hard protein corona (formed by either BSA or RNase) were taken up by the cells, although to a different extent (Fig. 5A). Huh7 internalized bare BSA or RNase to a lowest extent (Fig. 5A). Cells ingested about the same amounts of PS-NH 2 or Si-NH 2 bearing either BSA or RNase as hard protein corona (Fig. 5A). Interestingly, Huh7 took up significantly less Si-OH NPs (bearing either BSA or RNase as protein corona) than corresponding PS-NH 2 or Si-NH 2 (Fig. 5A). These data are in line with previously published research that shows that positively charged NPs may passively target tumor cells, even in vivo 8,13,47 .
To address the possible lysosomal leakage induced by amino-functionalized NPs, we used the lysosomotropic dye acridine orange (AO). AO uptake in lysosomes leads to red fluorescence, which decreases when the dye is Cells were incubated with BSA or RNase (both 50 µM) labelled with Atto633 (red dye) for 1 h. Additionally, cells were incubated with either Si-OH or Si-NH 2 or PS-NH2 NPs (all 50 µg/ml) bearing either BSA or RNase as hard protein corona (both 50 µM). Acidic organelles were stained with LysoTracker probe (Invitrogen, green dye), and the cells were analyzed using confocal microscopy. Colocalization is yellow.
leaking from this compartment into the cytosol. After treatment with NPs, the cells were stained with AO and analyzed by fluorescence microplate reader. Treatment with PS-NH 2 , but not Si-OH or Si-NH 2 NPs, induced a significant lysosomal permeabilization (Fig. 5B). It has been shown repeatedly that amino-functionalized NPs induce lysosomal damage due to the so-called "proton sponge" effect 12,13,18,19 . The postulated mechanism of the "proton sponge effect" involves sequestration of protons by the amines on the particle surface thereby keeping the proton pump working, which leads to the retention of one Cl − ion and one water molecule for each proton that enters the lysosome. This process leads to lysosomal swelling and to osmotic rupture 12 . Finally, this process will result in leakage of lysosomal enzymes, such as cathepsin B 12,13,18,19 . When cathepsin B comes in contact with mitochondrial membranes, it causes mitochondrial dysfunction 59,60 . To investigate whether polystyrene particles can perturb mitochondrial function, we used the fluorescent dye JC-1. JC-1 is a lipophilic, cationic dye that can selectively enter into mitochondria and reversibly change color from green to red as the membrane potential increases. In healthy cells with high mitochondrial ΔmΦ, JC-1 spontaneously forms aggregates with intense red fluorescence. On the other hand, in apoptotic or unhealthy cells with low ΔmΦ, JC-1 remains in the monomeric form, which shows only green fluorescence. As expected, only PS-NH 2 induced depolarization of the mitochondrial membrane as indicated by a decrease of the red-to-green fluorescence intensity ratio (Fig. 5C). In contrast, Si-NH 2 and Si-OH NPs did not induce any significant changes in the mitochondrial membrane potential when compared to controls (Fig. 5C).
Interestingly, PS-NH 2 bearing either BSA or RNase as hard protein corona showed significantly lower ability to induce lysosomal leakage (Fig. 5D). Formation of protein corona onto PS-NH 2 NPs did not affect The data expressed as mean ± SEM, n = 3 each. *p < 0.05, **p < 0.01. Comparison of lysosomal integrity (D) and mitochondrial potential (E) of Huh7 cultured with Si-OH, Si-NH2, or PS-NH2 NPs bearing BSA or RNase as hard protein corona or bare NPs. Cell were treated as in (B,C). The data expressed as mean ± SEM, n = 3 each. *p < 0.05, **p < 0.01. (F) Degradation of functionalized silica NPs in acidic conditions. Si-OH, Si-NH 2 or PS-NH2 NPs (all 100 µg/ml) were incubated in either a buffer simulating extracellular conditions with neutral pH 7.4 or mimicking lysosomes with pH 4.0 for 2.5 h. After incubation hydrodynamic diameter of NPs was characterized by dynamic light scattering (DLS) using a Zetasizer Nano. The data expressed as mean ± SEM, n = 3 each. **p < 0.01. mitochondrial function (Fig. 5E). These data are perfectly in line with previously published studies showing that the adverse effects triggered by NPs (including cytotoxicity) are significantly mitigated in the presence of the protein corona 37,52,53,61 . Perturbation of mTOR activity by nanoparticles. So far biological effects obtained with PS-NH 2 NPs were in line with previously published studies and supported "proton sponge" effect as a major cause of cytotoxicity 12,13,18,19 . However, we could not explain why Si-NH 2 (NPs that have approximately the same charge and surface functionalization as PS-NH 2 ) did not induce lysosomal leakage and subsequent cells death. Thus, we performed deep literature search for possible causes of the observed effects. Actually, PS NPs are hardly biodegradable 13,18,23,38,62,63 , whereas silicon based NPs exhibit considerable degree of biodegradability 32,34,64 . Importantly, it has been postulated that for biodegradable particles, the particle composition and degradation products might influence their biological effects 47,65,66 . These data and our results led us to a hypothesis that different biodegradability of PS-NH 2 and Si-NH 2 NPs could be a reason for opposite biological effects observed in our study.
