High-Throughput, Single-Cell Analysis of Macrophage Interactions with Fluorescently Labeled Bacillus anthracis Spores

Abstract

The engulfment of Bacillus anthracis spores by macrophages is an important step in the pathogenesis of inhalational anthrax. However, from a quantitative standpoint, the magnitude to which macrophages interact with and engulf spores remains poorly understood, in part due to inherent limitations associated with commonly used assays. To analyze phagocytosis of spores by RAW264.7 macrophage-like cells in a high-throughput, nonsubjective manner, we labeled B. anthracis Sterne 7702 spores prior to infection with an Alexa Fluor 488 amine-reactive dye in a manner that did not alter their germination, growth kinetics, and heat resistance. Using flow cytometry, large numbers of cells exposed to labeled spores were screened to concurrently discriminate infected from uninfected cells and surface-associated from internalized spores. These experiments revealed that spore uptake was not uniform, but instead, highly heterogeneous and characterized by subpopulations of infected and uninfected cells, as well as considerable variation in the number of spores associated with individual cells. Flow cytometry analysis of infections demonstrated that spore uptake was independent of the presence or absence of fetal bovine serum, a germinant that, while routinely used in vitro, complicates the interpretation of the outcome of infections. Two commonly used macrophage cell lines, RAW264.7 and J774A.1 cells, were compared, revealing significant disparity between these two models in the rates of phagocytosis of labeled spores. These studies provide the experimental framework for investigating mechanisms of spore phagocytosis, as well as quantitatively evaluating strategies for interfering with macrophage binding and uptake of spores.

Host macrophages play an important role in inhalational anthrax, which is a complex, multistep disease that begins with the deposition of inhaled Bacillus anthracis spores within the alveolar space of the lungs (9, 12, 43). Histopathological studies using several animal infection models revealed that subsequent to spore inhalation, B. anthracis breaches the alveolar epithelial barrier without any clinical evidence of lung injury (1, 3, 18, 27, 36, 50). In several animal models, spores were observed to be rapidly taken up by alveolar macrophages and transported to regional lymph nodes (1, 18, 27). Thus, macrophage-mediated transport of B. anthracis is predicted to be an important mechanism by which vegetative bacilli access the bloodstream, resulting ultimately in dissemination to the liver, spleen, and additional organs (9, 12).

In vitro models have been widely used to obtain important insights into possible outcomes of B. anthracis interactions with macrophages. However, drawing conclusions from these in vitro studies has been challenging, in part because a variety of cells have been employed, with some models using primary macrophages from an assortment of animals, tissues, and genetic backgrounds while other models have employed established macrophage-like cell lines (4, 8, 11, 20-22, 26, 32, 45, 46, 51). The relative extents to which these different macrophages and cells respond to and engulf B. anthracis spores, which is critical for the outcome of infection, have not been systematically investigated.

The accurate enumeration of spore engulfment by macrophages has been restricted, in part, by the inherent limitations of existing assays. Some in vitro models employ the gentamicin protection assay to study B. anthracis spores internalized within a population of cells (2, 8, 13). Macrophages are typically exposed to spores for a period of time, and after nonassociated spores are removed, extracellular bacilli are killed by the addition of gentamicin prior to enumeration of intracellular CFU. An intrinsic limitation of the gentamicin protection assay, as with any assay based on quantifying CFU, is that information can be obtained only about viable organisms, e.g., there can be no accounting for B. anthracis spores that were taken up by macrophages but rapidly killed prior to serial plating. Germination-deficient spores (6, 42, 44, 48, 49) that are phagocytosed into a host cell vacuole but are unable to germinate also cannot be detected by assaying for CFU. Gentamicin protection assays are also limited by their ability to reveal only the average number of viable B. anthracis spores per macrophage within a population, but they cannot provide information at the single-cell level about the percentage of macrophages that become infected or how many spores are taken up into individual macrophages. In addition to these innate limitations, several studies have indicated that gentamicin can enter macrophages, and possibly intracellular compartments (10, 16), and interfere with the germination of phagocytosed B. anthracis spores (2, 8). An alternative to the gentamicin protection assay has been the use of microscopy to observe the engulfment of spores by individual cells (14, 21, 37). However, microscopy-based approaches are limiting in terms of the number of cells that can be analyzed within a reasonable time frame, as well as the subjectivity of the user. Consequently, from a quantitative standpoint, the level to which spores interact and are engulfed by macrophages remains poorly understood.

