Coating Metal Nanoparticle Surfaces With Small Organic Molecules Can Reduce Non-Specific Cell Uptake

Abstract

Elucidation of mechanisms of uptake of nanoparticles by cells and methods to prevent this uptake is essential for many applications of nanoparticles. Most recent studies have focused on the role of proteins that coat nanoparticles and have employed PEGylation, particularly dense coatings of PEG, to reduce protein opsonization and cell uptake. Here we show that small molecule coatings on metallic nanoparticles can markedly reduce cell uptake for very sparsely PEGylated nanoparticles. Similar results were obtained in media with and without proteins, suggesting that protein opsonization is not the primary driver of this phenomenon. The reduction in cell uptake is proportional to the degree of surface coverage by the small molecules. Probing cell uptake pathways using inhibitors suggested that the primary role of increased surface coverage is to reduce nanoparticles’ interactions with the scavenger receptors. This work highlights an under-investigated mechanism of cell uptake that may have played a role in many other studies and also suggests that a wide variety of molecules can be used alongside PEGylation to stably passivate nanoparticle surfaces for low cell uptake.

Graphical Abstract

We show that, independent of PEGylation, increasing surface coverage of gold nanoparticles using a small molecule reduces their uptake by cells in vitro. This reduction is achieved by suppressing scavenger receptor mediated uptake and has no correlation with degree of protein opsonization.

A major challenge in using nanoparticles for targeted delivery is that NPs administered intravenously (IV) primarily accumulate and persist in the liver and spleen.^1–4 Considering the Kupffer cell population in the liver and the large macrophage population in the spleen, this accumulation is generally attributed to the phagocytosis of nanoparticles by macrophages in the body, though recent work has implicated other immune cells as well.^{3, 5,6} Decreasing uptake of nanoparticles by macrophages in vitro has been an active area of research in the hunt for long-circulating, liver-avoiding nanoparticles that could improve targeted delivery for diseases such as cancer. Many different passivating agents — polyethylene glycol (PEG),^7–9 chitosan,¹⁰ hyaluronic acid,¹¹ polyoxazoline,¹² among others — have been investigated for their ability to shield the nanomaterials from phagocytic cells in vitro. PEG is by far the most commonly used passivating agent. PEG acts as a barrier layer that prevents the direct interaction between nanoparticles and their environment. PEG functionalized nanoparticles are endocytosed by macrophages less than their unfunctionalized counterparts.¹³ In general, it has been reported that as the amount of PEG on the surface of the nanoparticle increases, the cellular uptake of these particles decreases. It is commonly thought that this phenomenon is due to PEG’s ability to reduce protein opsonization since it has been demonstrated that by increasing the density of PEG¹³ or the length of PEG,^14,15 there is a decrease – but never an elimination – of protein opsonization. Most reports focus on increasing the density of the PEG on the surface of the nanoparticle either through backfilling,¹⁶ using branched PEGs,⁹ using a hydrophobic core as a spacer,¹⁷ or by incubating the particles in an excess of PEG.¹³ It has also been demonstrated that the conformation of the PEG coating (brush vs mushroom) can play a significant role in NP uptake and, as a result, increasing PEGylation is not always beneficial.¹⁸ For all PEGylation strategies, fully protecting the surface of the NPs appears to be quite challenging as, for example, even with “dense” coatings it was found that cysteine residues on peptides or proteins can penetrate the PEG layer and attach to the surface of the particle.¹⁹

Recent reports have suggested that the protein corona hypothesis is more complex than simply the amount of protein adsorbed onto the surface of a nanoparticles.²⁰ The identity and abundance of proteins composing the corona have been observed to influence nanoparticles ability to avoid cell uptake.^21,22 Recognizing that composition may be more important than amount, protein corona fingerprinting has been undertaken with metallic nanoparticles in order to predict in vitro cell uptake based on the composition of the protein corona.²² It has also been demonstrated that for liposomes the identity of the proteins forming the corona depends on PEG length and this change in composition correlates with a change in cell uptake.¹⁴ Interestingly, this report focused on using apolipoproteins in the protein corona to enhance uptake by cancer cells expressing a high level of scavenger receptor class B type 1, while another study suggested that apolipoprotein J may be the active mediator of the low cell uptake achieved by PEGylation.¹⁵ Additionally, reports have investigated the role of the conformation of the proteins with evidence that suggests that denatured albumin on the surface of a nanoparticles could trigger its cell uptake in a phagocytic cell.²³ Overall, the protein corona is widely believed to be a key mediator of cell uptake but there remains a debate on how the corona mediates the uptake process.

In our previous publication on the controlled synthesis of gold nanoparticle aggregates using a small molecule crosslinker, pentaerythritol tetrakis-(3-mercaptopropionate) (PTMP), we noticed what appeared to be unusually low cell uptake in a human macrophage-like cell line (THP-1).²⁴ In transmission electron micrographs, most cells were void of nanoparticle aggregates and the cells that contained nanoparticle aggregates generally had one aggregate per cell. This result was in contrast to publications that showed endosomes full of gold nanoparticle aggregates^25–27 or solid gold spheres of similar size (60 nm).^28,29 In further work, we found that these aggregates have an extremely high level of surface coverage as measured by rate of dissolution in potassium cyanide (KCN) (Figure S1A) and hypothesized that this is responsible for their extremely low uptake (Figure S1B).

In order to study the effect of surface coverage in a controlled manner, we prepared 50 nm solid gold spheres with similar levels of PEGylation but varying surface coverage by PTMP. We found that uptake by three different cell lines correlated with the degree of surface coverage with the highest level of surface coverage closely matching the behavior of the aggregates. Interestingly, the degree of uptake could be decreased significantly below that observed for 50 nm solid gold spheres with a higher amount of PEGylation but no small molecule surface coverage. While amount of cell uptake was inversely proportional to surface coverage on the gold nanoparticles, there was no correlation with amount of protein opsonization and similar results were obtained in serum-free media. In particular, the fact that there was no change in the uptake trends when serum-free media was used strongly suggests that the inhibition of uptake by the small molecule coating is not driven by alterations in the composition of the protein corona. An inhibitor study identified the scavenger receptors as the predominant mechanism of uptake for the nanoparticles with incomplete surface coverage, suggesting that more complete surface coverage suppresses this mechanism – likely by inhibiting direct interaction between the scavenger receptors and the NP surface. Overall, we show that, for gold nanoparticles, in vitro uptake can be markedly suppressed by coating their surface with a small molecule and this likely occurs by suppressing uptake via the scavenger receptors and is independent of protein opsonization.