Firstly, we checked stability of NPs in buffers representative for the extracellular space and lysosomal compartments. All three types of NPs showed stability in a buffer simulating extracellular conditions with neutral pH 7.4 (Fig. 5F). Incubation of PS-NH 2 in buffer mimicking lysosomal environment with pH 4.0 did not result into particle core degradation (Fig. 5F). In contrast, Si-OH and Si-NH 2 NPs incubated in acidic conditions showed significant particle core degradation, unlike in neutral environment (Fig. 5F).
Membranes of acidic organelles are important sites for activation of mTOR, a serine/threonine protein kinase that controls cell growth and cell proliferation, as well as cell motility and survival through regulation of protein synthesis and transcription 20,67 . It is well known that amino acids activate mTOR Complex 1 leading to cellular growth through increased ribosome biogenesis and elevated rates of protein synthesis, while suppressing autophagy 68,69 . Active mTOR within the mTOR complex is phosphorylated on serine 2448 70 . It has been shown that polystyrene NPs surface-functionalized with amino groups, but not those with carboxyl groups, obstruct the mTOR signaling in leukemia cells 19 . Importantly, mTOR signaling has a critical role in the pathogenesis of hepatocellular carcinoma (HCC) 26,71 . Therefore, we hypothesized that PS-NH 2 and Si-NH 2 NPs might differently affect this crucial signaling axis due to different stability in acidic organelles. Control Huh7 cells, indeed, exhibited activated mTOR phosphorylated on serine 2448 (Fig. 6A). Treatment with PS-NH 2 , as expected, resulted in the inhibition of mTOR phosphorylation (Fig. 6A). Differently, Si-NH 2 NPs induced the mTOR phosphorylation (Fig. 6A). Moreover, PS-NH 2 bearing either BSA or RNase as hard protein corona showed significant decrease in phosphorylation of the mTOR (Fig. 6A) pointing to activation of mTORC1 signaling axis. Increase in phosphorylation of the mTOR induced by addition of BSA or RNase is not surprising, since mTOR represents a nutrient sensor that is crucially important for licensing cell growth driven by oncogenic PI3K-AKT signaling and for suppressing autophagy 21,72 .
We developed further our hypothesis, that Si-NH 2 NPs degrade in acidic organelles and this degradation is accompanied with mTOR activation. In contrast to the Si-NH 2 NPs, PS-NH 2 NPs were stable in lysosomes and resulted into lysosomal permeabilization leading to mTOR inhibition and subsequent cell death (Fig. 6B).
It is known that amino acids signal to the mTOR complex I growth pathway [73][74][75] . Specifically, stimulation of cells with positively charged amino acids (like arginine) activates mTOR signaling 73 . Furthermore, a putative lysosomal arginine sensor required for arginine to activate mTOR was recently identified 73,76 . Therefore, we thought that degradable Si-NH 2 NPs might work as positively charged amino acids activating mTOR signaling.
