Scavenger Receptor Mediated Endocytosis of Silver Nanoparticles into J774A.1 Macrophages is Heterogeneous

Abstract

We investigated the scavenger receptor mediated uptake and subsequent intracellular spatial distribution and clustering of 57.7 ± 6.9 nm diameter silver nanoparticles (zeta-potential = −28.4 mV) in the murine macrophage cell line J774A.1 through colorimetric imaging. The NPs exhibited an overall red-shift of the plasmon resonance wavelength in the cell ensemble as function of time and concentration, indicative of intracellular NP agglomeration. A detailed analysis of the NP clustering in individual cells revealed a strong phenotypic variability in the intracellular NP organization on the single cell level. Throughout the observation time of 24h cells containing non- or low-agglomerated NPs with a characteristic blue color coexisted with cells containing NPs with varying degrees of agglomeration, as evinced by distinct spectral shifts of their resonance wavelengths. Pharmacological inhibition studies indicated that the observed differences in intracellular NP organization resulted from coexisting actin- and clathrin-dependent endocytosis mechanisms in the macrophage population. Correlation of intracellular NP clustering with macrophage maturity marker (F4/80, CD14) expression revealed that differentiated J774A.1 cells preferentially contained compact NP agglomerates, whereas monocyte-like macrophages contained non-agglomerated NPs.

Nanotechnology seeks to utilize the unique physico-chemical properties of materials on the 1–100 nm length scale in a diverse range of applications. Nanoparticles (NPs) are important building blocks of this emerging technology and although the number of their applications in consumer products is rapidly increasing the current understanding of potential risks associated with the release of these materials in the environment for human and animal health is still insufficient.^1–4 This knowledge gap motivates fundamental studies into the intricate interactions between cellular systems and NPs. In this work we focus on the scavenger receptor mediated uptake and subsequent trafficking of silver NPs in the adherent mouse macrophage cell line J774A.1. Macrophages belong to the mononuclear phagocyte system (MPS),⁵ which plays a prominent role in the clearing of environmental microparticles or microorganisms from the body. Macrophages are equipped with a broad range of pattern recognition receptors that enable the cell to recognize unopsinized objects with specific surface features, which the body commonly associates with pathogens or debris.⁶ Since macrophages are “first responders” to all foreign materials, including NPs, mature in vitro murine macrophage cell lines, such as J774A.1 are commonly used model systems for studying various aspects of NP uptake and the potentially related toxicity under well defined environmental conditions.^7–16

Scavenger receptors are a sub-group of pattern recognition receptors that recognize negative surfaces and are known to trigger the internalization of micron-sized exogenous and endogenous objects, such as apoptotic cells, bacteria, and dust particles through phagocytosis.¹⁷ There is increasing experimental evidence that scavenger receptors also recognize nanoscale objects, including engineered NPs.^18–20 In fact, clathrin dependent, scavenger receptor A mediated endocytosis was found to be the major uptake mechanism for carboxydextran coated (and thus negatively charged) iron oxide NPs with diameters of 20 and 60 nm in human macrophages.²¹ Scavenger receptor mediated NP uptake comprises three fundamental stages: i.) NP binding, ii.) NP internalization, and iii.) intracellular transport and redistribution. Together these processes result in a spatial translocation of NPs from the extracellular space through the membrane towards special processing centers within the cell. Endocytosis is, however, not a single defined mechanism, instead multiple uptake mechanisms coexist and are coupled to various vesicle trafficking mechanisms.^{22, 23} Given this complexity, it is hardly surprising that many aspects of the endocytosis of metal NPs into macrophages,^{7, 16, 21, 24, 25} including the exact nature of the underlying cellular mechanisms and their dependence on the size, shape, and surface properties of the NPs, as well as the intracellular fate of the non-degradable NPs and their impact on the cell physiology, still pose important questions. Cell-nanoparticle interactions in biological media are complex since proteins, enzymes, or various small molecules can non-specifically adsorb to the NP surface to form a corona.^{26, 27} The latter can modify the intracellular uptake and distribution of the NPs.

We focus in this work on the scavenger receptor mediated endocytosis of silver NPs since these particles are a cause for acute concern due to their already widespread use as antimicrobial agents in consumer products and medical devices.^28–31 Furthermore, silver NPs have superb photophysical properties and large optical scattering cross-sections in the blue,^32–34 which makes the observation of their fate amenable in a conventional optical microscope.³⁴ The optical response of silver NPs is dominated by resonant electron density oscillations in the particles, which are referred to as plasmons.³⁵ The distance-dependent electromagnetic coupling^{36, 37} between plasmons in individual silver NPs provides a unique approach for monitoring the association of NPs into clusters during NP internalization and subsequent trafficking in real time.^38–43 If two particles approach each other to within approximately one particle diameter, their plasmons hybridize⁴⁴ and the coupled resonance red-shifts with decreasing interparticle separation.³⁶ If additional NPs attach to the dimer to form a larger cluster, the spectra further redshift and broaden with growing number of NPs in the cluster. The ability to track NPs within cells and to simultaneously determine their association levels using plasmon coupling microscopy^{38–40, 45} provides exciting new opportunities for monitoring the spatio-temporal distribution of NPs in living cells. For J774A.1 cells our studies show that NP uptake leads to an agglomeration of the NPs on the ensemble level, but that the degree of NP clustering varies significantly between individual macrophages due to different NP endocytosis and processing strategies within the cell population.

