Detection of silver nanoparticles in cells by flow cytometry using light scatter and far-red fluorescence

Authors

  • R. M. Zucker,

    Corresponding author
    • U.S. Environmental Protection Agency, Office of Research and Development, National Health and Environmental Effects Research Laboratory, Toxicology Assessment Division (MD-67), Research Triangle Park, NC 27711
    Search for more papers by this author
  • K. M. Daniel,

    1. U.S. Environmental Protection Agency, Office of Research and Development, National Health and Environmental Effects Research Laboratory, Toxicology Assessment Division (MD-67), Research Triangle Park, NC 27711
    2. Contractors to the USEPA
    Search for more papers by this author
  • E. J. Massaro,

    1. U.S. Environmental Protection Agency, Office of Research and Development, National Health and Environmental Effects Research Laboratory, NC
    Search for more papers by this author
  • S. J. Karafas,

    1. U.S. Environmental Protection Agency, Office of Research and Development, National Health and Environmental Effects Research Laboratory, Toxicology Assessment Division (MD-67), Research Triangle Park, NC 27711
    2. Contractors to the USEPA
    Search for more papers by this author
  • L. L. Degn,

    1. U.S. Environmental Protection Agency, Office of Research and Development, National Health and Environmental Effects Research Laboratory, Toxicology Assessment Division (MD-67), Research Triangle Park, NC 27711
    Search for more papers by this author
  • W. K. Boyes

    1. U.S. Environmental Protection Agency, Office of Research and Development, National Health and Environmental Effects Research Laboratory, Toxicology Assessment Division (MD-67), Research Triangle Park, NC 27711
    Search for more papers by this author

  • Disclaimer: Research described in this article was supported by the United States Environmental Protection Agency; it has been subjected to Agency review but does not necessarily reflect the views of the Agency. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.

Correspondence to: R.M. Zucker; U.S. Environmental Protection Agency, Office of Research and Development, National Health and Environmental Effects Research Laboratory, Toxicology Assessment Division (MD-67). E-mail: zucker.robert@epa.gov

Abstract

The cellular uptake of different sized silver nanoparticles (AgNP) (10, 50, and 75 nm) coated with polyvinylpyrrolidone (PVP) or citrate on a human derived retinal pigment epithelial cell line (ARPE-19) was detected by flow cytometry following 24-h incubation of the cells with AgNP. A dose dependent increase of side scatter and far red fluorescence was observed with both PVP and citrate-coated 50 nm or 75 nm silver particles. Using five different flow cytometers, a far red fluorescence signal in the 700–800 nm range increased as much as 100 times background as a ratio comparing the intensity measurements of treated sample and controls. The citrate-coated silver nanoparticles (AgNP) revealed slightly more side scatter and far red fluorescence than did the PVP coated silver nanoparticles. This increased far red fluorescence signal was observed with 50 and 75 nm particles, but not with 10 nm particles. Morphological evaluation by dark field microscopy showed silver particles (50 and 75 nm) clumped and concentrated around the nucleus. One possible hypothesis to explain the emission of far red fluorescence from cells incubated with silver nanoparticles is that the silver nanoparticles inside cells agglomerate into small nano clusters that form surface plasmon resonance which interacts with laser light to emit a strong far red fluorescence signal. The results demonstrate that two different parameters (side scatter and far red fluorescence) on standard flow cytometers can be used to detect and observe metallic nanoparticles inside cells. The strength of the far red fluorescence suggests that it may be particularly useful for applications that require high sensitivity. © Published 2013 Wiley-Periodicals, Inc.

The rapid development and commercialization of man-made nanomaterials have outpaced information regarding the potential hazards of these materials to the environment, humans, or other organisms [1]. This situation has prompted the formation of a National Nanotechnology Initiative that coordinates the efforts of the federal government to enable safe and responsible development of these promising new technologies, and the creation of an associated research strategy for the US Environmental Protection Agency (EPA) Office of Research and Development [2]. Nanoparticles may pose unique health risks beyond those created by larger particles of the same material due to their reactivity, small sizes, and increased surface areas [2-4]. Research on potential hazards of manufactured nanomaterials presents many technical challenges, one of which is a limited ability to detect and quantify nanoparticles in environmental media, tissues, or cells that may have been exposed to nanomaterials. The development of techniques to identify and characterize nanoparticles in cells and various media would be an asset for evaluating potential risks of manufactured nanomaterials.

Key to evaluating the potential hazard of man-made nanomaterials is the observation and measurement of nanoparticles in cells. A variety of techniques have been used to detect these particles in the environment, including transmission electron microscopy [5], scanning electron microscopy [6], and atomic force microscopy [7, 8]. While these techniques have been useful in determining the physical properties of nanoparticles, they have limitations including expense, limited availability and capacity, and lack of quantitative information regarding cellular interactions with nanoparticles.