First of all, confocal fluorescence microscopy of phosphorylated mTOR revealed that the majority of activated mTOR in cells treated with Si-NH 2 was associated with lysosomal marker LAMP1, whereas in cells treated with PS-NH 2 , the pmTOR staining was weak, diffuse, and did not colocalize with lysosomes (Fig. 6C). Interestingly, BSA alone as well as Si-OH bearing BSA as hard protein corona showed significant increase in phosphorylation of the mTOR (Fig. 6D). However, this phosphorylation increase was not accompanied by colocalization with lysosomal marker LAMP1 (Fig. 6C,E). Contrarily, cells treated with Si-NH 2 exhibited both phosphorylation increase of mTOR and colocalization with lysosomes ( Fig. 6C-E). These data show that the mTOR activation in Si-NH 2 -treated cells occurs at membrane of acidic organelles, whereas these organelles in PS-NH 2 -treated cells are defective and not capable of mTOR activation.
To support further our findings, we re-evaluated the effects of NPs using a second liver derived cell line (HepG2). Consistent with the findings in Huh7 cells, PS-NH 2 but not Si-OH NPs induced cell death of HepG2 cells (Fig. 7A). Moreover, Si-NH 2 NPs significantly enhanced HepG2 proliferation rate (Fig. 7A), and HepG2 cell treatment with Si-NH 2 bearing hard protein corona (formed either by BSA or RNase) resulted in higher proliferation rate in comparison with bare Si-NH 2 treatment (Fig. 7A). All these findings are perfectly in line with the data obtained with Huh7 cells depicted in Fig. 2D. Additionally, treatment with PS-NH 2 , but not Si-OH or Si-NH 2 NPs, induced a significant lysosomal permeabilization (Fig. 7B) which resulted in depolarization of the mitochondrial membrane (Fig. 7C). Further, majority of activated mTOR in HepG2 cells treated with Si-NH 2 was associated with lysosomal marker LAMP1, whereas in cells treated with PS-NH 2 , the pmTOR staining was weak, diffuse, and did not colocalize with lysosomes (Fig. 7D). Also similarly to Huh7 treatment (Fig. 6D), BSA alone as well as Si-OH bearing BSA as a hard protein corona showed significant increase in phosphorylation of the mTOR in HepG2 (Fig. 7E). This phosphorylation increase was not accompanied by colocalization with lysosomal marker LAMP1 (Fig. 7D,F). Contrarily, HepG2 cells treated with Si-NH 2 exhibited both phosphorylation increase of mTOR and colocalization with lysosomes ( Fig. 7D-F). Thus, all data presented in Fig. 7 obtained with the second liver derived cell line HepG2 replicated the findings in Huh7.
Number of studies show that mTOR senses amino acids through the RagA-D (in particular RagC) family of GTPases 75,77,78 . Importantly, RagC makes complex with mTOR that targets the lysosomal surface and this complex formation is necessary for mTOR pathway activation 74 . Thus to check the hypothesis that Si-NH 2 NPs might act as mTOR activator, we performed co-immunoprecipitation assay of RagC complexes. Indeed, only in cells treated Figure 6. mTOR signalling in Huh7 cells exposed to different nanoparticles. (A) Huh7 cultured with Si-OH, Si-NH2, or PS-NH2 NPs (all 50 µg/ml) bearing BSA or RNase (both 50 µM) as hard protein corona or bare NPs for 4 h. Activation of mTOR was analyzed by Western immunoblotting (full blots are presented in Supporting Information). Actin served as a loading control. Representative blots out of 3. (B) Scheme illustrating the mTOR involvement in NP signalling. (C) Activation of mTOR on lysosomal membranes was analyzed in Huh7 stimulated as in (A) and analyzed by confocal microscopy using LAMP1 antibody as marker of lysosomes (red) and pmTOR (green). Colocalization is shown in yellow. Increased cell fluorescence due to mTOR phosphorylation is presented (D) as corrected total cell fluorescence (CTCF). Quantifications performed using ImageJ are presented as means of n = 30 cells. *p < 0.05, **p < 0.01, # p < 0.05, ## p < 0.01. Colocalization analysis of pmTOR and LAMP1 from images (C) is presented in (E). Quantifications performed using ImageJ are presented as means of n = 30 cells. *p < 0.05, **p < 0.01, # p < 0.05, ## p < 0.01. with Si-NH 2 NPs, either bare or bearing hard protein corona, endogenous RagC co-immunoprecipitated with endogenous mTOR (Fig. 8A). Taken together, these data clearly show that Si-NH 2 NPs activate mTOR signaling whereas PS-NH 2 leads to mTOR inhibition. Since, mTOR is regulated by the Ragulator-Rag GTPases complex, which assesses free amino acid content within the lysosomal lumen by measuring the efflux of specific amino acids across the lysosomal membrane 72 , degradable Si-NH 2 NPs likely affect cellular functions in the same way.

