Lysosomal Dysfunction Caused by Cellular Accumulation of Silica Nanoparticles*

Nanoparticles (NPs) are widely used as components of drugs or cosmetics and hold great promise for biomedicine, yet their effects on cell physiology remain poorly understood. Here we demonstrate that clathrin-independent dynamin 2-mediated caveolar uptake of surface-functionalized silica nanoparticles (SiNPs) impairs cell viability due to lysosomal dysfunction. We show that internalized SiNPs accumulate in lysosomes resulting in inhibition of autophagy-mediated protein turnover and impaired degradation of internalized epidermal growth factor, whereas endosomal recycling proceeds unperturbed. This phenotype is caused by perturbed delivery of cargo via autophagosomes and late endosomes to SiNP-filled cathepsin B/L-containing lysosomes rather than elevated lysosomal pH or altered mTOR activity. Given the importance of autophagy and lysosomal protein degradation for cellular proteostasis and clearance of aggregated proteins, these results raise the question of beneficial use of NPs in biomedicine and beyond.

with the biomedical application of NPs at the cellular and organismic levels remain incompletely understood (1). Among the phenotypic changes reported to be associated with the biomedical application of NPs are cellular stress responses (i.e. redox imbalance, oxidative stress), DNA damage, and altered gene expression (8,9). Which of these phenotypes can be considered a direct consequence of cellular NP association or uptake and the underlying molecular mechanisms have remained in many cases unknown.
Upon cellular application, NPs initially interact with the plasma membrane, often followed by their internalization into the cell interior (10 -12) via clathrin-dependent as well as clathrin-independent endocytosis routes (i.e. via caveolae), which may require the membrane-severing GTPase dynamin (13,14). Due to the questionable specificity of many commonly used pharmacological tools toward these pathways (15) the precise mechanisms of cellular uptake of NPs often have remained elusive. After cell entry NPs are delivered to the endolysosomal system (16), where they may accumulate. Lysosomes play essential roles in cell physiology ranging from the degradation of malfunctional or aggregated proteins (e.g. via autophagy) or lipids to nutrient signaling and cellular growth control (17). For example, internalized growth factors such as EGF are sorted to late endosomes (18), which fuse with lysosomes to deliver their intralumenal content for degradation (19). It is thus conceivable that the cellular uptake and accumulation of NPs profoundly impacts on the function of the endolysosomal system and thereby on cell physiology.
Here we have investigated the mechanisms of uptake and the intracellular trafficking itinerary of silica (SiO 2 ) NPs (SiNPs) in human cervix carcinoma (HeLa) cells. We demonstrate that SiNPs are internalized largely via clathrin-independent endocytosis involving dynamin 2-dependent caveolar uptake followed by their targeting to and accumulation within lysosomes. We further show that intralysosomal accumulation of SiNPs severely perturbs autophagy-mediated protein turnover and degradation of internalized epidermal growth factor due to impaired cargo delivery via autophagosomes or late multivesicular endosomes to SiNP-filled lysosomes. Consistent with these results we show that lysosomal SiNP accumulation reduces the metabolic activity of cells although it does not induce cell death due to apoptosis or necrotic cell lysis. Our results raise the question of beneficial use of NPs in biomedicine and beyond.

Results
To address the mechanism of cellular uptake and the physiological consequence of NP accumulation in mammalian cells we prepared SiNPs covalently labeled with FITC by modified microemulsion synthesis (20). The resulting dye-labeled spheres were used as cores for the subsequent seeded growth of a silica shell of 3-nm thickness based on the Stöber method (21). The surface of FITC-labeled SiNPs was modified by amino functionalization with N-(6-aminohexyl)-aminopropyltrimethoxysilane (AHAPS), resulting in the covalent introduction of positively charged amino groups (Scheme 1) that facilitate cellular uptake and counteract aggregation of SiNPs (22). The resulting AHAPS-functionalized SiNPs were homogenously sized spheres with an average diameter of 75 Ϯ 2 nm and a zeta potential of 51 Ϯ 1 mV in ethanol and of 119 Ϯ 2-nm and a zeta potential of 25 Ϯ 1 mV in water (Fig. 1A, Table 1). AHAPSmodified SiNPs showed excellent colloidal stability in ethanol and in cell culture media.