Silver nanoparticles are being used in a variety of commercial products that could lead to unintended exposures of people or the environment [9-11]. The development of sensitive, rapid and economical methods to detect and quantify nanoparticles in cells, tissues and in environmental samples such as soil or water would help to evaluate their potential risks [2, 12, 13]. Previously, we have shown that flow cytometry and darkfield microscopy can detect different types of TiO2 nanoparticles in a human derived retinal pigment epithelial cell line (ARPE-19) [2, 14, 15] and other adherent cell types (unpublished). ARPE-19 cells incubated with TiO2 nanoparticles for 24 h showed a decreased forward scatter and increased side scatter in a dose-dependent manner between 1 and 100 μg/ml. Under darkfield microscopy, the TiO2 particles were observed inside the cytoplasm of ARPE-19 cells in peri-nuclear agglomerations that grew in size and number with increasing exposure concentration [2, 15].

The current experiments evaluated whether the presence of silver nanoparticles inside APRE-19 cells could also be detected using darkfield microscopy and flow cytometry [15]. Different sizes of silver nanoparticles (AgNP, 10, 50, and 75 nm) coated with either citrate or polyvinylpyrrolidone (PVP), were evaluated. Similar to TiO2, the incubation of different sizes of AgNP with APRE-19 cells resulted in a dose and size-related increase in flow cytometry side-scatter signal showing the relative uptake of the particles [2, 16, 17]. During the course of these experiments the 50 and 75 nm AgNP incubation resulted in an increased far-red fluorescence which was detected on several flow cytometers manufactured by different companies. This far red signal was not observed in a similar experiment using either TiO2 NP or submicron particles. We hypothesized that silver nanoparticles entered cells whereby they clumped and generated surface plasmon resonance (SPR) that interacted with intercellular fluorescence molecules to emit fluorescence in the far red wavelengths following the laser beam excitation.

Materials and Methods

Cell Culture

For microscopic observations, a human derived retinal pigment epithelial cell line (ARPE-19; ATCC, Manassas, Virginia) were plated in T75 culture flasks in a 1:1 mixture of Dulbecco's Modified Eagle's Medium and Ham's F-12 Nutrient Mixture (DMEM/F-12) with 10% fetal bovine serum (FBS). Cells were grown to confluence, trypsinized (0.05% trypsin, EDTA 0.02%, Sigma), and plated on chambered glass tissue culture slides (1 ml cell suspension per chamber, 2×105 cells/ml). After plating, cells were incubated for 24 h (37°C, 5% CO2) without nanoparticles, and then treated with nanoparticles for a further 24 h before staining and fixation [2, 14].

Morphology

The cells were transfected with organelle lights, Golgi GFP BacMam 2.0 (C10592) or endoplasmic reticulum ER-GFP BacMam (C10590) to identify structures within the cells, and were counterstained with CellMask™ Orange plasma stain (Invitrogen, Eugene Oregon, C10045) to indentify the cell cytoplasm area (Invitrogen, Eugene Oregon). The cells were fixed with warm 4% paraformaldehyde made in phosphate-buffered saline (PBS) and mounted with Prolong Gold with DAPI (P36935). After the mounting medium dried the slides were sealed and then observed with a combination of dark field and fluorescence microscopy. Pictures were taken with a 60x objective with an iris diaphragm.

For flow cytometry evaluation, ARPE-19 cells were plated in seven T25 culture flasks (5×104 cells/ml) in DMEM/F-12 with 10% FBS. Cells in flasks were incubated for 24 h without nanoparticles, and then treated with AgNP for 24 h. The proliferative cells were then trypsinized and centrifuged. Two milliliters of media were added to the pellet and placed on ice prior to flow cytometry analysis.

Silver Nanoparticles

ARPE-19 cells were incubated with one of three sizes (10, 50, or 75 nm) AgNP. The silver was coated with either citrate or PVP at concentrations between 1 and 100 μg/ml (NanoComposix San Diego CA).

TiO2, Nanoparticle Suspension, and Treatment

Three different types of TiO2 primary particles were evaluated including: 10 nm (Alpha Aesar), 25 nm (Degussa P25), and 200–400nm (MKN-TiO2). They were sonicated as previously described and added to cells in a similar manner to that described previously by our laboratory [14, 15].

Staining, Fixation, and Mounting

Cell viability was determined by detecting intact or compromised membranes using calcein-AM (green, live) and Propidium iodide (PI red, dead). A small aliquot of cells, usually 50-100 uL, were fixed with an equal amount of 4% paraformaldehyde (PF) in PBS for morphology studies. To visualize the nuclei, organelles, cell membranes and nanoparticles in the cells, the fixed cells were stained with DAPI, cell membrane orange, and transfection probes for the Golgi and endoplasmic reticulum (ER).

Microscopy

A Nikon E-800 microscope was used to observe dark field and fluorescence images. This microscope had space for five filter cubes. The fluorescence excitation cubes were for DAPI, FITC, TRITC, and Cyan GFP. The fifth space in the cube holder was intentionally left without a cube in order to acquire a clean dark field image without distortion from filters. However, the dark field image was bright and it could be observed through any of the four filter cubes. The dark field image was about 100 times brighter than the fluorescence image as measured by comparing their exposure times. The fluorescent and dark field images were obtained sequentially and then the images were combined with Nikon Elements 3.01 software. Co-localization of the optical system was established with 0.5 μm Tetra spec beads and also with 1 μm and 15 μm multi-wavelength ring beads (Molecular Probes, Eugene Oregon). The Xenon light source was used to acquire the dark field images. This light source has shorter wavelength excitation and is brighter than halogen light which should provide better resolution from the sample cells. A GG420 filter was put in the eyepieces to protect the user's eyes from possible UV damage from the Xenon light source that emits light below 420 nm [15].