To more thoroughly evaluate macrophage uptake of spores, we labeled B. anthracis Sterne 7702 spores with an Alexa Fluor 488 amine-reactive dye and demonstrated that labeling did not alter the germination, outgrowth, or heat resistance of spores. Using flow cytometry, we rapidly screened a large number of cells incubated with prelabeled spores to concurrently discriminate infected from uninfected cells, as well as internalized from surface-bound spores. We applied this approach to acquire several new insights into commonly used in vitro infection models. Finally, these studies indicated that spore uptake was not uniform but, instead, highly heterogeneous and characterized by subpopulations of infected and uninfected cells, as well as considerable variation in the numbers of spores associated with individual cells. These studies provide the experimental framework for investigating mechanisms of spore phagocytosis, as well as quantitatively evaluating strategies for interfering with macrophage binding and uptake of spores.

MATERIALS AND METHODS

B. anthracis spore preparation.

An aliquot (100 μl) of an overnight culture of B. anthracis Sterne 7702 grown in brain heart infusion broth (3.7% Bacto Brain Heart Infusion [BD Diagnostics, Franklin Lakes, NJ], Millipore deionized water, 0.5% glycerol) at 37°C was plated onto Difco sporulation medium agar plates (8 g/liter Difco nutrient broth [BD Diagnostics], 1 g/liter KCl, 250 mg/liter MgSO₄·7H₂O, 15 g/liter agar) and incubated at 30°C under ambient CO₂. After 3 days, each plate was washed with deionized water (10 ml), and the washes were pooled and filtered sequentially through 3.1- and 1.2-μm glass filters from National Scientific Company (through VWR, Rochester, NY) to remove vegetative cells and clumped spores (30). The spore suspensions were incubated at 65°C for 30 min to inactivate any remaining heat-sensitive organisms. The spores were washed three times with deionized water (centrifugation at 3,270 × g for 30 min), and stored in deionized water (10 ml) at 4°C. Spores were prepared on a weekly basis and used within 7 days. Prior to each experiment, we confirmed that the spores had not undergone germination during storage by assessing the resistance of the spores to heat, as described below.

Quantification of spores.

Spores were directly counted using a Petroff-Hauser Chamber 10 hemacytometer (Electron Microscopy Sciences, Washington, PA) under a light microscope (Nikon Alphaphot YS; Nikon, Mellville, NY). In parallel, spores were serially diluted on Luria-Bertani (LB) (10 g/liter Bacto Tryptone, 5 g/liter NaCl, 5 g/liter Bacto Yeast Extract, 15 g/liter Bacto Agar; BD Diagnostics) plates, which were incubated at 37°C. After 12 to 18 h, B. anthracis colonies were counted on plates with 30 to 300 colonies, from which CFU/ml were calculated. Several studies have reported that analyzing spore preparations by direct microscopic counting sometimes yields spore counts higher than those obtained by quantifying CFU (42, 44, 49), indicating that not all spores germinate in some preparations. The capacity of spores to germinate and outgrow was calculated by dividing the number of CFU by the number of hemacytometer-counted spores. Spores were used only if CFU were determined to be >90% of the number of hemacytometer-counted spores.

Alexa Fluor 488 labeling of spores.

B. anthracis spores were labeled using an amine-reactive Alexa Fluor 488 carboxylic acid succinimidyl ester (Molecular Probes-Invitrogen, Eugene, OR). Just prior to the labeling reaction, 1 ml of freshly prepared 1 M sodium bicarbonate was added to freshly prepared spores (1 × 10⁹ to 3 × 10⁹) in 9 ml of deionized water. To this mixture, 50 μg of the Alexa Fluor 488 amine-reactive dye, previously dissolved in dimethyl sulfoxide and aliquoted, was added to the spores, and the mixture was incubated at room temperature. The extent of spore labeling, as revealed by flow cytometry, was dependent on the duration of the labeling reaction (data not shown). Typically, spores were incubated with the dye for 3 h. Unreacted dye was removed by washing the spores three times with deionized water.