Results and Discussion

Previously synthesized²⁴ biocompatible nanoparticle aggregates had remarkably low cellular uptake in vitro. Electron micrographs of THP-1 cells differentiated into a macrophage phenotype and exposed to gold nanoparticle aggregates for 24 hours showed that the majority of cells contained no aggregates and those that did had only one or two aggregates. This low cell uptake was in contrast to other studies in our lab^{15a, 17} and in the literature^25–29 which showed high levels of uptake with multiple vesicles per cell containing numerous particles. We were interested in determining why these aggregates showed such low uptake into a phagocytic cell line. We found that these aggregates have an extremely high level of surface coverage as measured by rate of dissolution in potassium cyanide (KCN) (Figure S1A) and hypothesized that this is responsible for their extremely low uptake. Since the aggregates possess different morphology than standard spherical NPs and the degree of surface coverage cannot be readily tuned for the aggregates, we decided to investigate the influence of surface coverage with a small molecule in the context of solid spherical gold nanoparticles.

We developed a library of nanoparticles with variable surface coverage, characterized their physical properties, and measured their stability in media, salt, and cyanide before determining their cellular uptake. Our library of nanoparticles was based on 50 nm citrate-stabilized gold nanoparticles functionalized with variable amounts of a tetravalent thiol small molecule, pentaerythritol tetrakis (3-mercaptoproprionate) (PTMP). While PTMP has previously been used to assemble 5– 15 nm gold nanoparticles into aggregates, 50 nm gold nanoparticles are too large to create aggregates.²⁴ We were able to modulate surface coverage of solid 50 nm particles by varying the amount of PTMP in the reaction mixture (Scheme 1). Particles are referred to as high, medium, or low surface coverage particles depending on the amount of PTMP measured on the surface of the nanoparticles. Unfortunately, NPs only coated with PTMP are unstable in high salt solutions (like 1x PBS and media) and require PEGylation for stability in these environments. This prevents evaluating the impact of PTMP coating in isolation. Thus, to best study the effect of PTMP coating, high, medium and low surface coverage NPs were functionalized with PEG-maleimide which reacts with free thiols on the NP surface. The degree of PEGylation was kept to the amount necessary for stability in solution and was so minimal that we could not directly measure it (described further below and in Table 1). PEGylated 50 nm particles, citrate-stabilized 50 nm particles and 50 nm aggregates assembled from 5 nm particles were used as controls. PEGylated particles were used to understand how coverage with PTMP compared to the most commonly used passivating agent. Of note, the amount of PEG on the PEGylated particles was much higher than for any of the three PTMP-coated NPs as these NPs were not stable with the minimal amount of PEGylation necessary to stabilize the small molecule-coated NPs. Citrate-stabilized particles were used as a positive control since they are highly endocytosed.¹³ The 50 nm nanoparticle aggregates were used as a control due to their previously noted low cellular uptake and to ensure that the simplified model was able to match the characteristics and behavior of the nanoparticle aggregates.

Scheme 1:

Library Synthesis

Table 1:

Surface Coverage Library Characterization

Particle Size Coating	50 nm Citrate	50 nm PEG	Aggregates	50 nm High Coverage	50 nm Medium Coverage	50 nm Low Coverage
Hydrodynamic Diameter (nm)	52	59	50	66.3	74.6	74.3
Zeta Potential (mV)	−28	−10	−18	−26	−9	−6
Weight Percent PEG	-	8%	10%	<1.5%1	<3%1	<4%1
Weight Percent PTMP	-	-	14%	3.5± 0.8	2%	0.8 ±0.1
Percent Surface Area Covered2	-	70%	≤100%	≤100%	≤71%	≤29%

¹Below limit of quantitation by TGA which is 5% by weight. The maximum possible PEG was calculated as 5 – (weight % PTMP).

²See Figure S8 for explanation of assumptions and calculations

All of the particles were similar in size and charge (Table 1). While the amount of PEG on the PEGylated 50 nm particles was 8% by TGA, the high, medium and low surface coverage nanoparticles did not have enough organic material (PEG-Maleimide plus PTMP) to accurately quantify by TGA with our instrument. Based on the sensitivity of our system and the amount of material we were able to prepare, we know the particles contained <5% PEG by weight. In order to quantify the amount of PTMP on these nanoparticles we prepared the particles in elemental analysis grade water and performed ICPMS to quantify the amount of gold and sulfur. The calculated amounts of PTMP for the 50 nm gold nanoparticle aggregates were similar to that determined by TGA (16%)²⁴ suggesting that using ICPMS to analyze minute amounts of PTMP on the gold nanoparticle surface would be accurate. The high, medium and low particles contained 3.5%, 2% and 0.8% PTMP by weight, respectively, based on the average of four different batches of particles. We estimate that this corresponds to surface coverage of approximately >95%, 71% and 29%, respectively (See Figure S8 for assumptions and calculations). Importantly for this discussion, the amount of organic material was much less than that seen with the PEGylated particles. No evidence of aggregation was seen with these particles using TEM or UV measurements (Figure S3).

In order to confirm that the variable amounts of PTMP used to coat the particles did correlate to a change in surface coverage, the nanoparticles were treated with 0.1 M potassium cyanide. Cyanide stability did correlate with the variable amounts of PTMP, as the amount of PTMP on the nanoparticle increased the cyanide stability also increased (Figure 1A). As with the citrate-stabilized and PEGylated particles, the low surface coverage particles were stable for less than 1 hour with a 50% drop in absorbance seen within 5 minutes. The 50 nm medium surface coverage particles were stable for greater than 4 hours but had completely etched by 24 hours. The 50 nm high surface coverage particles were the most stable of the solid 50 nm nanoparticles with less than 20% decrease in absorbance over 24 hours in the presence of 0.1 M cyanide. Aggregates assembled from 5 nm subunits showed no change in absorbance over 24 hours (Figure 1A).

Figure 1:

Particle Stability.

50 nm particles were capped with PEG-SH or PTMP of different concentrations and the stability of these particles to was monitored over time by UV and compared to the absorbance of the sample in water over time. A. Citrate coated, PEGylated, and high low coverage particles behaved similarly dissolving within 1 hour of 0.1 M KCN exposure. Medium surface coverage particles were stable for four hours, but were dissolved by 24 hours. High surface coverage particles were the most stable to dissolution, showing only a slight reduction in absorbance at 24 hours. B. All particles except citrate-stabilized gold nanoparticles were stable in 1x PBS for 1 hour, but showed a decrease in absorbance as the particles coated the sides of the cuvettes. C. All particles behaved similarly in complete media and were stable up to 24 hours according to UV. Data points are reported as the ratio between the sample in the solution of interest compared to the sample in water, where 1 signifies no difference in the optical density and 0 represents a clear solution. Each data point is the average of 3 replications and error bars show the standard error mean of the experiment.