The most suitable Nikon lens when using a Nikon infinity-corrected microscope was a 60x Plan Fluor with an iris diaphragm to control the numerical aperture (NA) between 0.55 and 1.25. The lens had a sufficiently large magnification to observe cellular details while the background scatter could be controlled by adjusting the diaphragm. The lower NA yielded good depth of field for bright nanoparticles, and the higher NA (1.25) yielded bright fluorescence images of the cellular components. By balancing the fluorescence and dark field signals, a sequential image could be acquired with the same NA setting (approximately 0.8 NA). During the course of this study, the dark field images were obtained using the following dry lenses: Plan Apo 20x, (NA 0.75), Plan Apo 40x (NA 0.95), and Achromat 60x (NA 0.8) and the following oil lenses: 20x multi-immersion (MI, NA 0.75), Achromat 100x with iris (NA 1.25-0.55), and 60x Plan Fluor with iris (NA 1.25-0.55). NA values had to be below 0.95 with an oil darkfield condenser to provide proper illumination for good dark field images of nanoparticles. Most of the microscopy was accomplished by first observing the field with a 20x oil lens and then obtaining images with the 60x Plan Fluor oil lens with an iris adjustment to optimize both the fluorescence and dark field images [15].

Confocal images of nanoparticles within cells were made using a Leica SP1 confocal microscope as previously described [15, 18]. Briefly, a Plan Apo 63x lens with an iris was used with an air darkfield condenser in the light path on the Leica SP1. Fluorescence was acquired from an argon krypton laser emitting either a 488 or 568 nm lines. The configuration allowed enlargement of areas where the nanoparticles were located. Magnifications as high as 10,000x were achieved with this confocal microscope.

Flow Cytometry

Calibration and standardization

A BD FACSCalibur™ (BD Biosciences, San Jose, CA) flow cytometer containing a 488-nm laser, forward scatter (FSC) diode detector, and photomultiplier tube (PMT) SSC detector was used initially in this study. This instrument was checked daily for fluidic alignment using Molecular Probes 2.5-μm alignment beads (A-7302) for 488 lasers and Thermo Fisher 3um alignment beads (Cyto Cal plus violet (FC3AVL). The coefficient of variation for the Molecular Probes beads was under 2.5% for the fluorescent channels and under 3% for the scatter channels. Because the flow rate affects these measurements, these tests were performed at low flow rates. If the coefficient of variation was higher than 2.5% for the fluorescence channels, the cytometer was cleaned with 10% Clorox bleach followed by 10-50% Contrad 70. The cytometer was set up to measure SSC logarithmically and FSC linearly.

Gating logic

The gating logic consisted of counting cells within a large SSC (log) vs. FCS (linear) cytogram. This gating logic of SSC and FCS was used to eliminate the small debris particles from the histograms. The scatter and fluorescent histogram parameters were derived from this gated scatter region. Most of the samples contained minimal debris for low influence on the mean of the histogram distribution. The system was set up to count cells and trigger on FSC. Dynamic ranges of the PMTs were optimized to show maximum changes for the AgNP and TiO2 NP doses that were used. The highest dose of nanoparticles was measured first to set the range for the maximum SSC and far red fluorescence signals. The robust geometric means and standard deviations of the log histogram distributions were obtained for data analysis.

Data Analysis

To compare the changes in scatter or far red fluorescence between control and AgNP treated samples, the data were evaluated by one of two methods. The first method was a simple ratio calculated by dividing the scatter or fluorescence value of the treated sample by the value of the control. This is considered to be the mean fluorescence intensity method (MFI).

The second method was the stain index (SI) [19] or the separation index method [20]. The data are shown in Tables 1 and 2 and used as a comparison to the MFI ratio. In this SI method the difference between positive and negative populations is divided by two times the standard deviation (SD) of the negative population. Maecker describes the effective brightness of a reagent by this method which depends on the difference between the positive and the negative populations and the spread of the negative population. The stain index is a useful metric for the normalized signal over background. This normally defined stain index has been modified for our experiments as we measured only side scatter or a change in far red fluorescence after incorporation of silver nanoparticles into the cells.

Table 1. Comparison of 3 sizes of AgNP (10, 50, and 75 nm PVP coated AgNP) incubated at different doses (0, 10, 30, and 55 μg/mL) with APRE-19 cells for 24 h
NP size (nm)[Ag NP] μg/mLFSCSSCFL1(530/30)FL2(585/42)FL3(>670)SI FL3(>670)
  1. The scatter and fluorescence data (Fl1, Fl2, and Fl3) derived from a FACSCalibur is expressed as the intensity ratio of treated/control (MFI). As a comparison, the stain index described by Maeckler et al. [19] is shown for the far red fluorescence data (F3 > 670 nm). Both 50 nm and 75 nm AgNP show large far red fluorescence while only a minimal effect in the far red fluorescence channel was observed for 10 nm AgNP.