To study the mechanism of SiNP uptake, cultured HeLa human cervix carcinoma cells were incubated for 4 h at 37°C with FITC-labeled SiNPs, washed, and fixed. Analysis by confocal spinning disc microscopy revealed the accumulation of SiNPs in spherical intracellular organelles enriched in the perinuclear area. To unravel the mechanism of SiNP uptake, cells were treated with small interfering (si) RNAs to deplete them of endogenous clathrin heavy chain, a key essential component of clathrin-mediated endocytosis, flotillin 1, an integral membrane protein thought to contribute to clathrin-independent fluid-phase endocytosis via the CLIC/GEEC pathway (23,24), or of caveolin, the main structural component of caveolae that undergoes dynamin-mediated fission (25). RNA interference resulted in the efficient and specific down-regulation of the corresponding target proteins by 85-90% (Fig. 1B). SiNP endocytosis was greatly reduced in HeLa cells depleted of caveolin 1 (using either smart pool siRNAs or a distinct single siRNA) or of dynamin 2. Knockdown of flotillin 1 led a small although statistically significant reduction in SiNP uptake, whereas SiNP endocytosis proceeded unperturbed in the absence of clathrin. Thus, AHAPS-modified SiNPs enter HeLa cells primarily via dynamin 2-mediated caveolar endocytosis (Fig. 1, C and D) independent of clathrin.
Internalized SiNPs displayed a punctuate distribution with some enrichment in the perinuclear area, suggestive of their accumulation within intracellular organelles. Intracellular SiNPs colocalized with CD63 and LAMP1, transmembrane proteins that partition between late endosomes and lysosomes (26), but not with early endosomes containing early endosomal antigen 1 (EEA1) (Fig. 2, A and B). Limited colocalization was observed with the clathrin adaptor AP-1 (assembly protein 1) and with the mannose 6-phosphate receptor (MPR), proteins that cycle between endosomes and the trans-Golgi network (26) (Fig. 2, A and B). No colocalization of internalized SiNPs with the Golgi complex, mitochondria, or the endoplasmic reticulum was detected (data not shown), in agreement with Al-Rawi et al. (27). The accumulation of SiNPs in late endosomes/lysosomes was also observed in living cells incubated with LysoTracker Red (Fig. 2, A and B), an organic heterotricyclic 4-bora-3a,4a-diaza-s-indacene compound that accumulates within the acidic lumen of late endosomes/lysosomes. Ultrastructural analysis of HeLa cells by thin-section electron microscopy confirmed the prominent accumulation of SiNPs within the lumen of lysosomes (Fig. 2C). Lysosomal accumulation of SiNPs was paralleled by decreased metabolic activity as evidenced by the reduced ability of SiNP-treated HeLa cells to reduce the tetrazolium dye MTT (Fig. 3A). However, we did not detect pronounced cell death via apoptosis (Fig. 3B) or necrotic cell membrane permeabilization (Fig. 3C) induced by accumulation of SiNPs. Thus, dynamin 2-mediated caveolar endocytosis of SiNPs into HeLa cells resulted in their accumulation in late endosomes and lysosomes and in reduced metabolic activity of cells (consistent with Refs. 27 and 28).