In PBS, citrate-stabilized 50 nm particles showed a broadening of the UV absorption spectra and decrease in the absorbance of the nanoparticles within 5 minutes of exposure and by 4 hours the particles had completely aggregated and coated the bottom of the cuvette (Figure 1B). This was also confirmed by a rapid increase in hydrodynamic diameter by DLS (Figure S4). The 50 nm particles functionalized with PEG, the 50 nm low, medium, and high surface coverage nanoparticles, as well as the 50 nm aggregates showed no shift in absorbance that would indicate aggregation. Over 24 hours in 1x PBS, all samples showed a decrease in absorbance at 531 nm (Figure 1B). This decrease had been previously described in the literature³⁰ and is the result of nanoparticles interacting with the cuvette to form a pink nanoparticle layer on the plastic. Interestingly amongst the 50 nm particles, the low coverage particles were more stable by UV than the medium or high coverage particles, showing the least decrease in absorbance. Stability of particles in PBS did not correlate with degree of surface coverage.

As described previously, UV absorbance and light scattering can be used to determine if the nanoparticles aggregate upon exposure to complete media. These particles are considered unstable if there is a change in particle size after protein exposure or if there is a broadening of the UV absorbance peak over time. According to UV measurements, no nanoparticles showed a change in absorbance compared to the sample in water (Figure 1C) nor was there any broadening of the peak associated with nanoparticle aggregation (Figure S4). These particles could also grow in size without a shift in λ_max due to the adsorption of proteins onto the surface of the nanoparticles, but DLS measurements showed a minimal increase in hydrodynamic diameter over 24 hrs in media (Figure S4). There was no correlation between surface coverage and media stability.

Using our library of nanoparticles with variable surface coverage, we wanted to determine if nanoparticle surface coverage correlated with cellular uptake. We treated three different immortalized cell lines -- mouse macrophages, human breast cancer, and mouse hepatocytes-- with equal concentrations of gold nanoparticles for 1, 4, 16, and 24 hours. We measured the gold and magnesium content in the cells using ICPMS. Gold was used to determine the amount of nanoparticles in the cells. Magnesium levels are known to correlate with cell number³¹ and were used to ensure that differences in cellular uptake were not due to variability in cell number.

Macrophages, particularly Kupffer cells, are the primary phagocytic cell in the liver and are recognized as the major phagocytic cell type to consume nanoparticles.^{3, 32} RAW 264.7 cells, an immortalized mouse macrophage cell line, were used due to the prevalence of these cells as a model system in the literature.^33–35 Surface coverage of the nanoparticles inversely correlated with cellular uptake at 24 hours (Figure 2A, TEM images Figure S7). A time course was performed (Figure 2B) in order to determine if the reduction in cellular uptake was due to a decrease in cellular uptake, a delay in cellular uptake, or rapid exocytosis. At every time point after 1 hour, the relative levels of gold in the cells were similar, though over time the total amount of gold in the cells did increase. This suggests that there was neither a delay in cellular uptake nor a rapid exocytosis. Using intracellular magnesium levels as a marker for cell number,³¹ there were no differences between the control groups and any treatment group for any cell line at any time point so variable uptake was not due to differences in cell number. This study was repeated with a human breast cancer cell line (MDA-MB-231) as well as a mouse hepatocyte cell line (AML-12). In both cell lines there was an inverse correlation between surface coverage and cell uptake (Figure 2) and this behavior was both observed early on in the time course and persisted throughout the time course. This demonstrates that this phenomenon is not cell type specific as surface coverage can influence cell uptake in both phagocytic and nonphagocytic cells.

Figure 2:

Cell Uptake of Surface Coating Library

Raw 264.7, AML-12 and MDA-MB-231 cells (A, C, and E, respectively) treated with 3.6e8 nanoparticles/mL were analyzed by inductively coupled plasma mass spectrometry for gold content at 24 hours. Cell uptake showed an inverse correlation with surface coverage for the high, medium, and low surface coverage particles. Each bar is the average of 3 independent experiments each with 4 replicates, error bars are the standard error mean. The inverse correlation between surface coverage and cell uptake was evident at 5 minutes, 1 hour, 4 hours and 24 hours for Raw 264.7, AML-12 and MDA-MB-231 cells (B, D, E respectively). Each data point is the average of 3 independent experiments with 4 replicates, error bars are standard error mean.

Furthermore, we found that this effect is not unique to gold NPs and PTMP. Techniques to measure surface coverage exist for a limited number of classes of NPs. We chose to use silver nanoparticles since their surface coverage can also be assayed by KCN dissolution and found that silver nanoparticles with a high surface coverage of PTMP capped with minimal PEG-maleimide had markedly lower uptake by RAW macrophages as compared to PEGylated silver nanoparticles (Supplementary Figure S5). In addition, it was found that coating gold nanoparticles with a high surface coverage of trimethylolpropane tris(3-mercaptopropionate) (TTMP), a trivalent analogue of PTMP, also resulted in dramatic suppression of uptake by RAW macrophages (Supplementary Figure S6). Clearly, gold and silver nanoparticles and PTMP and TTMP are related but this data suggests that the ability of a small molecule surface coating to suppress cell uptake may be relevant for different NP types and chemical composition. In order to further expand these studies, assays for surface coverage in other materials are needed.

Overall, no correlation was observed between PBS or media stability and cellular uptake for our uptake studies of the low, medium and high coverage particles using RAW macrophages, MDA-MB-231 cancer cells and AML-12 hepatocytes. Only cyanide stability, a measure of surface coverage, correlated with cellular uptake. This suggests that exposed gold on the surface of the nanoparticle plays a role in cellular uptake. High cyanide stability was a particularly useful predictor of low cell uptake as both the high coverage particles and the aggregates had the lowest uptake out of the set by a significant margin. However, stability in KCN was not a meaningful predictor of cell uptake for particles with modest (medium coverage) or poor (low coverage, citrate stabilized and PEGylated) stability in KCN. While the medium particles had greater stability in KCN, their uptake was similar to the PEGylated particles which rapidly degraded in KCN. The PEGylated, low and citrate stabilized particles had similar rates of cyanide dissolution (defined as time to reach 50% of initial max) but much greater amounts of gold were found in the cells treated with citrate stabilized particles and the low coverage particles showed greater uptake than the PEGylated particles. We hypothesized that this might be due to the fact that ⁻CN is a particularly small molecule while cell uptake is most likely mediated by interactions between the particles and proteins which are much larger than ⁻CN. Thus, coatings that prevent access of ⁻CN are certain to block the protein interaction leading to cell uptake but coatings that permit ⁻CN to reach the particle surface may still block the protein interaction to a varying degree that is not predictable based on the rate of dissolution by ⁻CN.