 Control1.001.001.001.001.001.00
 101.021.281.001.012.231.01
10 nm300.912.211.390.932.040.84
 550.842.351.171.391.780.67
 Control1.001.001.001.001.001.00
 100.932.151.041.353.272.39
50 nm300.902.971.201.568.8810.18
 550.783.671.351.779.0610.41
 Control1.001.001.001.001.001.00
 100.952.721.011.072.411.59
75 nm300.873.941.131.3810.2410.76
 550.775.441.341.8114.2117.05
Table 2. Comparison of a Stratedigm S1000 with a FACSCalibur
 StratedigmFACSCalibur
 488 nm Laser live405 Laser live488 Laser fixed488 Laser live
[Ag NP] μg/mlSSCFL5(>740)FL5(>740)FL5(>740)SSCFL3(>670)
  1. Data are shown for SSC and far red fluorescence after incubation with different concentrations of 75 nm PVP-coated AgNP (0.5, 1, 2.5, 5, 10 μg/ml) for 24 h. Data are expressed both using intensity ratio method (A, MFI, ratio of treated/control) and the stain (separation) index (B, SI=Mean (treated)–Mean (control)/ (2x SD) control. The cells were excited with a 405 nm or 488 nm lasers on Stratedigm S1000 and with a 488 nm laser on the FACSCalibur. The cells were fixed with 4% PF and then measured on both of these instruments. The same populations of live or fixed cells were used in both flow cytometers. The 5 μg/ml sample had insufficient number of cells for some of the measurements and therefore they are reported as not done (Nd). The data illustrate that the far red fluorescence effect is greater using live cells with 488 nm laser exaction compared to the 405 nm laser excitation or with fixed cells using a 488 nm laser excitation. . Fixation and use of the 405 nm laser reduced the far red fluorescence by ∼90%.

A. Intensity ratio
01.001.001.001.001.001.00
0.51.531.341.561.221.261.18
11.881.631.821.381.391.3
2.52.497.892.481.671.893.65
52.7624.56NdNdNdNd
103.7589.353.62.512.7238.43
B. Stain index
01.001.001.001.001.001.00
0.50.890.440.000.390.260.02
11.480.810.060.520.380.04
2.52.508.870.211.580.780.33
52.9530.33NdNdNdNd
104.63113.712.375.721.664.69

Instruments: filters and lasers FACSCalibur

The FACSCalibur (BD Biosciences, San Jose, California) contained a 15 mW 488 laser and the following band pass filters: FL1- 530/30, FL2- 564/42, and FL3- 670LP, respectively [2, 17].

A LSR11 (BD Biosciences, San Jose, CA) contained the following three lasers: 405 nm (25 mW), 488 nm (20 mW), and 633 nm (20 mW). It was used primarily to confirm the FACSCalibur results. The 488 nm excitation was detected between 500 nm and 810 nm using the following five different filters: 530/30, 575/26, 610/20, 695/40, and 780/60. The 633 nm excitation was detected between 650 nm and 810 nm using the following three filters: 675/20, 730/45, and 780/60. The 405 nm excitation was detected using the following two filters: 440/40 and 525/50.

Stratedigm S1000 (Stratedigm, San Jose, CA) flow cytometer was set up to trigger on forward scatter using the 488 nm (50 mW) laser excitation. In some experiments, light scatter was collected from the 405 nm (100 mW) laser in addition to the 488 nm excitation. In this Stratedigm system the same 6 PMTs were used to collect emission from both the 405 nm and 488 nm excitation as these laser lines were spatially separated and the fluorescence from the two lasers did not interfere with each other. The filters in the Stratedigm 1000 system that are used for both lasers were the following: 450/40, 530/30, 580/30, 615/30, 690/40, and 740 long pass (LP).

An Attune flow cytometer (Life Technologies, Carlsbad, CA), containing 405 nm (50 mW) and 488 nm (20 mW) lasers was used to detect fluorescence for each laser in 3 PMTs. The filters for the 488 were 530/30 575/24 and a 640LP. The filters for the 405 nm laser were 450/40 522/30 and 603/48 nm. The last PMT with the 488 nm laser was a 640 nm LP, and the last PMT with the 405 nm laser was a 603/48 bandpass filter which inhibits light in the far red range.

Results

Different sizes and types of AgNP were studied on a FACSCalibur by incubating ARPE-19 cells for 24 hours with AgNP at concentrations between 0 and 100 μg/ml. The flow cytometry data showed a dose dependent increase in side scatter and far red fluorescence (>670 nm). Because we observed changes from control at the lowest doses levels initially tried, we evaluated the sensitivity of the far red signal between 1 and 20 μg/ml using the 75 nm Ag nanoparticles (Fig. 1). The results showed a 4.3-fold concentration dependent increase in side scatter while the far red fluorescence increased 39-fold. (Supporting Information Table 1 corresponding to Fig. 1).

Figure 1.

Side scatter (SSC-H; panel A) and far red fluorescence (FL >670; panel B) of ARPE-19 cells exposed to 0, 1, 3, 10, 15, or 20 μg/ml of 75 nm PVP AgNP for 24 h. The 15 μg side scatter point is not presented for figure clarity. There was a large increase in fluorescence between 3 μg/ml and 10 μg/ml, presumably due to AgNP aggregation inside the cell when used at and above 10 μg/ml (Fig. 4).