Results and Discussion

Verification of Scavenger Receptor Mediated Endocytosis

The silver NPs used in this work had a diameter of 57.7 ± 6.9 nm and were stabilized by assembling a monolayer of short polyethylene glycols (PEGs) functionalized with thiol and carboxylic acid residues at opposing ends onto the silver surface (see Methods). The resulting NPs had a zeta-potential of −28.4 mV and were stable against agglomeration in Hanks buffer containing Mg²⁺ and Ca²⁺ (HBSS). J774A.1 macrophages adherent to a glass slide were exposed to solutions of silver NPs of varying loading concentrations in HBSS buffer for 5 mins at 37°C. Subsequently, the NP solution was removed and the cells were washed with copious amounts of buffer and then cultured in a CO₂ incubator at 37°C for defined times before inspection. We limited the incubation times of the NPs with the cells i.) to minimize a contamination of the NP surface with opsonins or other ligands⁴⁶ which could trigger NP uptake through specific receptors (other than scavenger receptors) and ii.) to track NP uptake and intracellular trafficking along a defined synchronized time axis in a process that is not complicated by further continuous uptake of additional NPs.

In a first control experiment, We verified that silver NPs are taken up by J774A.1 macrophages and that the biding and uptake are scavenger receptor mediated. In Figure 1A we show a darkfield image of macrophages acquired 0.5h after the initial exposure to a solution of 4.5 µg/mL silver NPs. The intense blue-green color indicates an efficient binding and subsequent uptake of NPs into the macrophage. The transmission electron microscope (TEM) image of a representative macrophage section obtained under these conditions in Figure 1B confirms NPs and NP clusters located in the cytoplasm of the macrophage. Pre-incubation of the cells with the strongly negatively charged polymer polyinosinic acid (poly-I) almost completely inhibited NP binding (Figure 1C), whereas polycytidylic acid (poly-C) did not have a similar effect (Figure 1D). The selective inhibition of NP binding through poly-I is a characteristic sign of scavenger receptor mediated NP binding.⁴⁷

Figure 1.

Silver NP uptake into J774A.1 macrophages is scavenger receptor mediated. A. Darkfield image of a cluster of macrophages exposed to a 4.5 µg/mL solution of 57.7 nm Ag NP in HBSS for 5 mins and then incubated in growth medium for 0.5h. B. TEM image of a macrophage section which show clearly that the NPs are uptake into the cell interior. C. Preincubation of the cells with 100 µg/mL poly-I blocks NP uptake. The macrophages were incubated with poly-I and then exposed to silver NP under identical conditions as in A. D. Preincubation of the cells with 100 µg/mL poly-C does not influence NP uptake.

Time and Concentration Dependence of NP Uptake and Intracellular Clustering

Noble metal NPs are a unique “cargo” for uptake studies since their scattering spectra encode information about their association state, enabling the tracking of their clustering state as function of location and time through optical microscopy.^{38, 41, 45} This is highly relevant for elucidating the mechanisms underlying receptor mediated endocytosis of NPs since some studies have shown that the uptake is strongly size dependent and can require a clustering of NPs on the plasma membrane.^48–54 Furthermore, scavenger receptor mediated endocytosis involves the incorporation of NPs into vesicles at the plasma membrane. These vesicles are expected to fuse with other membranes and endosomal compartments during their subsequent intracellular trafficking, which can result in a local enrichment of NPs in a confined space and, thus, induce agglomeration.⁵⁵

In a first set of experiments we investigated the time and concentration dependence of NP clustering during scavenger receptor mediated endocytosis. We exposed the macrophages to three different concentrations of silver NPs in Hanks buffer (c₁ = 1.2 µg/mL, c₂ = 4.5 µg/mL, c₃ = 8.7 µg/mL) and then monitored for spectral shifts of the NPs at t = 1h, 4h, 8h, 24h by analyzing the intensity distribution of the scattered light on three wavelength channels (λ₁ = 490 ± 20 nm, λ₂ = 560 ± 14 nm, λ₃ = 650 ± 10 nm). We used a total intensity threshold to identify the NP containing areas in the recorded cell image and then determined the contribution from λ₁ – λ₃ in the identified image areas (see Methods). For each concentration and time point we evaluated at least 120 individual cells. Figure 2 summarizes the resulting ensemble-averaged intensity distributions as function of time for the investigated NP concentrations.