TEM.

Samples for transmission electron microscopy (TEM) were fixed in Karnovsky's fixative and submitted to the Center for Microscopic Imaging core facility at the University of Illinois for preparation and imaging using a Hitachi H600 transmission electron microscope.

Germination and outgrowth kinetics.

Identical numbers of labeled and unlabeled spores were cultured separately in LB broth at 37°C on a platform shaker (250 rpm). At 5, 15, and 60 min, aliquots were removed, diluted 1:20 in ice-cold Hank's balanced salt solution (HBSS) (Mediatech, Inc., Herndon, VA), washed once with ice-cold HBSS, and finally resuspended in a final volume of 1 ml of ice-cold HBSS. Each sample was divided equally and incubated for 30 min either at 65°C or on ice. B. anthracis organisms in each aliquot were quantified by plating serial dilutions and enumerating CFU, as described above. The percentage of spores that had undergone germination and outgrowth was calculated by dividing the number of CFU within samples incubated at 65°C by the number of CFU within samples incubated on ice.

Growth in liquid culture.

Identical numbers of labeled and unlabeled spores were cultured at 37°C in LB broth on a platform shaker (250 rpm). Optical density measurements were collected at 600 nm.

Heat resistance.

Identical numbers of labeled and unlabeled spores in HBSS were incubated at 65°C or on ice. At t = 30, 60, or 120 min, aliquots were removed, and B. anthracis organisms were quantified by plating serial dilutions and enumerating CFU, as described above. The fraction of heat-resistant spores was calculated by dividing the number of CFU recovered from samples heated at 65°C by the number of CFU recovered from samples incubated on ice.

Cell culture.

RAW264.7 cells (CRL-2278; ATCC, Manassas, VA) were maintained within a humidified environment at 37°C and under 5% CO₂ in RPMI-1640 medium (ATCC) containing penicillin (100 U; Gibco BRL, Grand Island, NY), streptomycin (0.1 mg/ml; Gibco BRL), l-glutamine (2 mM; Sigma), and fetal bovine serum (FBS) (10%; JRH Biosciences, Lenexa, KS). J774A.1 cells (TIB-67; ATCC) were maintained within a humidified environment at 37°C and under 5% CO₂ in Dulbecco's modified Eagle's medium (DMEM) (JRH Biosciences) containing penicillin-streptomycin, l-glutamine (4 mM; Sigma, St. Louis, MO), and FBS (10%). All tissue culture plasticware was purchased from Corning Incorporated (Corning, NY).

Spore interactions and uptake by mammalian cells.

Cells were seeded into 48-well plates or 8-well chambered slides (Nalge Nunc International, Rochester, NY) in order to achieve 80 to 95% confluence after 2 days of incubation and were incubated with DMEM containing penicillin-streptomycin (100 U penicillin, 0.1 mg streptomycin/ml), l-glutamine (4 mM), and FBS (10%) in a humidified environment at 37°C and under 5% CO₂. To accurately calculate the number of labeled spores needed to achieve the indicated multiplicity of infection (MOI), cells from several wells were counted using a hemacytometer immediately before each experiment. The cells were used only if greater than 90% of the cells excluded trypan blue; generally, greater than 95% of the cells within the monolayer excluded trypan blue. Prior to the addition of labeled spores, cells were washed at least three times with HBSS and then incubated in DMEM containing l-glutamine (4 mM). Where indicated, the medium was also supplemented with 10% FBS or RPMI-1640 medium or HBSS was substituted for DMEM. For binding experiments, cells were preincubated with cytochalasin D (10 μM) for 1 h prior to the addition of labeled spores. To synchronize the exposure of cells to spores, labeled spores were gently centrifuged (300 × g for 5 min) onto the surfaces of cells immediately after addition. The plates or slides were incubated within a humidified environment at 37°C and under 5% CO₂ for the indicated times prior to analysis.