There was a large non-linear increase in far red fluorescence between 3 μg/ml and 10 μg/ml but a proportionately smaller apparently linear increase in SSC (Fig. 1). At low doses of AgNP, the fluorescence initially increased slowly, but at higher concentrations, the far red fluorescence signal increased more rapidly than did the side scatter (Supporting Information Fig. 2 Supporting Information Table 1). This observation was repeated in all of our experiments using doses between 1 and 20 μg/ml. The data suggested that the particles contained in the cells might be interacting with light differently in the scatter channels and the far red fluorescence channels to give this different response. A similar effect was observed with the 50 nm and 75 nm silver nanoparticles using broader dose ranges (1–100 μg) (Supporting Information Fig. 1 and Supporting Information Table 2). The 10 nm particles differed from the 50 and 75 nm particles as they did not show a dose dependent increase in far red fluorescence and the magnitude of the fluorescence was much less than that observed with the larger particles.

A comparison of three sizes (10, 50, and 75 nm) AgNP-PVP coated is shown in Table 1. The SSC and fluorescence parameters increased after treatment with successively higher doses of AgNP-PVP. The increase in fluorescence was smaller at lower wavelengths, (PMTs Fl1 (530/30) and Fl2 (585/42)) than at higher wavelengths (Fl3 > 670 nm). At the highest doses, the relative increase FL3 (>670 nm, far red) fluorescence measured on FACSCalibur was 9 or 14 times larger than control values for the 50 nm and 75 nm AgNP-PVP, respectively. In contrast, the 10 nm PVP particles showed less than a two fold increase in far red fluorescence over the same dose ranges. The stain index method showed a 10- and 17-fold increase at the 55 μg/ml dose for the 50 and 75 nm particles, respectively.

In order to evaluate the influence of particle coating on cellular uptake, AgNP-PVP were compared to AgNP-citrate at a concentration of 30 μg/ml. Both types of AgNP showed a concentration-dependent increase in SSC and far red fluorescence (Supporting Information Table 3). The FSC and SSC parameters were similar for the two coating types. Both types of AgNP also exhibited increased SSC and decreased FSC with increasing particle size. Interestingly, 10 nm AgNP with either coating showed only a slight increase in the far red fluorescence. The far red fluorescence increase was greater for 75 nm than 50 nm particles, and it was greater for citrate than for similarly sized PVP coated particles (Supporting Information Table 3).

The amount of SSC (Fig. 2A) and far red fluorescence (Fig. 2B) varied as a function of particle size and concentration. For the same doses, the larger particles showed greater light scatter and far red fluorescence. The 10 nm particles showed the least SSC and almost no far red fluorescence. AgNP-citrate showed similar SSC values but higher far red fluorescence values relative to the AgNP-PVP (data not shown). The 10 nm silver particles showed lower SSC than the 50 nm or 75 nm particles, and almost no increase in far red fluorescence. Data shown in Figure 2 were limited to applied concentrations of less than or equal to 30 μg/ml, and fit with a linear regression (SigmaPlot for Windows Version 12.2.0.45, Systat Software, Inc. San Jose CA). When data were included at concentrations greater than 30 μg/ml, the curves became visibly nonlinear and appeared to saturate at higher concentrations. Previously, we fit linear functions to the relationship between TiO2NP concentration and SSC for doses of TiO2NP up to 30 μg/ml [2]. Above 30 μg/ml of AgNP the functions became asymptotic, suggesting that higher concentrations of AgNP approached the material limits that the cells could absorb.

Figure 2.

Data derived from the FACSCalibur machine with only PVP-coated Ag particles plotted. Side-scatter values are presented in Panel A, and far-red fluorescence in Panel B. Separate plots are provided for 10, 50, and 75 nm particles. Data were limited to concentrations less than or equal to 30 μg/ml. Linear regressions were fit to each plot using SigmaPlot (SigmaPlot for Windows Version 12.2.0.45; Systat Software, Inc. San Jose CA). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

The observation of a large far red fluorescence increased after treatment with AgNP was unusual, and therefore we decided to test four different flow cytometers in our geographical area to ascertain whether this observation was unique to the optical configuration of this specific FACSCalibur, or if it could be reproduced in other flow cytometers from different manufacturers with different configurations of lasers and detectors. The aim of using multiple flow cytometers was not to compare the efficiency of the filters in the flow cytometers or to compare fluorescence intensity between cytometers, but to demonstrate that the far red fluorescence effect was reproducible with the 5 different flow cytometers. It is not easy to compare different flow cytometers with different filters and different laser powers. A direct comparison of two different flow cytometers (FACSCalibur and Stratedigm 1000) in one laboratory (EPA) is presented in Table 2. These experiments were done in the author's laboratory where the instrument's performance was calibrated and standardized using equivalent procedures [2, 15]. At 10 μg/ml the MFI method demonstrated an 89 fold increase on the Stratedigm and a 38 fold increase on the BD FACSCalibur in the far red fluorescence channel. The comparable stain index showed a 114 fold increase on the Stratedigm and a 4.7 fold increase on the FACSCalibur. Data for all the channels are presented. Values for the stain index with many doses and parameters were less than 1, but in almost all cases there was a dose dependent increase in values using both methods of analysis.