Figure 2.

The plasmon resonance wavelength of silver NPs red-shifts as function of NP concentration and time. Relative intensities of the light detected from silver nanoparticles on three different wavelength channels (λ₁ = 490 nm (blue); λ₂ = 560 nm (green); λ₃ = 650 nm (red)). At t = 0h the macrophages were exposed to solutions of silver NPs with concentrations of c₁ =1.2 µg/mL, c₂ = 4.5 µg/mL, and c₃ = 8.7 µg/mL for 5 mins. After that, the cells were incubated in growth medium at 37°C. We recorded data points at t = 1h, 4h, 8h, 24h. 120 – 260 cells were evaluated per time point.

For c₁ the λ₁ channel dominates the scattering spectra at t = 1h, but its contribution continuously decreases as function of time, whereas the contribution from λ₂ overall increases. Between t = 8h and t = 24h the intensity in the λ₃ channel also slightly increases. For cells that were exposed to a solution of silver NPs of concentration c₂, λ₁ still represents the major contribution to the NP spectrum at t = 1h but its intensity is lower than for c₁. The decrease in intensity of the λ₁ channel at t = 1h is accompanied by an increase in the scattering intensity on the λ₃ channel. With increasing time, the intensity of λ₁ further decreases, whereas λ₃ increases. The intensity of λ₂ remains nearly constant over time, only between t = 8h and t = 24h we observe a small increase in λ₂. For the highest investigated NP concentration, c₃, the spectra are much more red-shifted than for c₁ and c₂. The intensity in the λ₂ channel is already dominating at t = 1h. With increasing time the intensities for λ₂ and λ₁ decrease, and the intensity for λ₃ increases.

The above analysis reveals a clear red-shift of the average plasmon resonance wavelength as function of time for all three investigated NP concentrations. This finding is consistent with a gradual clustering of the NPs during the uptake and subsequent intracellular trafficking. The comparison of the three panels in Figure 2 also shows that the resonance wavelengths evaluated at a constant time red-shift with increasing concentration of the initial NP loading solution, confirming a concentration dependence of the association process.

Heterogeneity in NP Uptake and Association

Although the ensemble-averaged resonance wavelength shows a general red-shift as function of both NP concentration (c_i) and time, the optical inspection of individual cells one and eight hours after exposure to c_i (Figure 3) reveals significant cell-to-cell differences in the color (and thus association state) of the uptaken NPs. In the case of c₁ we find macrophages that contain vividly blue NPs besides cells containing blue-greenish NPs at both t = 1h and 8h. A detail of special note is that in many cells containing primarily blue NPs, these particles are often confined to diffuse intracellular compartments. For c₂ we observe more frequently cells that contain NPs of green color distributed across the entire cytoplasm at t = 1h, which is indicative of small to medium sized NP agglomerates. These cells co-exist with cells that still contain primarily blue NPs. At t = 8h, the number of cells containing large NP clusters with orange-yellow color has increased, but cells containing only blue NPs are still frequent. If the loading concentration is further increased to c₃ we observe cells with compact orange NP agglomerates already frequently at t = 1h. But as observed for c₁ and c₂, the cells of the ensemble do not show a uniform NP association level, and macrophages containing non- or low-agglomerated NPs with a strong spectral response in the blue are still present. At t = 8h NP agglomeration is advanced in many cells, but significant color differences between individual cells remain. Even cells containing exclusively blue NPs (see inset for (c₃, t= 8h) in Figure 3) can still be identified.

Figure 3.

The NP association of silver NPs in J774A.1 cells is heterogenous. Representative dark-field images of J774A.1 macrophages exposed to silver NP solutions with concentrations of c₁ = 1.2 µg/mL, c₂ = 4.5 µg/mL, c₃ = 8.7 µg/mL for 5 mins. The images were acquired after incubation in a particle-free culture medium for t = 1h and 8h at 37°C. The differences in the color between individual cells indicates systematic variations in NP association levels due to phenotypic variations in NP uptake and or processing. The inset for c₃, t = 8h shows that even after exposure to a high concentration of NPs and incubation for several hours, some cells still contain exclusively NPs with scattering resonances in the blue.

The optical inspection of macrophages after NP exposure indicates a broad range of phenotypic variability in NP uptake and intracellular processing. This conclusion is further corroborated by monitoring NP association in individual macrophages in real time (Figure 4). Whereas the macrophage in Figure 4A does not show any significant NP agglomeration in the monitored time window of 15–45 min, the macrophage in Figure 4B exhibits an increase in yellow-orange scatterers, which confirms a redistribution of NPs within the macrophage.