The high level accumulation of SiNPs within lysosomes conceivably may perturb lysosomal function. Lysosomes are essential eukaryotic organelles that serve as endpoints for the degradation of internalized growth factors such as epidermal growth factor (EGF) (18) and for cytoplasmic material (i.e. aggregated proteins) targeted for autophagy, a pathway intimately linked to cellular growth control (29). Autophagy requires the posttranslational lipidation of cytoplasmic LC3 protein with phosphatidylethanolamine (a form termed LC3-II), resulting in its association with membranes and autophagosome formation. Modified LC3-II then undergoes degradation as autophagosomes fuse with lysosomes to enable degradation (29). Incubation of HeLa cells with increasing concentrations of SiNPs showed a concomitant increase in lipidated LC3-II protein (Fig.  4A) and in the number of LC3-positive autophagosomes (Fig.  4B). A similar cellular accumulation was seen for p62 (ϭSQSTM1) (29), another component of autophagosomes and substrate for autophagy-mediated lysosomal protein turnover (Fig. 4, A and C). As expected, LC3-positive autophagosomes displayed little overlap with LAMP1-containing lysosomes (Fig.  4E). An accumulation of LC3-containing autophagosomes was also seen in HeLa cells treated with iron-oxide NPs coated with polyethylene glycol (PEG) (Fig. 4D), suggesting that accumulation of autophagosomes may be a general phenotype induced by treatment of cells with high doses of NPs.
To further explore the mechanism underlying the observed autophagosome accumulation in SiNP-treated cells, we analyzed the effects of serum starvation, conditions that induce autophagosome formation, and of the vacuolar ATPase inhibi- tor bafilomycin A1, which blocks lysosomal degradation by dissipating the lysosomal proton gradient. Starvation of control cells resulted in increased levels of lipidated LC3-II due to the accumulation of LC3-positive autophagosomes, and this phenotype was exacerbated by concomitant application of bafilomycin A1 to prevent lysosomal turnover of autophagosomes (Fig. 5, A and C). Cells treated with SiNPs displayed high levels of lipidated LC3-II and accumulated LC3-positive autophagosomes under all conditions, largely irrespective of whether cells had been starved and/or treated with bafilomycin A1 (Fig. 5, A and C). Treatment of cells with non-fluorescent, unlabeled SiNPs also caused LC3-II to accumulate (Fig. 5B). These data suggest that lysosomal accumulation of SiNPs within cells is not due to enhanced autophagosome formation but may rather reflect defective autophagic flux.

TABLE 1 Characterization of SiNPs with (؉) and without (؊) FITC in the core
Data are from n ϭ 5 independent measurements (at 25°C, in DMEM (ϩ10% FCS) at 37°C) performed within the first hour after transfer into the corresponding medium. TRM, transmission electron microscopy; DLS, differential light scattering. Ø ϭ diameter; ϭ zeta potential. mTOR signaling, a key regulatory pathway of autophagy, as the cellular accumulation of SiNPs did not alter the levels of phospho-S6K1 or phospho-ULK1 (Fig. 6). We conclude that intralysosomal accumulation of SiNPs leads to the accumulation of p62-containing LC3-positive autophagosomes that apparently fail to undergo further lysosomal degradation as further explored in detail below (see Fig. 9).

Nanoparticle
Although these results indicate that cellular accumulation of SiNPs impairs degradation of cytoplasmic proteins targeted for lysosomal degradation via autophagy, they do not address whether growth factors internalized from the plasma membrane via endosomes can be degraded efficiently in SiNP-filled lysosomes. We, therefore, analyzed the uptake and degradation of Alexa647-labeled EGF. Endocytosis of EGF-Alexa647 into HeLa cells proceeded unaltered irrespective of whether or not these cells had been incubated with SiNPs before EGF addition (Fig. 7C, 30 min). SiNP-treated cells also showed normal surface levels of EGF receptors, and ligand binding to these receptors elicited a signaling response similar to that of control cells (Fig. 7, D and E). However, SiNP-containing HeLa cells displayed a strongly reduced ability to degrade internalized EGF-Alexa647 during the subsequent post-endocytic chase period of 60 or 120 min (Fig. 7C). Instead of being rapidly degraded, a significant fraction of internalized EGF-Alexa647 remained within punctuate structures identified by multicolor confocal microscopy as LAMP1-and CD63-positive late endosomes even after a 120-min chase (Fig. 7, F and G). This effect was specific for the lysosomal system as endocytosis and recycling of transferrin, an iron-carrying protein not degraded via lysosomes, from endosomes to the cell surface proceeded unperturbed (Fig. 7, A and B). Intralysosomal accumulation of SiNPs in HeLa cells, thus, disrupts the lysosomal degradation of internalized EGF and likely of other growth and differentiation factors.