Discussion
Previously, we have analyzed the mechanisms of functionalized NP uptake by different cell lines as well as cytotoxic potential of amino-functionalized NPs 13,38 . We and others showed that unsaturated amino groups on the surface of the NP are capable of sequestering protons in the lysosomes leading to activation of a proton pump v-ATPase and retention of water. This so-called "proton sponge effect" might finally result in lysosomal swelling to the point of leakage of the lysosomal content and lysosomal rupture associated with apoptotic cell death 12,18,19,38,39 .
To assess the effect of protein corona and particle functionalization on proliferation of cancer cells, we have used human liver-derived cell lines. It is worth noting here, that the liver cancer is the second most lethal cancer after pancreatic ductal adenocarcinoma in terms of 5-year survival rate 79 . Among all primary liver cancers worldwide hepatocellular carcinoma (HCC) accounts for up to 90%, and represents a major health problem 79,80 . Despite substantial advances in development of new therapies, HCC still possesses serious therapeutic challenge and targeted therapies only provide a modest benefit in terms of overall survival 80,81 .
Summarizing the data obtained in this study, we propose the following model of different NPs action on proliferation liver tumor cells (Fig. 8B). Si-NH 2 are rapidly internalized by the studied cells with subsequent localization to lysosomes (Fig. 8B). Due to inherited instability, Si-NH 2 NPs are rapidly degraded by lysosomal content. This degradation results in RagC-mTOR complex formation and targeting of this complex to the lysosomal surface (Fig. 8B) leading to mTOR activation. Activation of the mTOR pathway contributes to cell proliferation (Fig. 8B). Protein corona represents in this case additional source of nutrition and supports mTOR activation (Fig. 8B).
Different to Si-NH 2 NPs, PS-NH 2 NPs treated cells exhibit proton accumulation in lysosomes associated with lysosomal destabilization and damage of the mitochondrial membrane (Fig. 8B). Moreover, PS-NH 2 NPs inhibit proliferation of liver derived tumor cells. At the molecular level, PS-NH 2 NPs obstruct mTOR signaling (Fig. 8B) leading to cell death.

Conclusions
Using non-biodegradable PS and biodegradable silica NPs as a model we have analyzed if surface functionalization and biodegradability might be used to control important cellular processes. We have found that by means of particle surface functionalization one could affect lysosomal stability and regulate essential cellular processes through control of mTOR kinase activation. Importantly, biodegradability of NPs plays a crucial role in regulation of essential cellular processes. Thus, biodegradable silica NPs having the same shape, size and surface functionalization showed opposite cellular effects in comparison with similar PS NPs. NPs surface-functionalized with amino and having different degree of biodegradability might provide a reliable tool to control the mTOR activation.
Si-OH or Si-NH 2 functionalized ~30 nm silica NPs were purchased from nanoComposix (San Diego, CA). Amino-functionalized 30 nm PS NPs were purchased from Nanocs (New York, NY). The particles were characterized by the medium particle size, polydispersity index (PDI), and zeta potential by using a Zetasizer Nano (Malvern Instruments); the particles were dispersed in PBS, pH 7.4, by sonification before each experiment.