Figure 4.

Real-time monitoring of NP association reveals differences in NP processing in individual macrophages. While the macrophage in A. does not show a systematic increase of NP agglomeration, an increase in the number of yellow-orange scatterers, indicative of agglomeration, is observed for the macrophage in B. The cells were first incubated with a NP concentration of c₂ = 4.5 µg/mL, cleaned in prewashed HBSS buffer and transferred into the optical microscope where the tracking experiments were performed at 37 °C. The given times stamps specify the incubation time in NP-free medium.

The prolonged existence of cells containing vividly blue NPs in Figure 3 is especially remarkable. To ensure that these NPs were indeed taken up and not just adherent to the cells, we combined plasmon coupling microscopy with conventional fluorescence microscopy to check for colocalization of the NPs inside of the cells with endosomal compartments. We analyzed cells initially exposed to silver NPs of concentration c₃ after incubation for t = 24h. Figure 5A shows a representative darkfield image and Figure 5B shows the corresponding fluorescence image after staining with Lysotracker Red, which is fluorescent only in strongly acidic compartments, such as the lysosome. The compartments containing blue NPs colocalize well with the lysosome marker, confirming that the NPs have indeed been translocated into the cell interior through a vesicular trafficking process.

Figure 5.

Correlation of darkfield and fluorescence microscopy after treatment with Lysosome Tracker Red. A. Color darkfield image. B. Monochromatic darkfield image with overlaid fluorescence image. Blue-shaded cell regions containing an enrichment in NPs strongly colocalize with the lysosome tracker.

Overall, our optical studies show that NPs are efficiently uptaken by the macrophages and that the exact intracellular distribution of the NP varies substantially between individual cells, indicative of significant variability in the uptake and processing mechanisms in different cells.

Cytotoxicity of the Silver NPs

We monitored the influence of silver NPs on the viability of the macrophages using a live/dead assay (see Methods) t = 1h and t = 24h after exposure to silver NPs for 5 mins. We did not observe any systematic decrease in cell viability for the NP concentration range of relevance in this work (Figure 6). We conclude that under the chosen experimental conditions the effect of the silver NPs and of potentially released silver ions^{56, 57} on the cell viability is low. Based on this data we exclude silver nanoparticle induced toxicity as source for the heterogeneity observed in the intracellular NP organization.

Figure 6.

Silver NP uptake does not affect cell viability. The histogram shows the cell viabilities obtained after exposing macrophages to NP concentrations of 0.0 µg/mL, c₁ = 1.2 µg/mL, c₂ = 4.5 µg/mL, c₃ = 8.7 µg/mL for 5 mins and subsequent incubation for 1 hour or 24 hours, respectively.

Probing NP Uptake Mechanisms through Pharmacological Inhibitors

Our studies show that the silver NPs used in this work bind to scavenger receptors on J774A.1 cells and are subsequently taken up into the cell interior. In this section we will characterize the uptake process in further detail by determining the contributions from different endocytosis mechanisms. In a first set of experiments we checked for passive translocation of NPs across the membrane at 4 °C under otherwise identical conditions as before. At 4 °C all active endocytosis is stalled, and we did not detect any indications of NP uptake. Instead, the total scattering intensity of the cells was significantly lower than for experiments performed at 37 °C, and NPs were exclusively detected attached to the apical cell surface. These observations confirm that the NP uptake is the result of energy dependent endocytosis.

In the next step, we systematically blocked various endocytosis pathways using pharmacological inhibitors. An overview of the applied pharmacological inhibitors and the affected uptake mechanisms is given in Table1. We verified in control experiments with selected reagents (see Methods) that the inhibitors efficiently affected the internalization of positive controls under the chosen experimental conditions. In the case of NPs we detected unequivocal deviations from non-inhibited uptake in the association and spatial distribution of NPs only for inhibitors that affect clathrin-mediated or actin-dependent endocytosis. Figure 7 shows representative darkfield images of the cells obtained under these conditions. When the actin-mediated endocytosis of NPs was inhibited by cytochalasin D, we exclusively observed the “blue” cell phenotype (Figure 7A). NP agglomerates with a red-shifted spectral response were absent, indicating that the formation of larger NP agglomerates is actin-dependent. In contrast, the inhibition of clathrin-mediated endocytosis (Figure 7B) led to an overall decrease in the amount of NPs with a spectral response in the blue. NP agglomerates were also present, but the intracellular distribution of the NPs and clusters changed. Especially in the presence of chlorpromazine, the NPs were often found to be concentrated in specific subcellular compartments with prominent concentric, “doughnut”-like shapes, which were absent without inhibition. These compartments were found to be acidic using the pH sensitive fluorescence dye Lysotracker Red. We conclude that although the inhibition of clathrin recycling through chlorpromazine did not inhibit the formation of intracellular NP agglomerates, the observed change in the spatial distribution of the NP agglomerates within the cells suggests some interplay of clathrin- and actin-based endocytosis machineries in the formation of NP agglomerates under uninhibited conditions.^{58, 59}

Table 1.