What might be the reason for the failure of SiNP-filled lysosomes to degrade cytoplasmic autophagic substrates or internalized growth factors? Several scenarios can be envisioned. Lysosomes containing AHAPS-modified SiNPs may be dysfunctional because their proton gradient is disrupted due to the accumulation of charge, sequestration of ions required for acidification (i.e. Cl Ϫ ), or loss of membrane integrity. Such changes would eventually result in the inactivity of intralysosomal hydrolases (17). Alternatively, SiNP-filled lysosomes may be unable to receive cargo from upstream donor compartments such as autophagosomes and late endosomes, i.e. due to the inability of their limiting membrane to undergo remodeling required for fusion, resulting in the spatial segregation of lysosomes from their target substrates (19). We tested these possibilities by first analyzing lysosomal pH in living cells. Ratiometric imaging using pH-sensitive Oregon Green488 coupled to 10-kDa dextran (30) revealed an intralysosomal pH of 4.5 Ϯ 0.1 in control cells and of 3.9 Ϯ 0.1 in SiNPloaded cells (Fig. 8, A-C). The slightly lower pH of SiNPloaded lysosomes possibly is a secondary consequence resulting from the failure of these lysosomes to degrade intracellular proteins, which serve as the main lumenal buffer for protons. Consistent with this and with the fact that most lysosomal hydrolases are activated by low pH, we observed elevated cathepsin B and L activities in SiNP-loaded cells (Fig. 8, D-F). These data show that dysfunction of SiNP-filled lysosomes is not the result of impaired lysosomal acidification or inactivity of intralysosomal hydrolases.
We, therefore, finally tested the hypothesis that impaired autophagy and lysosomal degradation result from impaired cargo delivery to lysosomes from upstream donor compartments such as autophagosomes and late endosomes. If this indeed were the case, one would expect that substrates for lysosomal proteolysis such as LC3-II, p62, and internalized EGF accumulate in compartments lacking SiNPs as well as functional lysosomal hydrolases. Multicolor confocal imaging indeed revealed the near complete lack of colocalization between LC3-, p62-, and SiNP-containing lysosomes (e.g. positive for cathepsin B/L), whereas SiNPs displayed a profound overlap with lysosomal cathepsin B and L hydrolase activities (Fig. 9, A and B). To directly test the hypothesis that impaired autophagy and lysosomal degradation result from impaired cargo delivery to lysosomes from upstream compartments, we used tandem RFP-GFP-tagged LC3. Fusion of RFP-GFP-LC3 with lysosomes results in quenching of GFP fluorescence and, thus, has been widely used to monitor autophagic flux and autolysosome formation. Serum starvation induced autolysosome formation in control cells, and this was completely blocked by treatment of cells with the vacuolar ATPase inhibitor bafilomycin A1 as expected (Fig. 9C, no NPs). By contrast, autolysosome formation was significantly impaired in SiNP-treated cells irrespective of whether cells were fed or serum-starved (Fig. 9C, ϩ100 g/ml NPs).

Discussion
We demonstrate using siRNA-mediated cellular depletion of key endocytic proteins that positively charged SiNPs enter cells largely (though perhaps not exclusively) via dynamin 2-dependent caveolar internalization rather than clathrin-mediated endocytosis in contrast to a previous study (12). We also observed a small inhibitory effect of flotillin 1 depletion on SiNP uptake that may reflect a secondary caveolin-independent endocytic pathway or, alternatively, the use of shared components of caveolar-and CLIC/GEEC-mediated endocytosis described recently (31).