Mammalian cell viability.

Propidium iodide (PI) (1 μg/ml) uptake by RAW264.7 or J774A.1 cells was measured using flow cytometry, as previously described (31). Because most of the uptake and binding experiments were conducted in the absence of serum, we confirmed that the percentages of PI-positive RAW264.7 or J774A.1 cells were essentially identical in the presence or absence of FBS over 6 h (data not shown), indicating that cell viability did not decrease in the absence of FBS.

Fluorescence quenching of Alexa Fluor 488-labeled spores.

Stock solutions of trypan blue were made in PBS, pH 7.2, and filtered using a 0.22-μm filter prior to use. Where indicated, trypan blue was added (at the indicated final concentration; typically, 0.5% in PBS, pH 7.2) to Alexa Fluor 488-labeled spores or to mammalian cells exposed to labeled spores. The samples were incubated for 5 min on ice and analyzed immediately by flow cytometry or microscopy.

Flow cytometry.

Mammalian cells were detached from tissue culture wells using cell dissociation buffer (Sigma), collected by centrifugation, and resuspended in 400 μl of ice-cold HBSS in 5-ml polypropylene tubes (Becton Dickinson, Franklin, NJ) for flow cytometry analysis. Analytical flow cytometry was carried out using a Beckman Coulter Epics XL-MCL flow cytometer equipped with a 70-μm nozzle, a 488-nm line of an air-cooled argon ion laser, and a 400-mV output. The band pass filter used for detection of spores was 525/10 nm. Cell analysis was standardized for side/forward scatter and fluorescence by using a suspension of fluorescent beads (Beckman Coulter Inc., Fullerton, CA).

At least 10,000 events were detected for each experiment (>2,000 events per min). Events were recorded on a log fluorescence scale, and the density plots were generated using FCS Express 3.00.0311 V Lite Standalone. Data were expressed as forward scatter and side scatter, which are indicators of relative size and cellular complexity or granularity, respectively, and the fluorescence intensity. For some experiments, the data were plotted as the geometric mean of the fluorescence intensity (MFI). The percentage of cells with bound and/or internalized B. anthracis spores was calculated by dividing the number of viable cells with greater than background autofluorescence by the total number of viable cells.

Viable RAW264.7 cells (typically 90 to 98% of the total events) were readily identified by their high forward scatter (Fig. 1A) and lack of PI staining (data not shown). A second distinct population (2 to 8%) of dead RAW264.7 cells was routinely detected with lower forward scatter (which indicates a smaller size, as indicated in Fig. 1A) and positive PI staining (indicating nonviable cells; data not shown). Over the course of individual experiments, we observed no detectable increase in cell death in the presence of labeled spores, as indicated by PI uptake (data not shown). Finally, sample debris (as indicated by lower forward and side scatter and a lack of PI staining) represented a small fraction (1 to 2%) of the detected events. Based on these data, the data from subsequent experiments were gated to include only viable cells while excluding nonviable cells, cellular debris, and spores not associated with cells.

FIG. 1.

Flow cytometry discriminates RAW264.7 cells and B. anthracis spores. The data are depicted as density plots of the forward and side scatter of RAW264.7 cells (A) and labeled spores (B) prepared from B. anthracis Sterne 7702. The number in each quadrant represents the percentage of either RAW264.7 cells (A) or spores (B).

Quantification of cell-associated viable B. anthracis.

Cells exposed to B. anthracis were washed three times with HBSS and then lysed by resuspending and vortexing the cell pellet in sterile tissue culture grade water (Sigma) for 5 min at room temperature. To determine CFU, serial dilutions of the suspensions were plated on LB agar plates as described above. Microscopy revealed no detectable spore clumping following lysis of the cells.

Fluorescence-activated cell sorting.

A MoFlo Multi Laser Sort high-speed sorting flow cytometer from Dako Cytomation (Fort Collins, CO) was used to sort cells in the R. J. Carver Biotechnology Center at the University of Illinois. Using a 100-μm nozzle at 30 lb/in², subpopulations of infected RAW264.7 cells, arbitrarily defined by their relative fluorescence, were sorted until 10,000 cells from each subpopulation had been distributed into separate wells of a 96-well tissue culture plate. The cells were lysed, as described above, and the entire lysate was dilution plated to enumerate CFU.