Summary of the pharmacological inhibitors used and their functions.

Inhibition chemicals	Functions
Monodansyl cadaverine	inhibits transamidase activity and clathrin-coated pit formation71
Chlorpromazine	blocks clathrin disassembly and recycling72
Amantadine	blocks the budding of clathrin-coated pits73
Nystatin	disrupts lipid-raft74
Genistein	inhibits the phosphorylation of caveolin75
Methyl-β-cyclodextrin	sequesters cholesterol from the membrane74
Colchincine	depolymerizes microtubles76
Rottlerin	inhibits protein kinase C77
Wortmannin	inhibits phosphatidylinositol (PI)-3-kinase78
Cytochalasin B	depolymerizes actin microfilament71
Cytochalasin D	depolymerizes actin microfilament79

Figure 7.

Pharmacological inhibitors indicate that clathrin-dependent and actin-dependent endocytosis processes participate in silver NP uptake. A. In the presence of cytochalasin D (blocks actin-dependent endocytosis) only the blue macrophage phenotype is observed. B. Treatment with chlorpromazine (blocks clathrin-dependent endocytosis) leads to the formation of large NP agglomerates in spatially defined compartments, some of which are magnified in the insets.

Interestingly, we found that cytochalasin D also induces significant changes in the spatial distribution of the NPs on the macrophage surface as determined by scanning electron microscopy. Figure 8 displays representative SEM images of cells after exposure to a loading solution of 4.5 µg/mL for 5 mins and subsequent incubation in NP-free culture medium for 20 mins with and without cytochalasin D. Top and bottom rows show representative SEM images of high and low NP association levels, respectively, for both conditions. Whereas in the absence of cytochalasin D (Figure 8A & C) we frequently detected large NP clusters on the cell surface, the NP association on the cell surface was much less prominent for cytochalasin D treated cells (Figure 8B & D). One possible interpretation of the differences in NP clustering observed in our optical and electron microscopic studies is the existence of two uptake mechanisms. In the first mechanism individual 57.7 nm diameter NPs and small NP clusters are directly taken up via a clathrin-dependent but actin-independent pathway, which avoids an accumulation of NPs on the cell surface, whereas a second actin-dependent mechanism involves the association of NPs into larger clusters on the cell surface. Similar as in conventional phagocytosis the surface-formed NP clusters are finally engulfed in an actin-dependent step. Although the two identified endocytosis mechanisms coexist in the macrophage ensemble, our data indicates that - in the absence of inhibitors - the two mechanisms are mutually exclusive in individual cells.

Figure 8.

Cytochalasin D reduces NP clustering on the macrophage cell surface. After exposure to 4.5 µg/mL nanoparticle solution for 5 mins, the cells were cultured for 20 mins before fixation and inspection in the SEM. Left column: SEM images of macrophages obtained in the absence of cytochalasin D. Representative images for high (A.) and low (C.) NP density are shown. Right column: SEM images obtained in the presence of 10 µg/mL cytochalasin D. Representative images for high (B.) and low (D.) NP density are shown. Only in the absence of cytochalasin D large NP clusters are observed. All scale bars represent 1µm.

Correlating NP Processing with F4/80 and CD14 Surface Antigen Expression

Macrophages mature from monocytes and can develop diverse dynamic phenotypes depending on their environment.^60–62 Detailed phenotyping of J774A.1 macrophages has revealed that these cells are on an intermediate stage of development between monocytes and macrophages.^{63, 64} An inspection of the cells in Figure 3 shows that the J774A.1 cell population is heterogeneous with regard to the morphology of the individual cells. Some cells are round and compact, whereas others are stretched and elongated. The observation of this morphological heterogeneity is consistent with the findings of previous studies and indicates different differentiation states within the cell culture.^{60, 65, 66} We wondered whether the observed heterogeneity in the silver NP uptake and processing observed in this study is correlated with different macrophage differentiation states. To test this hypothesis we augmented the colorimetric information about the NP aggregation obtained from darkfield microscopy with relative measures of the macrophage maturity markers CD14 and F4/80 as obtained through fluorescence immunostaining. While F4/80 is characteristic of mature macrophages,⁶⁷ CD14 is preferentially expressed by less differentiated monocytes.^{61, 62}