The most important conclusion that emanates from our work is the fact that positively charged SiNPs and by extension other SiNPs that accumulate in lysosomes (28), although generally perceived as being well tolerated by mammalian cells, accumulate in lysosomes over extended periods of times (we have followed them for days without loss of signal) causing impairment of lysosomal function, a pathway not addressed in prior studies (for example, see Refs. 27 and 28). Several lines of evidence support this. First, we show that lysosomal accumulation of SiNPs causes the accumulation of LC3-and p62-positive autophagosomes that fail to undergo lysosomal degradation, suggesting that degradation of cytosolic proteins via autophagy is perturbed. Second, we demonstrate that SiNP-treated cells partially fail to degrade exogenous ligands such as growth factors (i.e. EGF), which instead accumulate in late endosomes. By contrast, endosomal recycling proceeds unperturbed, indicating that SiNP-induced alterations within the endosomal system are specific for degradative sorting. Degradation of internalized growth and differentiation factors is an important means of mammalian cells to control cell growth and entry into mitosis (18). Third, we show that failed lysosomal proteolysis of EGF or autophagosomal substrates does not result from altered mTOR signaling, elevated intralysosomal pH, or general damage of lyso-somal membranes. Furthermore, qualitative cathepsin activity assays argue against SiNP-induced impairment of lysosomal protease activity, although quantitative effects of SiNPs on cathepsin trafficking or activation cannot be ruled out completely.
Rather, we favor a model according to which lysosomes containing high levels of SiNPs fail to receive cargo from late endosomes and autophagosomes, i.e. due to defective fusion of these organelles with lysosomes ( Fig. 10) as evidenced by the observed spatial segregation of non-degraded EGF or p62 and lysosomal proteases (Fig. 9, A and B) and the impaired autophagic flux and autolysosome formation (Fig. 9C) in SiNP-treated cells. At the moment we can only speculate why SiNP-containing lysosomes maybe fusion-defective. One possibility is that the rigid structure of SiNPs, perhaps paired with electrostatic interactions of their positively charged surface with negatively charged glycolipids lining the lumenal face of lysosomal membranes (32), may prevent membrane remodeling processes that are required for SNARE-mediated fusion with late endosomes or autophagosomes (33). Consistent with this possibility, it has been observed that membrane fission at the plasma membrane is regulated by the specific lipid environment (34). However, other mechanisms of SiNP action on lysosomal fusion cannot be ruled out.
Given that lysosomal function is key to many aspects of cell physiology, most notably the clearance of degradation-prone (i.e. internalized growth factors), aggregated malfunctional proteins (17,29,35), we hypothesize that the adverse effects of SiNPs on the metabolic activity of cells are a direct consequence of lysosomal dysfunction. Accumulation of NPs in cells, in particular in the brain, apart from the effects on cell metabolism may potentially also affect the clearance of aggregated proteins via autophagy and, thus, favor neurodegenerative disorders such as Morbus Alzheimer, Huntington disease, or Parkinson disease (35). Irrespective of whether some or all of these effects are of clinical rele- vance, our data call for a reassessment of the risk associated with the application of SiNPs in biomedicine.

Experimental Procedures NP Synthesis and Characterization (Transmission Electron Microscopy, Dynamic Scattering, Zeta Potential)
The synthesis and characterization of SiNPs were performed as previously described (22). Super-paramagnetic Fe 2 O 3 -NPs were a kind gift of MagForce AG (Berlin).

Cell Culture and Transfection
HeLa cells were grown in low glucose Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (v/v) fetal bovine serum (FBS), 1% glutamine, and 1% penicillin/streptomycin and cultured at 37°C, 5% CO 2 . SiRNA transfections were done with Oligofectamine (Invitrogen). HeLa cells were transfected with the mRFP-GFP-LC3 plasmid using lipo-fectamine2000 (Invitrogen). After transfection, cells were   treated with siNPs for 24 h. Because of the fact that the GFP signal is quenched in acidic compartments, this probe makes it possible to differentiate between autophagosomes (GFP-positive and RFP-positive or yellow puncta) and autolysosomes (GFP-negative and RFP-positive or red puncta). Twenty cells were analyzed per assay in ImageJ.

Electron Microscopy
Glutaraldehyde-fixed HeLa cells treated with SiNPs were dehydrated and embedded in epoxy resin for subsequent electron microscopic analysis.