This led us to investigate the question of whether the protein interaction being blocked by the small molecule coating is one between the particles and proteins in solution or between the particles and proteins on the cell surface. Many studies have shown that particles, including PEGylated particles, are rapidly coated by proteins in solution and proposed that this protein corona drives cell uptake.^{13, 16,17, 19} Reports have both focused on the total amount of protein^{13, 16,17, 36} and the particular types of proteins adsorbed to the surface as predictors of cell uptake.^14,21,22

We began our studies by exposing equivalent concentrations of protein to equivalent numbers of nanoparticles and separating the proteins adsorbed to the surface of the nanoparticle by electrophoresis to determine if the amount of protein adsorbed or an obvious composition change correlated with cell uptake. Equal numbers of particles were mixed with equal volumes of serum in PBS. After isolating the proteins bound to the surface of the particles, electrophoresis was used to separate the isolated protein. Protein was visualized using a fluorescent protein stain. Interestingly, there was no trend between the amount surface coverage and the amount of protein adsorbed on the surface of the nanoparticle for the high surface coverage particles, medium surface coverage particles, low surface coverage particles, or the aggregates, in fact the largest amount of protein was found on the medium surface coverage particles (Figure 3A). Interestingly, this matches well with our experience handling the particles as the medium surface coverage particles were difficult to synthesize reproducibly and were the least stable member of the library, frequently sticking to a variety of surfaces. There is evidence in the field that suggests that the identity of the proteins adsorbed on the surface of the nanoparticles mediates the uptake of the nanoparticles by cells;^{15, 21} we next compared cell uptake in the presence and absence of proteins to see if there was a marked difference which would suggest that a particular protein was driving uptake. However, the removal of serum proteins from the media did not change the relative amounts of gold nanoparticles endocytosed by the cells (Figure 3B). These results support the hypothesis that the correlation between surface coverage and cell uptake is not mediated by adsorption of proteins from solution.

Figure 3:

Cellular Uptake Mechanisms of the Nanoparticle library in RAW 264.7 cells.

A) RAW 264.6 cells were treated with 3.6e8 nanoparticles/mL for equal numbers of gold nanoparticles with variable surface coverage for 24 hours in the presence or the absence of FBS. Cell uptake trends were maintained but the amount of gold detected in the cells was higher in the cells treated with serum free media. Bars are the average of 3 independent experiments, error bars are standard error mean. B) Gel Electrophoresis results (displayed here at 270° rotation) show no trends with the protein adsorbed onto the particle and the amount of surface coating. C) Treatment of RAW 264.7 cells with a variety of endocytosis inhibitors in serum-free media suggests that the dense surface coating of a nanoparticle is able to inhibit cell uptake through the scavenger receptor pathways. Each bar is the average of at least 2 independent experiments each with three replicates. D) The major targeted pathway is noted for each inhibiter.

We hypothesized that if solution proteins were not the primary mediators, the surface coverage must be disrupting a direct interaction with the cells. Thus, we inhibited different endocytosis pathways that have been shown to participate in nanoparticle uptake, such as clatherin-mediated endocytosis, caveolae- mediated endocytosis and phagocytosis,^37–41 to identify if there was a primary pathway that the high surface coverage might be avoiding. Clean delineation of the primary cell endocytosis pathway is challenging. Small molecule inhibitors are likely to have broad off target effects, especially with lengthy incubation periods, furthermore these lengthy incubation periods may result in the adoption of compensatory mechanisms. Finally, in the case of nanoparticle uptake studies, there are no positive nanoparticle-based controls available to demonstrate complete inhibition of a single pathway. With these limitations in mind, a cell uptake inhibition study was designed with short incubation periods and multiple inhibitors modeled on previously published assays in the nanoparticle field.⁴² Cells were treated in the absence of serum proteins with various endocytosis pathway inhibitors for one hour before being incubated with citrate stabilized, PEGylated, high surface coverage, or low surface coverage particles for 1 hour. Of note, we excluded the medium surface coverage particles from these studies because they were difficult to synthesize reproducibly and these cell studies required large amounts of nanoparticles. High surface coverage particles were included to provide a benchmark for low cell uptake. Our primary finding was that the scavenger receptors are the predominant pathway for all of the particle types. The scavenger receptors are a group of cell surface proteins that actively facilitate the uptake of a variety of molecules, predominantly negatively charged macromolecules.⁴³ The receptors are known to self- and cross-hybridize to activate uptake. Poly I is a non-specific inhibitor of the entire class.⁴⁴ Uptake by the scavenger receptors can include cavaeolae- and clatherin-mediated endocytosis as well as several other mechanisms. Thus, it is not surprising that for particles where poly I showed the largest effect multiple other inhibitors also showed a more moderate effect. We found that uptake of the citrate, PEG and low surface coverage particles was most significantly suppressed by pretreatment with Poly I. Uptake of the low-surface coverage particles was also slightly suppressed by inhibitors of caveolae- and clatherin-mediated endocytosis, while uptake of PEGylated particles was suppressed by an inhibitor of clatherin-mediated endocytosis (Chlorpromazine) to the same extent as Poly I pre-treatment. The citrate particles showed reduced uptake following pretreatment with several inhibitors in addition to Poly I (Chlorpromazine, Cytochalasin D, Wortmannin, and Methyl-B-Cyclodextran), with Methyl-B-cyclodextran (Caveolae-mediated endocytosis) showing the second largest effect. As expected, uptake of the high-surface coverage particles was below the limit of detection for all treatments. We speculate that given the heterogeneity of scavenger receptors mediated uptake that different surface coatings, such as PEG or small molecules, may suppress uptake by inhibiting different interactions within these pathways and that high surface coverage with a small molecule coating is sufficient to essentially eliminate this mechanism of uptake in vitro.