We performed the experiments in an identical fashion to that described before for monitoring the NP uptake but limited our analysis to one silver NP concentration (c₂) and incubation time (t = 4h). Figure 9 shows correlated darkfield and fluorescence images for macrophages with labeled F4/80 (top row) or CD14 (bottom row) obtained under these conditions. The representative images show that cells containing a higher degree of NP agglomeration (i.e. that have green or orange color) show higher expression levels of F4/80, whereas the cells containing non-clustered NPs (i.e. that are blue) show higher expression levels of CD14. These findings confirm that one important parameter that contributes to the phenotypic variability with regard to the NP processing is the maturation state of the macrophages. Taking into account the results of our endocytosis inhibition studies, we can conclude that scavenger receptor mediated NP uptake by differentiated J774A.1 macrophages is associated with the formation of NP clusters through the described actin-dependent endocytosis process, whereas monocyte-like J774A.1 cells contain smaller NP agglomerates or even individual NPs, taken up through the clathrin-dependent but actin-independent endocytosis. This behavior is overall consistent with the general trend observed for the internalization of larger objects that mature macrophages show a more effective phagocytosis than monocytes.^68–70

Figure 9.

Phenotypic variations in NP uptake are correlated with F4/80 and CD14 expression levels. Left: Darkfield image of macrophages with silver NPs (c₂ = 4.5 µg/mL, t = 4h). Right: Corresponding fluorescence image labeled for F4/80 (top) or CD14 (bottom). Macrophages containing NPs with blue color show low F4/80 intensity and high CD14 intensity, whereas macrophages containing NP agglomerates, as indicated by their green color, show high F4/80 and low CD14 intensity.

Conclusion

Due to their superb photophysical properties and their clustering dependent spectral responses, silver NPs facilitate a direct quantitative optical tracking of their uptake and intracellular fate. In this manuscript we investigated the interactions of negatively charged silver NPs with J774A.1 murine macrophages under sufficiently low NP concentrations to avoid a measurable impact on cell viability. Our findings demonstrate that the uptake of the silver NPs is mediated by scavenger receptors, which enable an efficient binding of the NPs to the macrophage surface. Furthermore, ensemble averaged spectral characterizations of the NPs confirmed that their internalization is associated with a time- and concentration-dependent association into NP clusters. A more detailed investigation of the NP association patterns on the single cell level revealed dramatic cell-to-cell fluctuations in the spatial clustering of the NPs within the cells. We characterized the internalization of the NPs using pharmacological inhibitors and found indications of two coexisting uptake mechanisms: actin- and clathrin- dependent endocytosis. The clathrin-dependent endocytosis leads to non-agglomerated NPs within the cells with a strong spectral response in the blue, while actin-dependent endocytosis leads to the formation of compact NP clusters whose spectral responses are significantly red-shifted. The NP organization on the macrophage surfaces, as characterized in the SEM, indicates that the actin-dependent uptake involves NP clustering on the cell surface. Pronounced differences in the expression of macrophage maturity markers imply that the non-agglomerated NP phenotype is preferentially associated with monocyte-like J774A.1 cells, while the NPs in more mature macrophages show a higher degree of aggregation.

The observed heterogeneities in NP uptake and processing are important, since J774A.1 is a commonly used model system for the nanotoxicological characterization of engineered NPs. Our findings suggest that an explicit consideration of the heterogeneity in the nanoparticle-cell interactions can further improve the reliability of nanotoxicological predictions obtained in vitro with J774A.1 and other macrophage cell lines.

Methods

Nanoparticle preparation & characterization

Citrate stabilized silver NPs from British Biocell International (BBI) (Ted Pella, catalog number 15707-1SC) were used for all experiments. The average diameter of the NPs, as determined by TEM, was 57.7 ± 6.9 nm, and the average aspect ratio was 1.17 ± 0.15. TEM images and size and shape distributions of the NPs are shown in Figure S1. The TEM image confirms that the NPs were roughly spherical and had a smooth surface. The citrate ligands were exchanged with PEGs by incubating the NPs overnight in a 10 mM aqueous solution of (HS-(CH₂)₁₁-(OCH₂CH₂)₆-OCH₂-COOH), then washed by centrifugation, and resuspended in HBSS buffer (137.93 mM NaCl, 5.33 mM KCl, 4.17 mM NaHCO₃, 0.441 mM KH₂PO₄, 0.338 mM Na₂HPO₄, 0.407 mM MgSO₄, 0.493 mM MgCl₂, 1.26 mM CaCl₂, 5.56 mM D-glucose) before incubation with the cells. UV-Vis spectra of the NP solution acquired before and after incubation with the cells indicated that the PEG-functionalized NPs were stable (Figure S2). Zeta potential measurements of PEG-functionalized NPs in water were performed using a Malvern NANO-ZS90 Zetasizer at 25 °C.