DIC/fluorescence microscopy.

Cells were washed with HBSS, fixed with 2% paraformaldehyde in HBSS, and then visualized using a Delta Vision RT microscope (Applied Precision, Issaquah, WA) outfitted with a 490/20 excitation-528/38 emission filter set and an Olympus Plan Apo 60× oil objective with a numerical aperture of 1.42 and a working distance of 0.17 mm. Differential interference contrast (DIC) images were collected for all fields. Images were processed using SoftWoRX Explorer Suite. As previously described (37), cells were examined to differentiate whether cell-associated spores were intracellular or extracellular and to determine the number of spores associated with a cell. For each cell, we collected a minimum of 25 to 30 optical z sections at 0.5-μm intervals. By carefully examining the series of optical planes, we counted the number of spores associated within a single cell. Typically, five or six randomly chosen cells within 10 to 12 separate fields were analyzed for each experimental condition.

Statistics.

All data are representative of those from three or more independent experiments. Error bars represent standard deviations. P values were calculated with Student's t test using paired, one-tailed distribution. P values of <0.05 indicate statistical significance. All statistics, including means, standard deviations, and Student's t tests, were calculated using Microsoft Excel (version 11.0).

RESULTS

Discrimination of RAW264.7 cells from B. anthracis spores.

To evaluate the macrophage uptake of B. anthracis spores, we employed an in vitro model comprising RAW264.7 macrophage-like cells incubated with spores prepared from B. anthracis Sterne 7702. When analyzed by flow cytometry, RAW264.7 cells and spores exhibited clear differences in forward and side scatter (Fig. 1A and B), indicating that within an infection, cells and nonassociated spores could be readily differentiated.

Fluorescent labeling and quality control assessment of modified B. anthracis spores.

To monitor spore interactions and uptake by RAW264.7 cells, we developed a protocol for labeling spores with an Alexa Fluor 488 amine-reactive dye to form stable dye-protein conjugates. By prelabeling spores prior to infection, the goal was to eliminate the requirement for further manipulation (e.g., fixing, permeabilization, and staining) of the cells or spores prior to analysis. After the spores were labeled, flow cytometry revealed that the fluorescence emission from the labeled spores was considerably greater than that from unlabeled spores (Fig. 2A). Moreover, the labeled spores were characteristically detected as a symmetric fluorescent population, which indicates that the population was homogeneously labeled (Fig. 2A).

FIG. 2.

Prelabeling does not alter spore properties. Alexa Fluor 488-labeled spores and unlabeled spores from the same preparation of spores were compared by measuring fluorescence intensity by flow cytometry (A), side and forward scatter by flow cytometry (B), the capacity to germinate and outgrow (C), the rate of germination and outgrowth (D), growth in liquid culture (E), and resistance to heat (F). Each preparation of labeled spores was evaluated for fluorescence intensity, side and forward scatter, the capacity to germinate and outgrow, and resistance to heat. Preparations of labeled spores were evaluated for the rate of germination and outgrowth and growth in liquid culture primarily during the development and optimization of the spore-labeling protocols. The error bars (C to F) indicate standard deviations. In panels C to F, there were no significant differences (i.e., P > 0.05) between data collected with labeled spores and unlabeled spores.

To confirm that the labeling procedure had not altered spore properties, each preparation of labeled spores was subjected to quality control assessment. These experiments revealed that labeled spores were essentially identical to unlabeled spores in terms of forward and side scatter (Fig. 2B), as well as overall gross morphology, as indicated by TEM (data not shown). The labeling procedure did not affect the germination or growth properties of B. anthracis, as labeled and unlabeled spores were indistinguishable in their capacities to germinate and outgrow in the presence of a germinant (Fig. 2C), rates of spore germination and outgrowth after exposure to a germinant (Fig. 2D), and growth in liquid culture (Fig. 2E). Finally, the labeling procedure did not induce germination, as labeled and unlabeled spores demonstrated nearly identical resistance to heat (Fig. 2F). These results validated the notion that labeling spores did not discernibly alter their germination, outgrowth, or heat resistance properties.