Conclusions

In vivo off-target accumulation of nanoparticles is often attributed to the phagocytosis of nanoparticles by macrophages in the liver. It has been thought that PEGylation was required to decrease macrophage uptake in vitro. We show here that gold nanoparticles (both aggregates and solid spheres) with high surface coverage of a small molecule and a minimal amount of PEGylation have a dramatic reduction in uptake by macrophages, cancer cells and hepatocytes as compared to solid gold spheres coated with only PEG, even though the PEGylated particles have a greater density of PEG on their surface compared to the particles with the small molecule coating. This low cellular uptake did not correlate to PBS or media stability, nor did it correlate to the amount of protein adsorbed to the surface of the nanoparticles. Instead, when a series of solid spheres was prepared with a range of surface coverage, cell uptake in vitro was inversely correlated with the degree of surface coverage as measured by stability during cyanide dissolution, suggesting that this type of coating may be a powerful tool for optimizing in vivo circulation as well as studying in vitro mechanisms of uptake. An initial investigation of the mechanism of uptake suggested that increasing surface coating with a small molecule reduces uptake by avoiding the scavenger receptors, a set of non-specific endocytosis pathways. Understanding how uptake by the scavenger receptors is inhibited is of interest for future work as it could permit the rational design of nanoparticles that are able to avoid macrophage uptake in vivo. Furthermore, we speculate that other reports that reduce cell uptake in vitro by increasing the density of PEGylation on the surface of nanoparticles, such as by backfilling or in situ PEGylation, are also observing this phenomenon and that the primary factor that influences in vitro cell uptake of gold nanoparticles is surface coverage.

Methods

All materials were used as supplied. Citrate-stabilized gold colloid suspensions (5 nm and 50 nm) and citrate stabilized silver colloid suspensions (40 nm) were purchased from Ted Pella. Phenol free-RPMI 1640 media, DMEM media without sodium pyruvate, and 10x PBS (without magnesium and calcium) were purchased from Life Technologies. DMEM media containing sodium pyruvate (ATCC 30–2002), DMEM-F12 media (ATCC-30–2006), AML-12 cells (ATCC CRL-2254), MDA-MB-231 cells (ATCC HTB-26), and RAW 264.7 cells (ATCC TIB-71) were purchased from ATCC. PEG-Maleimide was purchased from Nanocs. Pentaerythritol tetrakis (3-mercaptopropionate) and PEG-SH purchased from Sigma were handled under nitrogen at all times. Trimethylolpropane tris(3-mercaptopropionate) was purchased from Sigma and handled under nitrogen at all times. NOTE: If TTMP or PTMP is exposed to air and then used to synthesize particles, the size of those particles cannot be accurately predicted.

Synthesis of 50 nm SH-PEG₂₀₀₀ Particles:

A solution of PEG₂₀₀₀-SH in diH₂O (20mg/mL) was prepared. For each reaction, 7.5 mL of diH₂O containing 5.4 × 10¹⁰ particles/mL 50 nm citrate-stabilized gold nanoparticles were added dropwise to 567.5 μL of PEG-SH solution. Four batches of particles were shaken at room temperature for 2 hours before being concentrated and combined for washing. Particles were concentrated by repeated centrifugation at 10,000 g for 10 minutes until 15 mL of particles were combined into one 2 mL Eppendorf tube. The particles were washed by centrifugation (10,000g for 10 minutes) three times with 2 mL diH₂O. After washing, both tubes were concentrated to a combined volume of 1.5 mL and split into two equal volume batches, one batch was used for TGA analysis and 1 batch was diluted to 15mL and used for characterization, stability assays and cell experiments.

For the PEGylated silver nanoparticles, this procedure was followed with only the following modifications: 7.5mL of diH₂O containing 40 nm citrate-stabilized silver nanoparticles at stock concentration (OD 0.654 at λ_max=409 nm) was used instead of gold nanoparticles. Two batch of particles were synthesized, combined, and washed 3 times with diH₂O by pelleting the particles via centrifugation (10,000g for 10 minutes). The supernatant was removed and the particle pellet was resuspended in 2 mL diH₂O.

Synthesis of 50 nm High Surface Coverage Particles:

A solution of pentaerythritol tetrakis (3-mercaptopropionate) (PTMP) in ethanol (400 μg/mL) was prepared, for each reaction 1.74 mL PTMP solution was added to 2.598 mL diH₂O. 7.5mL of diH₂O containing 50 nm citrate-stabilized gold nanoparticles was added dropwise to the diH₂O-ethanol solution. The reaction solution was shaken for 2 hours on a tabletop shaker at top speed. After shaking, the reaction was left on the benchtop at room temperature for 24 hours. A solution of PEG₂₀₀₀-Maleimide in diH₂O (20 mg/mL) was prepared and for each reaction 6.456 mL was added to the reaction solution after it had stood for 24 hours at room temperature. The reaction was shaken at top speed on a table top shaker for 2 hours at room temperature. Four batches of particles were synthesized and combined before washing the particles 3 times with diH₂O. The particles were washed by pelleting the particles via centrifugation (10,000g for 10 minutes). The supernatant was removed and the particle pellet was resuspended in 2 mL diH₂O. The 2 mL of particles was separated into two batches each of 500 μL, one batch was used for TGA analysis and 1 batch was diluted to 15mL and used for characterization, stability assays and cell experiments.

For the high surface coverage silver nanoparticles, this procedure was followed with only the following modifications: 7.5mL of diH₂O containing 40 nm citrate-stabilized silver nanoparticles at stock concentration (OD 0.654 at λ_max=409 nm) was used instead of gold nanoparticles. One batch of particles was synthesized, combined, and washed 3 times with diH₂O by pelleting the particles via centrifugation (10,000g for 10 minutes). The supernatant was removed and the particle pellet was resuspended in 2 mL diH₂O.

Synthesis of 50 nm High Surface Coverage TTMP Particles:

The concentration of trimethylolpropane tris(3-mercaptopropionate) (TTMP) was selected such that the number of thiols was equivalent to the PTMP high surface coverage particles. Therefore, a solution of TTMP in ethanol (400 μg/mL) was prepared, and for each reaction 1.89 mL TTMP solution was added to 2.598 mL diH₂O. 7.5mL of diH₂O containing 50 nm citrate-stabilized gold nanoparticles was added dropwise to the diH₂O-ethanol solution. The reaction solution was shaken for 2 hours on a tabletop shaker at top speed. After shaking, the reaction was left on the benchtop at room temperature for 24 hours. A solution of PEG₂₀₀₀-Maleimide in diH₂O (20 mg/mL) was prepared and for each reaction 6.456 mL was added to the reaction solution after it had stood for 24 hours at room temperature. The reaction was shaken at top speed on a table top shaker for 2 hours at room temperature. The particles were washed 3 times by pelleting the particles via centrifugation (10,000g for 10 minutes). The supernatant was removed and the particle pellet was resuspended in 2 mL diH₂O.