In addition to the above flow cytometers located in the authors laboratory, the following flow cytometers were also tested: a 2nd BD FACSCalibur, an Invitrogen Attune, and a BD LSR11 (Supporting Information Tables 4, 5). All of these flow cytometers showed a concentration-dependent increase in far red fluorescence and side scatter. The far red fluorescence was detected with either red (633 nm) or blue (488 nm) laser excitation (Fig. 3 and Supporting Information Tables 4, 5.) Because the LSR11 had 5 filters covering the fluorescence range of 500 nm to 810 nm, it was possible to detect fluorescence in different ranges with greater accuracy than with the FACSCalibur, which had only 3 detectors across the equivalent range. Essentially a spectral response could be measured with the 5 filters across the 500 nm-810 nm range. In the LSR11, a similar amount of the far red fluorescence was detected in the far red fluorescence channel with both the 633 nm and 488 nm excitation. The maximum signal with the LSR 11 occurred using a 750-810 nm band-pass filter (Fig. 3, Supporting Information Table 5). The 405 nm laser on the LSR11 or the Attune had band pass filters that limited fluorescence detection to wavelengths below 650 nm, and thus a large increase in fluorescence was not observed. However, using the Stratedigm optical path 100 mW 405 lasers could be used to detect far red fluorescence. The far red fluorescence observed from a 100 mW 405 nm laser was over 90% decreased compared to the fluorescence observed from a 50 mW 488 nm laser. (Table 2) From identical samples the Stratedigm flow cytometer was able to detect over twice the fluorescence in the far red channels as that observed from the FACSCalibur with the 488 nm excitation (Table 2). This was likely due to a combination of a better optical path and a stronger laser (50 mW vs. 15 mW) of the Stratedigm compared with that of the FACSCalibur.

Figure 3.

LSRII detection of control cells and cells treated with 20 μg/ml of 75 nm PVP-coated AgNP using both 488 nm and 633 nm lasers. Both lasers show a maximum increase in the far red range (>670 nm). The maximum intensity was observed with the 750–810 bandpass in the Cy7 range for both lasers. In the histograms the red is the Ag treated cells while the black is the control cells for each parameter. The 488 nm laser fluorescence was detected in 5 PMTs while the 633 laser was detected in 3 PMTs. For clarity of the figure only the PE (585/42) and PerCP-CY55 (695/40) and PE–Cy7 (780/60) are shown for the 22 mW 488 nm excitation (PMT's 2, 4 and 5). The Alexa 700 and APC-Cy7 (780/60) are shown for the excitation with the 25 mW 633 laser. Note that there is almost no fluorescence increase below 600 nm (PE-A). Side scatter (SSC) is shown to illustrate the small magnitude of change compared to the larger change with the fluorescence parameters. The histogram data are presented as an area measurement. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

The far red fluorescence that was observed in these experiments appeared to be related to the physiology of live cells. Fixing the cells with paraformaldehyde caused over a 90% decrease in the far red fluorescence. Similar to the LSR11 spectral response, the Stratedigm 1000 flow cytometer with 5 PMTs and a 488 nm laser was used to detect different fluorescent emission across the 500 nm to 800 nm spectral ranges. In a similar manner to the LSR2, the biggest increase in fluorescence occurred from the PMT that collected light in far red fluorescence PMT (>700 nm). This far red fluorescence phenomenon obtained with AgNP was compared to 3 different types of TiO2NP. It was observed that there was no increase in far red fluorescence with any of the TiO2NP samples. It appeared that the increased far red fluorescence was unique to AgNP and did not occur with TiO2 nanoparticles (Supporting Information Table 6).

Morphology

AgNP sedimented onto ARPE-19 cells and then apparently passed through the cell membranes into the cytoplasm. Cells were transfected with Golgi GFP or endoplasmic reticulum GFP and counterstained with cell membrane orange to identify intracellular structures. Once the nanoparticles entered a cell, they appeared to migrate to the endoplasmic reticulum (ER) and surround the nucleus. In many cells, the nucleus was seen as a distinct object that was outlined with nanoparticles (Fig. 4). The particles did not appear to penetrate into the nucleus. The particles also appeared to aggregate within the cytoplasm (Fig. 4). The amount of aggregation was dependent on the concentration and size of particles added to the cultures (Fig. 4, Supporting Information Fig. 3). When examined by dark field microscopy, ARPE-19 control cells contained no particles, while AgNP were observed in the cells at all doses tested (3, 10 and 30 μg/ml). A comparison between the 3, 10 and 30 μg/ml illustrated that there were more particles and they were more clumped at the higher concentrations (Fig. 4A). At the lowest doses (3 μg/ml), fewer aggregates were observed in the cytoplasm of each cell (Fig. 4B). More aggregates were clearly discernible at 10 μg/ml (Fig. 4C) and 30 μg/ml (Fig. 4D) than at 3 μg/ml.

Figure 4.

Combined dark field and fluorescent images of ARPE-19 cells incubated in slide chambers for 24 h with three different concentrations (3, 10, and 30 μg/ml) of 75 nm AgNP. Each panel {control (A), 3 μg/ml (B) 10 μg/ml (C), and 30 μg/ml (D) show successive higher concentrations of white nanoparticles with higher doses. Cells were fixed with 4% PF and mounted with Prolong Gold (Molecular Probes) containing 10 μg/ml DAPI. The cell was stained with cell mask plasma membrane orange stain (C10045). Cells and nanoparticles are pseudo colored with cell nuclei appearing blue, Golgi appearing green, cell cytoplasm appearing orange, and AgNP or larger coarse aggregates appearing as white dots/spots. A Plan Fluor 60x lens with an iris diaphragm that could be adjusted between 0.55 and 1.25 NA to reduce background for dark field imaging was used. Images of fluorescent cells and dark field AgNP were acquired sequentially and then combined using Nikon Elements 3.0 software. Magnification 600x.