Autophagy Assays
Procedure for Fluorescence Microscopy-Cells were incubated with 20 g ml Ϫ1 and 100 g ml Ϫ1 SiNPs for 24 h. 4 h before incubation was stopped, serum-free medium was added together with 100 nM bafilomycin A1 (Sigma) to the corresponding samples. After washing with ice-cold PBS, cells were paraformaldehyde-fixed and processed for immunostaining with LC3. Digitonin (Invitrogen) was used to permeabilize the cells for LC3 staining. Images were taken with a confocal micro- scope LSM710 (Zeiss). Approximately 40 cells were analyzed per assay in ImageJ under identical threshold conditions.

Lysosomal pH Measurements
Lysosomal pH was determined by ratiometric fluorescence imaging using a pH-sensitive Oregon Green 488 dye coupled to a 10-kDa dextran (OGD; Molecular Probes). A standard pulse-chase protocol was used to specifically target the fluorophore to lysosomes. Therefore, cells were grown on MatTek glass-bottom dishes to a confluence of 80%. The PBS-washed cells were treated with 20 g of ml Ϫ1 SiNPs (ϪFITC) for 4 h in DMEM (ϩ10% FBS). After two additional PBS washing steps, 0.5 mg/ml OGD was loaded onto the cells and incubated overnight in DMEM (ϩ10% FBS). The next day cells were washed with PBS, and the pH-sensitive dye was chased into lysosomes for 2 h at 37°C in DMEM (ϩ10% FBS). Control experiments were conducted to prove the localization of Oregon Green to LAMP1-and CD63-positive structures. The fluorophore was excited at a wavelength of 440 and 488 nm, respectively, and ratiometric fluorescence images were acquired with an inverted microscope (Zeiss Axiovert 200 equipped with a 100 ϫ 1.30 NA oil immersion objective) connected to a Polychrom II monochromator (TILL photonics). After passing a 535 Ϯ 20-nm filter, the emitted light was captured with a Sensicam CCD camera (PCO Inc., Kelheim, Germany). For each sample (control and ϩNPs) at least 10 different cells were measured in imaging solution (10 mM glucose, 2 mM CaCl 2 , 1 mM MgCl 2 , 135 mM NaCl, 5 mM KCl, 10 mM HEPES, pH 7.4). Image analysis was performed using a self-programmed Macro for Fiji ImageJ. Therefore, regions of interest (ROIs) were defined as areas above a certain fluorescence threshold in the acquired images at 488-nm excitation to then calculate the ratio of the mean intensity of the 488 and the 440 channel for each ROI. To finally evaluate the pH for each measurement, an in situ pH calibration was conducted at the end of each experiment by treating the cells with isotonic K ϩ -based solutions (5 mM NaCl, 1 mM CaCl 2 , 115 mM KCl, 1.2 mM MgSO 4 , 10 mM glucose, 25 mM of either HEPES, MES, or potassium acetate) ranging in pH from 3.5 to 7.0 and supplemented with 10 M concentrations of both nigericin (Tocris Bioscience) and monensin (Sigma). The resulting fluorescence intensity ratio (488 nm/440 nm) was fitted with a sigmoidal function and used to interpolate the pH value from the experimental ratio data.

Analysis of Cell Viability and Cytotoxicity
MTT Assay to Measure Metabolic Activity-HeLa cells were seeded in 96-well plates at a density of 10,000 cells per well. The next day cells were incubated for 24 h with different concentrations of SiNPs (ϮFITC). Afterward, cells were washed twice with PBS. The addition of MTT (3-[4,5-dimethyltiazol-2-yl] 2,5-diphenyl-tetrazolium bromide), and following procedure was carried out according to manufacturer's instructions (Invitrogen). The absorbance was measured at 562 nm with a Safire2 plate reader (Tecan AG).
TUNEL Apoptosis Assay-Apoptosis was assayed by the TUNEL Apoptosis Detection kit (Invitrogen Click-iTunel AlexaFluor 647 Imaging Assay) according to the provided manual. Briefly, HeLa cells were seeded in 96-well plates. Asynchronously growing cells were treated in the presence or absence of SiNPs at the indicated concentration for 24 h. Cells were washed, fixed with 4% paraformaldehyde, permeabilized in PBS, 0.25% Triton X-100, and processed for TUNEL staining according to the manufacturer's protocol. Cells treated with DNase I for 15 min served as a positive control. Samples were analyzed by epifluorescence microscopy and quantified.