Synthesis of 50 nm Medium Surface Coverage Particles:

A solution of PTMP in ethanol (4 μg/mL) was prepared by dilution of a 400 μg//mL PTMP in ethanol. For each reaction 1.74 mL PTMP solution was added to 2.598 mL diH₂O. 7.5mL of diH₂O containing 50 nm citrate-stabilized gold nanoparticles was added dropwise to the diH₂O-ethanol solution and the synthesis proceeded as described above. Four batches of particles were synthesized and combined before washing the particles via dialysis. Particles were dialyzed because attempts to wash the particles caused the particles to irreversibly stick to the plastic. To remove excess linker, the particles were dialyzed against DiH₂O using 20,000 MWCO cellulose membrane (Spectrum). DiH₂O was changed at 1 hour, 2 hours, 4 hours and 16 hours. At the last diH₂O change the particles were dialyzed for 2 hours and then transferred to a clean vial. In order to concentrate the particles solutions after dialysis the nanoparticles were dialyzed against a concentrated PEG₈₀₀₀ solution using a 1000 MWCO regenerated cellulose membrane (GE Life Sciences). The dialysis vials contained a maximum of 8 mL of particle solution, in order to increase throughput multiple vials were used in individual solutions of PEG₈₀₀₀ to concentrate the particles. Every day in the morning and in the evening the vials were refilled with particles, over 12 days 80 mL of dialyzed particles were concentrated into a volume of 20 mL. Particles were split into two batches, 1 batch was used for TGA and the second batch was diluted to 15mL and used for characterization, stability assays and cell experiments. These nanoparticles were difficult to make consistently with some batches being stable after dialysis and others being unstable after just 24 hours.

Synthesis of 50 nm Low Surface Coverage Particles:

A solution of PTMP in ethanol (0.04 μg/mL) was prepared by serial dilution of a 400 μg/mL solution. For each reaction 1.74 mL PTMP solution was added to 2.598 mL diH₂O. 7.5mL of diH₂O containing 50 nm citrate-stabilized gold nanoparticles was added dropwise to the diH₂O-ethanol solution and the synthesis proceeded as described for the High Surface Coverage PTMP particles. The particles were washed as described for the Medium Surface Coverage PTMP particles. For TGA analysis, the particles were pelleted by centrifugation (3000 g for 45 minutes) in a glass tube, any particles that stuck to the tube were dislodged with DCM and transferred to the TGA.

Synthesis of Nanoparticle Aggregates:

Nanoparticle aggregates were synthesized as described previously²⁴ but at 15 times scale. Briefly a solution of pentaerythritol tetrakis (3-mercaptopropionate) (PTMP) in ethanol (400 μg/mL) was prepared and for each reaction 1.74 mL was added to 2.598 mL of diH₂O. 7.5 mL of diH₂O containing 5 nm citrate-stabilized gold nanoparticles (5 × 10 ¹³ particles/mL) was added dropwise to the diH₂O-ethanol solution. The reaction solution was shaken for 2 hours on a tabletop shaker at top speed. After shaking, the reaction was left on the benchtop at room temperature for 24 hours. A solution of PEG₂₀₀₀-Maleimide in diH₂O (20 mg/mL) was prepared and for each reaction 6.456 mL was added to the reaction solution after it had stood for 24 hours at room temperature. The reaction was shaken at top speed on a tabletop shaker for 2 hours at room temperature. The particles were washed by pelleting the particles via centrifugation (10,000g for 10 minutes). The supernatant was removed and the particle pellet was resuspended in 2 mL milliQ diH₂O. Four batches were synthesized in separate vials and combined during washing. Particles were filtered using a 25 mm 0.2 μm polycarbonate track etch filter in a Swinnex housing. The filter was washed with 1.5 mL of milliQ diH₂O. The filtrate was pelleted via centrifugation (10,000g for 10 minutes) and resuspended in 1 mL of milliQ diH₂O. The 1 mL of particles was separated into two batches each of 500 μL, one batch was used for TGA analysis and 1 batch was diluted to 15mL and used for characterization, stability assays and cell experiments.

Characterization:

To determine particle concentration and hydrodynamic diameter, particles were diluted 100 times in DiH₂O, mixed, and injected into the Nanosight (Malvern Instruments). To determine zeta potential, 10 μL of particles were diluted into 1.4mL of DiH₂O and analyzed using ZetaPals (Brookhaven Instruments). UV absorbance was measured on Ultrospec 3000pro (GE Life Sciences). In order to visualize the aggregates, 2 μL of aggregates solution were dried onto a formvar stabilized 200 mesh copper carbon grid purchased from TED Pella. TEM images were taken using FEI Tecnai 12 Twin.

Composition Analysis:

In order to determine the amount of PTMP and PEG-maleimide, TGA and ICPMS were used. Concentrated particle solutions were dried in the platinum sample holder under vacuum at 105 °C. The sample was loaded into the sample pan of the TGA (Q50, TA Instruments) and the sample was heated to 200 °C and held for 30 minutes to completely dry the sample and then heated up to 800 °C at a rate of 20 °C /minute. The starting material weight was the weight at the end of the drying period and the final gold weight was the weight at 800 °C. For ICP-MS, each member of the nanoparticle library was synthesized on a 500 μL AuNP scale in elemental analysis grade water (Fisher Optima). Particles were all washed 3 times with elemental analysis grade water and resuspended in 500 μL of elemental analysis grade water. Elemental analysis grade water was required to reduce the background levels of sulfur that interfere with this study. 250 μL of each synthesized nanoparticle was added to a metal free tube (SCP Science) and were digested by adding 1 mL of 68% HNO3 (BDH Aristar) with 1% HCl (BDH Aristar). The sample was diluted to 5 mL with 2% HNO₃ 1% HCl solution and analyzed on an Agilent 8800 in oxygen mode to determine the amount of sulfur. 500 μL of that sample was further diluted to 10 mL to measure gold content. A standard curve was made using a serial dilution of a 1ppm sulfur standard (Spex Certiprep) and a 100 ppm gold standard (Spex Certiprep) in 1% HCl and 2% HNO₃. Data were analyzed quantitatively in a spreadsheet program.