Interactions of B. anthracis spores with RAW264.7 cells.

We next evaluated the use of flow cytometry to study the uptake of labeled spores by mammalian cells. Importantly, flow cytometry readily differentiated labeled spores and RAW264.7 cells as two discrete and nonoverlapping populations based on differences in side scatter and fluorescence in the 525/10-nm channel (data not shown). Thus, in the presence of labeled spores, flow cytometry could be used to specifically analyze mammalian cells without interference from noninteracting spores.

When RAW264.7 cells were exposed to labeled spores (MOI = 10) and immediately analyzed by flow cytometry, a small but detectable population of RAW264.7 cells (<7%) formed interactions with spores in the short time period (<5 min) between mixing and analysis (data not shown). In contrast, two distinct populations of RAW264.7 cells were evident after 1 h at an MOI of 1, 10, or 100 (Fig. 3A to D). At an MOI of 1, cells were detected with (21%) or without (79%) associated labeled spores (Fig. 3B). The percentage of RAW264.7 cells with associated spores increased as a function of the MOI (Fig. 3C and D). The percentage of RAW264.7 cells that were associated with spores was reproducible over multiple independent experiments. The percentages ± standard deviations of cells infected with spores (i.e., the average numbers of cells, obtained from multiple [n] independent experiments, each performed in duplicate or triplicate on different days with different spore preparations, yielding fluorescence above the autofluorescence of the cells alone) were as follows: MOI of 1, 17.1 ± 5.3 (n = 3); MOI of 10, 54.0 ± 11 (n = 4); and MOI of 100, 92.2 ± 6.2 (n = 4). We nearly always detected viable RAW264.7 cells without associated spores, even at an MOI of 10 or 100.

FIG. 3.

Flow cytometry analysis of spore-cell interactions. RAW264.7 cells were incubated with labeled spores at an MOI of 1 (B and E), 10 (C and E), or 100 (D and E) for 1 h at 37°C and analyzed by either flow cytometry (A to D) or DIC/fluorescence microscopy (E). The numbers shown in the upper right and upper left corners in panels A to D represent the percentage of RAW264.7 cells with or without associated Alexa Fluor 488-labeled spores, respectively. (E) Each symbol represents the number of spores associated with a single RAW264.7 cell at an MOI of 1 (squares), 10 (triangles), or 100 (circles). n is the number of cells analyzed at each MOI, avg is the mean number of spores/cell at each MOI, and the percentage is the number of cells that had at least one associated spore.

Our flow cytometry results were validated by DIC/fluorescence microscopy, which revealed that 30%, 65%, and 100% of the total RAW264.7 cells had associated spores at an MOI of 1, 10, or 100, respectively (Fig. 3E). These data correspond well to data collected by flow cytometry, which yielded 21%, 67%, and 99% at MOI 1, 10, or 100, respectively (Fig. 3B to D). DIC/fluorescence microscopy of individual cells revealed an average of 0.52, 0.90, or 9.8 spores per cell at an MOI of 1, 10, or 100, respectively (Fig. 3E), which is consistent with an earlier study reporting that not all spores are taken up by RAW264.7 cells (21).

The broadness of the flow cytometry histograms, especially at an MOI of 100 (Fig. 3D), suggested that the number of spores associated with individual RAW264.7 cells was highly variable. DIC/fluorescence microscopy confirmed that there was considerable variation in the number of spores associated with individual RAW264.7 cells (Fig. 3E), a phenomenon that has been noted previously in studies with macrophages (34), with ranges of 0 to 3, 0 to 5, or 2 to 22 spores/cell at an MOI of 1, 10, or 100, respectively. The variability in the number of associated spores was not due to either the clumping or the uneven distribution of spores (data not shown).