Lactate Dehydrogenase Activity-based Necrosis Assay-Toxicity due to membrane permeabilization (necrosis) was assayed by determination of lactate dehydrogenase activity using the TaKaRa LDH Cytotoxicity Kit. HeLa cells were seeded in 96-well plates. Asynchronously growing cells were treated in the presence or absence of SiNPs at the indicated concentration for 24 h. The supernatant (100 l) was incubated with lactate dehydrogenase assay reagent according to the manufacturer's protocol and measured at 490 nm (plate background absorbance) using Tecan Safire plate reader. Values were normalized to drug/media background value, and toxicity was calculated as a % of a 100% Triton X-100-lysed cell control (100% necrosis).
EGFR Signaling-HeLa cells at a confluence of 90% were incubated with 20 g/ml SiNPs for 4 h, then washed and starved in DMEM for 2 h. The samples were then stimulated for 0 and 30 min with 500 ng/ml unlabeled EGF in DMEM supplemented with 10 g/ml cycloheximide. Subsequently cells were washed and lysed (1% Triton X-100, 20 mM HEPES pH 7.4, 100 mM KCl, 2 mM MgCl 2 , 1 mM PMSF, 0.03% protease inhibitor mixture (Sigma); PMSF and protease inhibitor mixture were added freshly). Phosphatase inhibitor mixture (Sigma) was added for the analysis of phosphorylated proteins. Lysates were spun at 20,000 ϫ g for 10 min to get rid of cell nuclei, and debris and supernatants were then analyzed by SDS/PAGE and immunoblotting.
EGF Degradation Assay-After SiNPs incubation (as described above), cells were washed three times with PBS and then conducted to starvation for 2 h in DMEM followed by incubation with 100 ng/ml EGF-Alexa647 (Molecular Probes) for 30 min on ice. Cells were then washed 2 times with ice-cold PBS and incubated in DMEM for 30, 60, and 120 min at 37°C. Immediately afterward, cells were washed on ice and processed for immunocytochemistry. To evaluate EGFR levels, coverslips were paraformaldehyde fixed right after the 30 min on ice incubation with EGF-Alexa647.
Transferrin Recycling Assay-After incubation with SiNPs, cells were washed 3 times with PBS and then starved for 1 h in DMEM followed by incubation with 20 g/ml Tf-Alexa647 (Molecular Probes) for 30 min on ice. Subsequently cells were washed 2 times with ice-cold PBS and incubated in DMEM for 5, 15, and 60 min at 37°C. Immediately afterward, cells were washed on ice and processed for immunocytochemistry. To evaluate transferrin (Tf) receptor levels, coverslips were para-formaldehyde-fixed right after the 30 min on ice incubation with Tf-Alexa647.
LysoTracker-Cells were grown overnight on coverslips, washed once with PBS, and incubated with 20 g/ml SiNPs in DMEM (ϩ10% FBS) supplemented with 50 nM LysoTracker Red DND-99 (Invitrogen) for 4 h at 37°C. Afterward cells were washed thoroughly with PBS and imaged live in HBSS, 10 mM HEPES at 37°C and 5% CO 2 .
Cathepsin B/L Lysosomal Protease Activity-Cells were seeded on MatTek glass-bottom dishes (MatTek Corp.) and cultured overnight at 37°C and 5% CO 2 . Washed cells were then incubated with 20 g/ml SiNPs in DMEM (ϩ10% FBS) for 4 h, washed thoroughly, and subsequently treated according to the manufacturer's protocol. The total incubation time with Magic Red TM Cathepsin B/L (both from Immunochemistry Technologies) was 60 min at 37°C and 5% CO 2 . Hoechst was added directly to cells 15 min before imaging.
Data Analysis and Statistics-Data are expressed as the means Ϯ S.E. of at least three experiments. The statistical significance was assessed by using the unpaired t test.