Cyanide Stability:

225 μL of the nanoparticle solutions in water were aliquoted into plastic cuvettes containing 225 μL phenol free RPMI 1640 complete cell growth media with 10% FBS. Serum containing media was used to prevent the salt induced coating of the plastic cuvettes⁴⁵ and to more closely mimic the surface of nanoparticles in an in vitro system. The particle solution was mixed with 50μL aqueous 1M KCN and examined by UV-Vis at 5 minutes, 1h, 4 h, and 24 h. Warning, KCN should be used in the chemical hood, kept away from acids, and neutralized by mixing any solutions containing KCN with an excess of H₂O₂ in the hood.

Salt Stability:

Nanomaterials were aliquoted into plastic cuvettes. 50 μL 10X PBS (Potassium chloride 2 g/L, Monopotassium phosphate 2.4 g/L, Sodium chloride 80g/L, Disodium phosphate 14.4 g/L, Tris Ultrapure 24.2 g/L) was added to 450 μL nanoparticle solutions. Each cuvette was examined by UV-Vis at 5 minutes, 1 hour, 4 hours and 24 hours and compared to a sample diluted with H₂O.

Media Stability:

Nanomaterials were aliquoted into plastic cuvettes and mixed with an equal volume of phenol free RPMI 1640 media containing 10% serum. Particles were monitored by UV and DLS at 5 min, 1 hour, 4 hours and 24 hours.

Protein Isolation:

2 mL of diH₂O containing 1.8×10¹⁰ particles/mL was mixed with 1 mL of 30% FBS in 3x PBS and incubated at room temperature on a shaker for 2 hours. Protein was isolated as previously described¹³ with some modifications. Briefly, after 2 hours the particles were washed three times with 1x PBS with 0.05% v/v Tween-20 to remove free or loosely bound proteins while stabilizing the nanoparticles. For each wash, particles were pelleted by centrifugation (20,000 g for 20 minutes), the supernatant was discarded, and particles were resuspended in 750 μL 1x PBS with 0.05% v/v Tween-20. After the third wash, the supernatant was discarded and approximately 15μL volume remained in the tube. 8μL of 4x LDS buffer (Invitrogen) and 4μL of 500 mM DTT were added each vial and then incubated at 70°C for 1 hour to reduce and denature proteins on the surface of the particle. The reactions were centrifuged for 20 minutes at 18,000g to pellet the particles and the supernatant containing the free proteins was transferred to a new tube. At this point the total volume of the particles was variable due to slight differences in washing; all supernatants were diluted to a final volume of 50 μL for subsequent analysis.

Protein Gel:

6.5 μL of the protein solution was transferred to a new tube and mixed with 2.5 μL 4x LDS and 1 μL of 500 mM DTT before incubation at 70°C for 30 minutes. 4–12% Bis-Tris gels (Life Technologies) were loaded with the10 μL protein samples and resolved at 200V for 50 minutes. The protein ladder (PageRuler unstained protein ladder, ThermoFisher) was loaded in the first lane. The gel was removed from the cassette. The gel was fixed by two 30 minute incubations in fresh fixing solution (40% ethanol, 10% acetic acid). After decanting the fixing solution, the gel was rinsed in diH₂O for 5 minutes. The gel was immersed in 1X Krypton Protein Stain (Thermo Fisher) overnight. After decanting the protein stain, the gel was first immersed in a destaining solution (5% acetic acid) for 5 minutes and then immersed in fresh diH₂O for two 15 minute incubations. The bands were visualized using a Typhoon 9200 with the excitation set to 532 nm and emission at 580 nm.

Cell Counting via ICPMS:

As described previously,³¹ briefly, cells were grown to confluency, trypsinized and washed three times with 1x magnesium free PBS. Cells were counted using the hemocytometer and aliquoted into vials such that each vial contained 1 × 10⁶, 5 × 10⁵, 2.5 × 10⁵, 1 × 10⁵, 5 × 10⁴, 1 × 10⁴ or 1 × 10³ cells. Cells were frozen in −20°C before being digested with 1mL of Aqua Regia (warning extremely corrosive) and then diluted to 10mL with 2% HNO₃ for ICPMS analysis. Magnesium concentration was determined by ICPMS Analysis on Agilent 8800 Series using a concentric nebulizer, Scott type spray chamber, and torch with a fixed quartz injector torch. A CX interface was used. Plasma power was 1500 W. A standard curve was made using a serial dilution of a 1ppm magnesium standard (Spex Certiprep) and a 100 ppm gold standard (Spex Certiprep) in 1% HCl and 2% HNO₃. In order to determine cell number in later experiments a standard curve was generated that correlated the amount of magnesium and the number of cells in a vial.

Cellular Uptake and Time Course ICPMS:

RAW 264.7 (TIB- 71) cells were cultured in complete DMEM media (ATCC 30–2002, with 10% FBS, 1% Penicillin and Streptomycin). AML-12 (CRL-2254) cells were cultured in complete DMEM-F12 media (ATCC-30–2006 with 10% FBS, 1% Penicillin and Streptomycin). MDA-MB-231 cells were cultured in complete DMEM, high glucose media (Gibco with 10% FBS, 1% Penicillin and Streptomycin). For each experiment 50,000 cells were added to each well of a 24 well plate and left to attach for 24 hours in the incubator at 37°C and 5% CO₂. Each plate contained all nanoparticle treatment conditions for one cell line and one time point. At 24 hours, 32 hours, 44 hours or 47 hours after plating, the media was removed, and 490μL of fresh media and 10μL of nanoparticles (1.8 × 10¹⁰ particles/mL) were added to each well. At 48 hours after plating, the media containing the nanoparticles was removed and the wells were each washed three times with 1 mL magnesium free PBS. After the last rinse, the PBS was removed and the plates were placed into the −20°C freezer until ICPMS analysis. In order to determine the non-specific binding of the nanoparticles at 48 hours, cells were treated with nanoparticles and the media was immediately removed and the cells were washed three times with 1 mL magnesium free PBS. For ICPMS analysis, each well was washed with 500 μL of agua regia followed by 500μL 2% HNO₃ and diluted to 5 mL with 2% HNO₃. A standard curve was made using a serial dilution of a 1ppm magnesium standard (Spex Certiprep) and a 100 ppm gold standard (Spex Certiprep) in 1% HCl and 2% HNO₃. To determine the amount of nanoparticles specifically endocytosed, the amount of gold measured in the non-specific uptake was subtracted from the experimental results.