Cell culturing

The murine macrophage J774A.1 cell line was purchased from American Type Culture Collection (ATCC). The cells were cultured in advanced Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 50 units/mL penicillin, and 50 µg/mL streptomycine in an incubator at 37°C, 5% CO₂ and 95% relative humidity. To subculture, cells were detached from the flask substrate by gentle scraping, aspirated and dispensed into fresh complete culture medium. Cells to be imaged under dark-field microscopy or fluorescence microscopy were grown on glass coverslides to approximately 30% confluency.

Monitoring NP uptake

A glass slide containing adherent macrophages was covered by 0.5 mL of NP containing solution (c = 1.2 µg/mL – 8.7 µg/mL) in HBSS buffer and incubated for 5 mins at 37°C. Then the cells were washed with pre-warmed HBSS buffer and the NP containing solution was replaced by culture medium. The cells were then incubated for variable periods of time (t = 1h – 24h) before optical characterization by the darkfield microscope equipped with a cage incubator.

Effect of homopolymeric nucleic acids on nanoparticle uptake

Cells were pre-incubated with 100 µg/mL polyinosinic acid (poly-I) or 100 µg/mL polycytidylic acid (poly-C) containing culture medium for 0.5h and then incubated with a solution of 4.5 µg/mL silver NPs in HBSS buffer which contained identical amounts of either of poly-I or poly-C for 5 mins. After the slide containing the cells were rinsed with HBSS, the cells were culture for another 0.5h in the poly-I or poly-C containing medium. Finally the slide containing cells were integrated into a homemade chamber and transferred to an optical microscopy for inspection. Images were recorded under darkfield white-light illumination. Control slides were treated in an identical fashion, but addition of poly-I and poly-C was omitted.

Pharmacological inhibition of NP uptake

Macrophage cells (J774A.1) were preincubated with different pharmacological pathway inhibitors at 37°C (clathrin-mediated endocytosis: 200 µM monodansyl cadeverine for 20 mins, 10 µg/mL chlorpromazine for 1h, 500 µM amantadine for 0.5h; caveolae/lipid raft-mediated endocytosis: 200 µM genistein for 1h, 50 µg/mL nystatin for 15 mins, 2.5 mM methyl-beta-cyclodextrin (MβCD) for 2h; pinocytosis: 10 µg/mL colchicine for 0.5h; macropinocytosis, 2µM rottlerin for 0.5h; fluid-phase endocytosis, 400 nM wortmannin for 0.5h; phagocytosis and macropinocytosis: 10 µg/mL cytochalasin B for 2h, 10 µg/mL of cytochalasin D for 2h), then incubated with a pre-warmed solution of 4.5 µg/mL NPs in inhibitor-containing HBSS buffer for 5 mins. Subsequently, the unbound particles were washed away with copious amount of pre-warmed HBSS buffer, and cells were cultured for another 1h in DMEM supplemented with same amount of inhibitors before optical inspection.

The experimental conditions for the inhibition studies were based on published procedures and for selected inhibitors for which reliable positive controls were available we validated the efficacy in internalization assays. Alexa-Fluor 488-Transferrin (Life Technologies), FITC-Dextran (Life Technologies), and 1µm polystyrene beads (Sigma-Aldrich) were used as markers for clathrin-mediated pathway, fluid-phase endocytosis, and phagocytosis. The control experiments (Figures S3 & S4) confirmed that the inhibitors efficiently block the uptake of the positive controls under the chosen experimental conditions.

TEM sample preparation

Cells grown on the culture flask were treated with 4.5 µg/mL Ag particles in HBSS for 5 mins and then cultured in particle-free medium for 4h. Then the cells were scraped off, suspended, washed with HBSS, and fixed with 2.5% glutaraldehyde in 1×PBS (pH = 7.4) for 1h at room temperature. After washing 3 times with PBS for 10 mins each time, the cells were post-fixed with 1% OsO₄ for 0.5h in dark. Then the cells were further washed 3 times with 3.6% NaCl for 5 mins each time, rinsed with distilled water for 5 mins, and stained en bloc with 5% uranyl acetate for 1h. After washing with distilled water, the cells were dehydrated in an ascending ethanol series (50%, 75%, 90%, 100%) and finally embedded in EMbed 812 (Electron Microscopy Sciences, Hatfield PA). Blocks were polymerized, sectioned and mounted on copper grids. The grids were allowed to dry and imaged on a JEM2010 Transmission Electron Microscope (JEOL CO Peabody, MA).

SEM sample preparation

The cell culture, nanoparticle particle treatment procedure, pharmacological inhibitor treatment procedure were identical to sample preparation for the optical studies with the exception that the J774A.1 cells were grown on 1 × 1 cm silicon substrate under the same culturing conditions. The cells on silicon substrates were fixed by 4% formaldehyde solution for 15 mins under room temperature followed by 3 times ice-cold 1×PBS wash for 5 mins each time. The substrates were rinsed with DI water and gently air dried before inspected by a Zeiss SUPRA 40VP SEM using an acceleration voltage of 1.5 kV.