Fluorescence labeling. BSA and RNase were labeled with Atto 633 (Invitrogen) according to the original procedure provided by the manufacturer. In short, 1 mg ATTO 633 N-hydroxysuccinimidyl (NHS)-ester in DMSO was mixed with the 1 mg of protein in PBS buffer at pH 8.3 and incubated at room temperature for an hour. Under this conditions unprotonated ε-amino groups of protein lysine residues reacted with NHS-ester resulting in stable dye-protein conjugate. Free dye was removed using dialysis tube with a MW cutoff of 3500 Da. Colocalization is shown in yellow. Increased cell fluorescence due to mTOR phosphorylation is presented (E) as corrected total cell fluorescence (CTCF). Quantifications performed using ImageJ are presented as means of n = 30 cells. *p < 0.05, **p < 0.01, ## p < 0.01. Colocalization analysis of pmTOR and LAMP1 from images (D) is presented in (F). Quantifications performed using ImageJ are presented as means of n = 30 cells. **p < 0.01.  Cell culture and measurement of cell viability. Cell viability was analyzed by WST-1 assay (Roche, Mannheim, Germany), which is based on the cleavage of tetrazolium salt WST-1 by cellular mitochondrial dehydrogenases, producing a soluble formazan salt. This conversion occurs only in viable cells, thus allowing accurate spectrophotometric quantification of the number of metabolically active cells in the culture. Cells were seeded onto 96-well plates at a density of 20 000 cells per well and treated with different NPs and proteins of indicated concentrations. 24 h after the treatment, WST-1 reagent was added to each dish and incubated for 2 h at 37 °C to form formazan. The absorbance was measured using a Tecan-Spectra ELISA plate reader (Mannedorf, Switzerland) at 450 nm. Readings were done in quadruplicates; three independent experiments were performed for each measurement.

SCIEntIfIC
Spin labeling. BSA and RNase were labeled with the His-specific 3-(2-bromoacetamido-methyl)-proxyl spin label (3-(2-Bromoacetamido)-2,2,5,5-tetramethyl-1-pyrrolidinyloxy), according to the procedure described in 55,83 . Briefly, each protein (0.7 mM) was incubated in 100 mM sodium acetate buffer at pH 5.1 in the presence of a 10-fold molar excess of the spin label. The mixture was stirred for 40 h at 40 °C in the dark; it was subsequently filtered through a Vivaspin20 ultrafilter (cutoff = 3000 Da, Vivascience-Sartorius, Germany) and washed three times with PBS pH 7.4 in order to discard the excess of unreacted spin label. The labeled protein was then incubated in the presence of different concentration of Si-OH or Si-NH 2 NPs. The suspensions were stirred in a thermostatic stirrer at 25 °C for 1 h and then filtered through a Vivaspin20 filter (cutoff = 30000 Da for RNase and cutoff = 100000 Da for BSA, Vivascience-Sartorius, Germany). The sample obtained by resuspension of the silica particles bearing the adsorbed protein in 500 μL of PBS at pH 7.4 was placed in a capillary tube for subsequent EPR measurements. EPR measurements. The EPR spectra were measured using a Bruker X-/Q-band E580 FT/CW ELEXSYS spectrometer. For the measurements the ER 4122 SHQE Super X High-Q cavity with TE011mode was used. The samples were placed into quartz tubes with a diameter of 2 mm. The experimental parameters were: micro-wave frequency 9.8756 GHz, microwave power 1.500 mW, modulation frequency 100 kHz, modulation amplitude 0.2 mT, and the conversion time 60 ms.
RNase activity assay. Different concentration of Si-OH or Si-NH2 NPs were incubated with RNase in PBS pH 7.4 for 1 h. In order to discard the excess of unbound protein, the mixture was filtered through a Vivaspin20 ultrafilter (cutoff = 30000 Da, Vivascience-Sartorius, Germany) and washed three times with PBS pH 7.4.
Subsequently, RNase activity was assessed using Ambion ® RNaseAlert ® Lab Test kit (Thermo Fisher Scientific), according to the manufacturer's instruction. The kit contains a fluorescent substrate that emits a green fluorescence if it is cleaved by RNase, the fluorescence can be visually detected by short-wave UV illumination or measured in a fluorometer. Solutions with active RNase produce a green glow in the assay, whereas solutions without RNase activity do not fluoresce. Following staining, samples were analyzed using a fluorescence microplate reader (Tecan Infinite ® 200 PRO). Readings were done in quadruplicates.