Cellular uptake Study – TEM:

RAW 264.7 (TIB- 71) cells were cultured in complete DMEM media (ATCC 30–2002, with 10% FBS, 1% Penicillin and Streptomycin). AML-12 (CRL-2254) cells were cultured in complete DMEM-F12 media (ATCC-30–2006 with 10% FBS, 1% Penicillin and Streptomycin). MDA-MB-231 cells were cultured in complete DMEM, high glucose media (Gibco with 10% FBS, 1% Penicillin and Streptomycin). For each experiment 5 × 10⁶ cells were plated in a T75 tissue culture flask and left to attach for 24 hours in the incubator at 37°C and 5% CO₂. After 24 hours to attach, cells were given 10 mL of fresh media and treated with a final concentration of 3.6 × 10⁸ particles/mL. Particle concentration was determined from the Nanosight measurement. After 24 hours at 37°C, the media containing excess nanoparticles was removed. The cells were washed with 5 mL of 1x PBS. Cells were removed from the flask using 0.25% Trypsin for 5 minutes at 37°C. The trypsinization was terminated by the addition of fresh media. The cell-media mixture was centrifuged at 130g for 5 minutes to pellet the cells. The cell pellet was washed with 10 mL of 1x PBS and centrifuged to pellet. The cell pellet was then resuspended in 1 mL of EM grade 2% glutaraldehyde in 0.1 M Cacodylate buffer at pH 7.4. The fixed cells were stored overnight at 4°C. The next day the cells were embedded in 10% gelatin. After the gelatin became firm, the gelatin was minced and fixed in 2% glutaraldehyle in 0.1 M Cacodylate buffer at pH 7.4 for 30 min at room temperature. The cell pellets were washed three times with 0.1 M Cacodylate buffer, pH7.2, post-fixed with 1% OsO₄ in 0.1 M Cacodylate buffer for 30 min and washed three times with 0.1 M Cacodylate buffer. The samples were then dehydrated through 60%, 70%, 80%, 95% ethanol, 100% absolute ethanol (twice), propylene oxide (twice), and were left in propylene oxide/Eponate (1:1) overnight at room temperature. The vials were sealed. The next day the vials were left open for 2–3 hours to evaporate the propylene oxide. The samples were infiltrated with 100% Eponate and polymerized at ~64°C for 48 hours. Ultra-thin sections (~70 nm thick) were cut using a Leica Ultra cut UCT ultramicrotome with a diamond knife and picked up on 200 mesh copper EM grids. Grids were stained with 2% uranyl acetate for 10 minutes followed by Reynold’s lead citrate staining for 1 minute. Electron microscopy was done on an FEI Tecnai 12 transmission electron microscope equipped with a Gatan Ultrascan 2K CCD camera. The TEM was operated at 120 KeV.

Cell Uptake -- Serum Study:

For each experiment 50,000 cells were added to each well of a 24 well plate and left to attach for 24 hours in the incubator at 37°C and 5% CO₂. After 24 hours, complete media was removed, each well was washed with 1mL of PBS, and 490μL of fresh media (with or without serum) and 10μL of nanoparticles (1.8 × 10¹⁰ particles/mL) were added to each well. At 48 hours after plating, the media containing the nanoparticles was removed and the wells were each washed three times with 1 mL magnesium free PBS. After the last rinse, the PBS was removed and the plates were placed into the −20°C freezer until ICPMS analysis. In order to determine the non-specific binding of the nanoparticles, cells were incubated in media (with or without serum) for 24 hours before nanoparticles were added. The media containing nanoparticles was immediately removed and the cells were washed three times with 1 mL Magnesium free PBS. After the last rinse, the PBS was removed and the plates were placed into the −20°C freezer until ICPMS analysis. For ICPMS analysis, 500μL of 68% HNO₃ with 1% HCl was added to each well and transferred to a 15 mL metal free tube (SCP Science). Samples were then diluted with 4.5mL of 2% HNO₃ with 1% HCl. A standard curve was made using a serial dilution of a 1ppm magnesium standard (Spex Certiprep) and a 100 ppm gold standard (Spex Certiprep) in 1% HCl and 2% HNO₃. Data were analyzed quantitatively in a spreadsheet program.

Cell Uptake Inhibitor Study:

Experiments were completed in a 6 well plate. One million RAW 264.7 cells were plated in each well and left to adhere overnight in complete DMEM media (ATCC 30–2002, with 10% FBS, 1% Penicillin and Streptomycin). After 24 hours, complete media was removed, each well was washed with 1mL of PBS, and 2 mL DMEM basal media (ATCC 30–2002) with inhibitor was added to each well at the concentrations in Table 2. After 1 hour at 37°C in the presence of the inhibitor 40 μL gold nanoparticles in water (1.8 × 10¹⁰ nanoparticles/mL) were added to each well for a final concentration of 3.6 × 10⁸ nanoparticles/mL. Plates were exposed to nanoparticles for 1 hour at 37°C, the media containing the nanoparticles was removed, and the wells were washed three times with 1 mL magnesium free PBS per wash. In order to determine the non-specific binding of the nanoparticles, cells were treated with inhibitor as above, but the gold nanoparticles were added after 2 hours of exposure to the inhibitor, the media containing nanoparticles was immediately removed, and the particles were washed three times with 1 mL Magnesium free PBS. After the last rinse, the PBS was removed and the plates were placed into the −20°C freezer until ICPMS analysis. For ICPMS analysis, 500μL of 68% HNO₃ with 1% HCl was added to each well and transferred to a 15 mL metal free tube (SCP Science). Samples were then diluted with 4.5mL of 2% HNO₃ with 1% HCl. A standard curve was made using a serial dilution of a 1ppm magnesium standard (Spex Certiprep) and a 100 ppm gold standard (Spex Certiprep) in 1% HCl and 2% HNO₃. In order to calculate the amount of gold in the cells at 1 hour, the average amount of Au in the non-specific control samples was subtracted from the amount of gold measured in the experimental group for each inhibitor.

Table 2:

Inhibitor Dose

Inhibitor	Dose
NaN3	10mM39
Genistein	37 uM37
Chlorpromazine	30 uM37
Cytochalasin D	4 uM37
Nocodazole	10 uM40
Wortmannin	0.1 uM40
Poly I	5 mg/mL41
Methyl – B – Cyclodextran	20 uM38