Darkfield microscopy

The cell slides were imaged on an inverted microscopy (Olympus IX71) under darkfield illumination through a high numerical aperture (NA) oil condenser (NA = 1.2–1.4). The illumination light was provided by 100 W Xenon Lamp (Till Oligochrome) which enabled rapid interchange between three different excitation filters (λ₁ = 494 ± 20 nm; λ₂ = 560 ± 14 nm; λ₃ = 650 ± 10 nm) and whitelight excitation. The scattering light was collected through a 60× oil objective (NA = 0.65) and captured on an electron multiplying CCD (EMCCD, Andor Ixon⁺). The light source and EMCCD were synchronized by digital delay generator (Stanford Research System, Model DG 645). The exposure time for each monochromatic image was 0.05s and the cycle time for a full set of four images was 0.4s. Digital color images were recorded under white light illumination using an Olympus SP310 digital camera attached to the microscopy through an eyepiece adapter.

Colorimetric imaging and image analysis

All image processing was performed with custom written Matlab codes. The whitelight and corresponding monochromatic images recorded in the three wavelength channels (λ₁, λ₂, λ₃) were background corrected (pixel-by-pixel) with the average pixel intensity calculated from cell-free regions of the same image. Subsequently, the monochromatic images were corrected for differences in the lamp intensity in (λ₁, λ₂, λ₃). After the cells had been localized in the whitelight image, the intensity in the whitelight image (I_w) and in the three monitored wavelength channels was determined for each pixel within the detected cell areas. If I_w lay above the cell background threshold, the pixels were included in the data analysis. For each investigated loading concentration and time point, at least 120 cells in multiple areas were inspected, recorded, and analyzed. The relative scattering intensities (I₁, I₂, I₃) were calculated by integrating the intensities on the three wavelength channels (λ₁, λ₂, λ₃) for all selected pixels in all investigated cells and subsequent division through the sum of the total intensities on all three channels.

Cell viability measurements

After incubation of the cells with different concentrations of silver NPs for defined times, commercial live/dead cell viability assay (Life Technologies, USA) was added to the cells according to the manufacturer’s directions and maintained in the incubator for 30 mins. Then the cells were washed with HBSS to remove the excess dye. The assays consists two components: calcein acetoxymethyl (calcein AM) and ethidium homodimer-1(EthD-1). Nonfluorecent cell-permeable calcein AM is converted to intensely fluorescent calcein by intracellular esterase in living cell, while EthD-1 enters cells with damaged membranes and produces a bright red fluorescence in dead cells. Living cells were detected with excitation and emission filters centered at 480 ± 17 nm and 510 ± 10 nm, respectively, and a 495 nm dichroic mirror. The EthD-1 signal for dead cells were measured with excitation and emission centered at 500 ± 25 nm and 610 ± 10 nm, respectively, and a 525 nm dichroic mirror.

Optical correlation of NPs with lysosomes tracker

Lyso Tracker Red DND-99 (Invitrogen) was employed to localize the acidic organelles. J774A.1 cells were first exposed to a pre-warmed solution of 4.5 µg/mL silver NPs for 5 mins in HBSS, after which unbound particles were washed away with copious amount of pre-warmed HBSS buffer. Subsequently the cells were cultured for another 24h in nanoparticle-free culture medium before incubation with 1µM Lyso Tracker Red solution for 0.5h under growth condition. The cell slides were washed and observed using a fluorescence microscopy using an appropriate filter set (excitation: 563 ± 9 nm; emission: 600 ± 10 nm; 580 nm dichroic). Dark-field scattering images were recorded under white light illumination as described.

Fluorescent immunolabeling of the surface marker

J774A.1 cells were grown on glass coverslides to approximately 30% confluency. Cell were fixed by 2% formaldehyde solution for 15 mins at room temperature followed by three times washing with ice-cold 1× PBS for 5 mins each time. The cells were then incubated in ice-cold 1× PBS buffer containing 1% BSA for 1h to block nonspecific binding at room temperature and subsequently washed three times with copious amounts of ice-cold 1× PBS buffer for 5 mins. The cells were then incubated with monoclonal anti-F4/80 antibody conjugated with FITC (diluted 1:10 in 1× PBS) or monoclonal anti-CD14 antibody conjugated with FITC (diluted 1:10 in 1× PBS) in a water vapor saturated atmosphere for 2h at room temperature. Finally the cell slides were washed three time with 1× PBS buffer for 5 mins each time and then integrated into a homemade flow chamber and transferred to a combined darkfield and fluorescence microscopy for inspection. The fluorescence signal was detected with excitation and emission filters centered at 480 ± 17 nm and 510 ± 10 nm, respectively, and a 495 nm dichroic mirror.