Analysis of nanoparticle and protein uptake. ATTO NHS-esters readily react with amino groups of proteins. Cells were treated with either BSA or RNase labelled with Atto633, or NPs with adsorbed proteins for 1 h. Si-OH, Si-NH 2 or PS-NH 2 NPs were incubated with Atto633-labelled BSA or RNase for 1 h in PBS pH 7.4. The suspensions were stirred in a thermostatic stirrer at 25 °C for 1 h and then filtered through a Vivaspin20 filter (cutoff = 30 000 Da for RNase and cutoff = 100 000 Da for BSA, Vivascience-Sartorius, Germany). The samples obtained by resuspension of the silica or PS particles bearing the adsorbed protein in PBS at pH 7.4 were used to treat cells. The LysoTracker probe (Invitrogen) was used to label lysosomes in cells, as described in the manufacturer's protocol. Cells were monitored on a Bio-Rad MRC-1024 laser scanning confocal microscope (Bio-Rad, Cambridge, MA). Image analysis. ImageJ software (NIH) was used for image processing and fluorescent micrograph quantification. Cellular fluorescence intensity was calculated by normalizing corrected total cell fluorescence (CTCF) of the full area of interest to average a single cell fluorescence. The net average CTCF intensity of a pixel in the region of interest was calculated for each image utilizing a previously described method 4,84 . Colocalization analysis of phosphor-mTOR and LAMP1 was done using colocalization plugin in ImageJ. Quantification of mitochondrial membrane potential. Cells were incubated with NPs for 4 h followed by measuring of mitochondrial membrane potential (ΔmΦ). After 4 h of incubation, cells were loaded with 2 µM JC-1 (Invitrogen), a lipophilic cationic fluorescence dye with a dual emission wavelength for 30 min, in order to analyze the depolarization of the ΔmΦ. At low concentrations (due to low ΔmΦ) JC-1 is predominantly a monomer resulting in a green fluorescence with emission of 530 nm. At high concentrations (due to high ΔmΦ) the dye aggregates, yielding an orange emission of 590 nm. Thus a decrease in the aggregate fluorescent count displays a mitochondrial membrane depolarization whereas an increase exhibits a hyperpolarization. Following the staining, cells were analyzed using a fluorescence microplate reader (Tecan Infinite ® 200 PRO). Readings were done in quadruplicates. Mean and standard deviation of JC-1 aggregate/monomer ratios were plotted for 3 independent experiments for each treatment.

Assessment of lysosomal integrity by
Cell extracts and western immunoblot analysis. Aliquots of whole cell lysates 85,86 containing equal amounts of protein were separated by SDS-PAGE, transferred to PVDF membrane, probed with specific antibodies against phospho-mTOR (Cell Signaling, catalogue no. 2971 S) and detected as described 85,86 . Actin (Thermo Fisher Scientific) staining served as loading control.
Co-immunoprecipitation. After 4 h post NP treatment, cells were lysed with lysis buffer from immunoprecipitation kit (Abcam, catalogue no. ab206996). RagC-mTOR complexes were co-immunoprecipitated from the precleared cell lysates with appropriate Ab as described in the manufacturer's instruction. After pre-clearing with Protein A/G Sepharose ® beads, the lysates were immunoprecipitated with anti-RagC antibody (Cell Signaling, catalogue no. 3360) for 12 hr and washed. The resulting protein complex was eluted from the beads with Laemmli protein sample buffer for SDS-PAGE (Bio-Rad) and resolved on SDS-PAGE with specific antibody against mTOR (Cell Signaling, catalogue no. 4517).

Statistical analysis.
Quantitative results are expressed as mean ± SEM or SD. The statistical significance of differences between the groups were determined using ANOVA Fisher's LSD and Newman-Keuls tests. Differences were considered statistically significant at *p < 0.05.
Experiments utilizing multi-well microtitre plates (e.g. cell viability, RNase activity, lysosomal integrity, mitochondrial membrane potential) were conducted in accordance with guideline on randomization, spatial arrangement of samples and sampling number 87 . Readings were done in quadruplicates. Data were plotted for 3 independent experiments for each treatment.
For quantitative fluorescence microscopy analysis (uptake of nanoparticles, mTOR colocalization and phosphorylation) we used rigorously defined guidelines for accuracy and precision quantification 88,89 . The sample size determination was based on a statistical method described in 90 , which determines sample size for 95% confidence interval and 0.9 statistical power equal to 30. Therefore, n = 30 cells were used in quantification. Furthermore, to meet acceptable standards of data presentation 91 , quantitative fluorescence microscopy data were shown by displaying the full dataset as scatterplots.