A comparison between the 10 nm and the 75 nm particles at the same dose (10 μg/ml) showed that the 75 nm particles demonstrated a brighter signal relative to the 10 nm particles by microscopy, presumably due to their larger size and/or greater agglomeration (Supporting Information Fig. 3).

Discussion

Far Red Fluorescence

This study demonstrated that silver nanoparticles (50 and 75 nm) induced a strong far red fluorescence signal between approximately 700–800 nm in response to laser stimulation of cells in a flow cytometer. One hypothesis to explain this observation is that surface plasmon resonance (SPR) from silver nanoparticles is generated inside the cells. Surface plasmon resonance is defined as the collective oscillation of valence electrons in a solid that is stimulated by incident laser light. It has been shown that both Ag and Au nano clusters strongly absorb and emit laser light by SPR [21-26]. In our studies the intensity of the signal observed from laser light excitation was a function of both the AgNP size and agglomeration. While 50 and 75 nm nanoparticles caused strong fluorescent signals, the 10 nm particles did not (Supporting Information Fig. 3). The smaller particles probably yielded smaller agglomerates and therefore a smaller plasmon surface to interact with the laser light. Larger particles most likely clumped into larger silver nano-clusters which induced a larger plasmon surface to interact with laser light. The interaction of the nanoparticles with molecules in the cytoplasm or the intra-cellular matrix can influence the fluorescent signal and produce a very strong fluorescence signal [21-26]. These AgNP located in a nano cluster give rise to a sea of freely moving electrons that display surface plasmon resonance. These nanoparticles can concentrate light into local electromagnetic hot spots that lead to increased absorption of incident light [23, 24]. Thus, the local environment surrounding the silver particles may affect the SPR, which in turn can affect the magnitude and/or wavelength of the resulting emission.

Flow Cytometry

Silver nanomaterials were observed inside cells by increased side scatter using flow cytometry, and by dark field microscopy. Incubation of AgNP with ARPE-19 cells increased SSC and decreased FSC in a dose dependent manner as the concentration of silver increased between 1 μg/ml to 100 μg/ml. The magnitude of the increase in SSC after treatment with AgNP was only about 1/3 of that observed using similar sized, highly reflective TiO2NP [2, 14, 17]. This suggests that flow cytometry (FCM) may be a technique to observe many types of nanoparticles, some that are highly reflective like TiO2 or some that have other properties (SPR) like silver or gold nanoparticles. The relationship between AgNP, dose, and increase in side scatter suggests that the amount of nanomaterials absorbed inside the cells is related to the mass concentration that is applied to the cells. Whether the AgNP are single or clustered in the cell will affect the degree of side scatter and possibly far red fluorescence. Using both the MFI intensity ratios and the stain index (SI), it was shown that AgNP could increase the signal around 100x background. Although the two methods indicated some differences, the general trends showed that the far red fluorescence was apparent and reproducible using both methods of data analysis. Microscopically, when higher concentrations of particles were added to the cells, larger clumps of AgNP were formed (as shown in Fig. 4), which reflect more light than the equivalent number of individual single particles. Since the SSC parameter can be easily measured using a standard flow cytometer, it may be possible to correlate the absorbed nanoparticles with observable cellular toxicity and viability in future experiments.

The entry of AgNP into cells was influenced by both particle size and coating. The AgNP coated with either PVP or citrate in three different particle sizes (10, 50, and 75 nm) had different penetration rates into ARPE-19 cells. Side scatter usually is approximately proportional to a particle's size. However, it can also be dependent upon the particle's composition, refractive index, and the wavelength of light used for excitation. For the same mass concentration of AgNP added to cells the larger particles (75>50>10 nm) scattered more light than the smaller particles. AgNP coated with citrate scattered more light than AgNP coated with PVP. Flow cytometers have been used in many diverse applications to measure both light scattering and fluorescence from particles or biological cells. The SSC parameter is more sensitive than FSC to detect small sub-micron particles whether they are inside cells or in suspension. Side scatter increases and forward scatter decreases upon interaction of cells with laser light [2, 17]. Gold nanoparticles attached to antibodies were shown to have greater than ten-fold increase in SSC compared to control [27, 28]. Silver attached to antibodies greatly increased the fluorescence and scatter signals. Nanoparticles attached to bacteria or to molecules in the circulating blood increase their resolution of detection [29-33]. This study suggests that flow cytometry may be applied to detect many types of sub-micron and nanoparticles if they are bound to antibodies or can be incorporated into the cell with various types of transport vehicles.

The observation that a strong far-red fluorescent signal was generated from AgNP nanoparticles and detected with standard flow cytometers presents a unique opportunity to detect particles composed of noble metals (such as Ag and Au) that are susceptible to the surface plasmon resonance phenomenon. These approaches may help to assess the relationship between nanoparticle concentration and cellular toxicity. Such experiments to measure cellular uptake of nanoparticles could potentially be performed quickly and easily using a flow cytometer.

Microscopy

Darkfield microscopy has been used previously to study nanoparticles [2, 14-16, 24, 34-37]. Cells treated with AgNP were evaluated using a microscope containing a xenon light source and special dark field objectives. Nanoparticles suspended in the media gradually settled onto ARPE-19 cells at the bottom of the cell culture plates and passed through the cell membranes. The nanoparticles could be observed inside the cells and arranged in clusters around the nuclei in the vicinity of the endoplasmic reticulum. (Fig. 4, Supporting Information Fig. 3) Aggregation was dependent in part on the concentration of particles added to the growing cells such that higher concentrations of AgNP resulted in larger and more numerous clumps of nanoparticles in the cytoplasm.

Incubation of 10 nm and 75 nm nanoparticles at the same concentration showed that the cells treated with the larger nanoparticles (75 nm) had a stronger side scatter signal and more numerous particles by darkfield microscopy than did cells incubated with the smaller particles (Supporting Information Fig. 3). Using the flow cytometer, greater side scatter and far red fluorescence signals were observed with the 75 nm particles than the 50 nm or 10 nm particles. This can likely be attributed to an increased number of larger particle aggregates in the cytoplasm than were observed microscopically.

It has been observed that light emitted from the same area may be green, red, and yellow yielding the term to define these multiple colors as blinking [26, 37-42]. This change in color may be due to the Ag charge being continually modified and modifying the local environment. Occasionally, it has been reported that by using confocal microscopy, the light emitted from areas where nanomaterial aggregations occurred the light appeared to oscillate between green, red, and yellow; a phenomenon others have referred to as “blinking” [37, 42]. Blinking has been observed in clumps of Ag particles that are unfixed in our laboratory and this observation may possibly relate to the fluorescent effect that is being reported in this communication using the flow cytometers.

Fixation of the cells with paraformaldehyde reduced the far red fluorescence signal in a flow cytometer by approximately 15-fold, possibly by decreasing the degree of oscillation of the Ag molecules that generate the plasmon or denaturing proteins or other biomolecules that might influence or participate in the plasmon resonance. This fixation also appears to reduce blinking from AgNP. Further research is necessary to elucidate the mechanism of fluorescence signals generated from Ag nanoparticle clusters located within the cells.

Modified flow cytometers have been used to study surface enhanced Raman scattering (SERS) [43-49]. Since Raman spectroscopy signals are typically greater than 1,000 nm; these flow cytometers have been modified with spectral detection capabilities and EMCCD cameras to detect the fluorescence in the spectral ranges above 1,000 nm. Using this equipment, Nolan and collaborators described a typical SERS experiment as utilizing a plasmonic nanoparticle, usually gold or silver, as a starting substance that developed surface plasmon resonance when illuminated with laser light of the correct wavelength. In contrast, we observed far red fluorescence below 1000 nm utilizing PMT's to detect the fluorescent light. It is conceivable that the fluorescence we observed may be at the leading edge of SERS spectra. Unlike the SERS, flow cytometers, far red fluorescence was observed on conventional flow cytometers and thus should be readily available in many scientific laboratories

A number of different type of experiments have been attempted to demonstrate that silver nano particles can yield far red fluorescence using methods other than flow cytometry. This effect has not been duplicated on a standard research fluorescence microscope, a confocal fluorescence microscope, in a spectrophotometer or on a slide with AgNP in solution. Fixation of cells containing AgNP reduced the effect by 90%. It is possible that this far red fluorescence observation is the result of nanoparticles migrating to the endoplasmic reticulum, and clumping which is followed by an interaction with some unknown molecules in this structure to induce far red fluorescence. It is very strange that we cannot visualize this far red fluorescence effect on research grade confocal or fluorescence microscopes on live or fixed cells. One may need a more powerful light source, like a 2 photon laser and use live cells to visualize the far red fluorescence of these nanoparticles. This effect of nanoparticles in cells will be investigated in a subsequent experiments dealing with intercellular transport of nanoparticles and cellular morphology.

Summary

Ag nanoparticles can be detected inside cells using a flow cytometer as increased side scatter and far red fluorescence signals. The uptake of nanoparticles in cells is dependent on the size and surface characteristics. The 75 nm AgNP showed the largest increase in side scatter, far red fluorescence and nanoclusters. The far red fluorescence signal could be over 100 x background levels at some concentrations of AgNP, but was decreased when live cells were fixed. The results suggest that silver particles may interact with biomolecules in living cells to create fluorescence signals that are detected in the far red range (>670 nm). The phenomenon may be related to Ag plasmon surface resonance.

Acknowledgments

Thanks are extended to John Rogers, Chris Lau, and Carl Blackman for their helpful comments and encouragement for many aspects of the paper and to Jim Birk of BD for restoration of a flow cytometer that was inactive for three years after a hurricane damaged an EPA Laboratory. Thanks are extended to Nancy Fisher and John Lay of UNC for allowing us to use their flow cytometry equipment for these studies. We wish to thank Keith Tarpley for helping us prepare the figures and tables used in this manuscript. None of the authors have any conflict of interest with any of the companies.

Ancillary