To validate the relationship between cell-associated fluorescence and the relative numbers of spores associated with individual cells, RAW264.7 cells were incubated with labeled spores (MOI, 1, 10, or 100) for 1 h and then immediately fractionated using fluorescence-activated cell sorting by arbitrarily gating subpopulations of RAW264.7 cells according to their relative fluorescences (as illustrated in Fig. 4A). At each MOI, 10,000 cells from each subpopulation were collected and lysed, and B. anthracis CFU were enumerated by plating serial dilutions of the entire lysate. These experiments revealed that the subpopulations of infected cells with the lowest fluorescence (subpopulations 1 to 3) yielded approximately 0.25 to 1.00 CFU/cell, regardless of whether subpopulations were collected at an MOI of 1, 10, or 100 (Fig. 4B). The subpopulation with intermediate fluorescence (subpopulation 4) yielded approximately 1.5 to 2 CFU/cell, and this number rose to approximately 5 to 6 CFU/cell for subpopulations with higher fluorescence (subpopulations 5 to 7). From experiments in which we collected cells gated to have no associated spores, we recovered less than 0.05 CFU per cell. An inherent limitation of this approach is that the number of spores/cell calculated from CFU enumeration does not account for spores that, while associated with the collected cells, had been killed or were unable to germinate. Nonetheless, three independent experiments reproducibly demonstrated that cell-associated fluorescence increases in relation to the number of B. anthracis spores recovered from cells. Moreover, these data demonstrated that CFU recovered from individual subpopulations defined by fluorescence were generally independent of the MOI.

FIG. 4.

Analysis of the relationship between RAW264.7 cell fluorescence and recoverable B. anthracis CFU. (A) RAW264.7 cells were incubated without (black) or with labeled spores at an MOI of 1 (olive), 10 (blue), or 100 (dark green). After 1 h, the fluorescence of viable cells was analyzed by flow cytometry. The macrophages that yielded fluorescence above background autofluorescence were arbitrarily divided into gated subpopulations (labeled 1 to 10) defined by their relative fluorescence intensities. (B) RAW264.7 cells exposed to spores at each MOI were sorted until 10,000 cells from each gated subpopulation were collected within separate wells of a 96-well tissue culture plate. At each MOI, RAW264.7 cell subpopulations were isolated only when more than 1% of the total cells were present within that gated subpopulation. Thus, cells gated within subpopulations 1 to 7 were sorted and collected from experiments conducted at an MOI of 100, subpopulations 1 to 4 at an MOI of 10, and subpopulations 1 to 3 at an MOI of 1. There were not enough cells within subpopulations 8 to 10 to be collected and analyzed. The RAW264.7 cells from each well were collected by centrifugation and lysed, and associated B. anthracis spores were determined by plating serial dilutions of the entire lysate to enumerate CFU. The number of spores per cell was calculated by dividing the number of CFU by 10,000 cells. The data shown are representative of a single experiment in which the entire lysate from each subpopulation was plated, which precluded the calculation of error for the experiment.

Discrimination of intracellular and cell surface-associated B. anthracis spores.

To discriminate intracellular spores from those that remain surface associated during infection, cells exposed to labeled spores were analyzed by flow cytometry in the presence or absence of trypan blue, a membrane-impermeable, Alexa Fluor 488 fluorescence-quenching agent (17, 38). Trypan blue has been used effectively to monitor the uptake into mammalian cells of several fluorescently labeled infectious agents (5, 25, 39). In the presence of 0.05 or 0.5% trypan blue, the fluorescence of labeled spores was quenched to nearly the levels of autofluorescence characteristic of unlabeled spores (Fig. 5A), indicating that the Alexa Fluor 488 label was accessible for quenching. Importantly, trypan blue did not detectably alter the capacity of labeled spores to respond to the presence of germinant, as the extents of spore germination and outgrowth were essentially identical in the presence or absence of trypan blue (Fig. 5B). Moreover, trypan blue did not induce germination, as labeled spores demonstrated similar heat resistance in the presence or absence of the reagent (Fig. 5C).

FIG. 5.

Trypan blue quenches the fluorescence of Alexa Fluor 488-labeled B. anthracis spores. Labeled spores were incubated on ice with trypan blue in HBSS and then analyzed for fluorescence intensity by use of flow cytometry (A), for the capacity to germinate and outgrow (B), and for resistance to heat (C). In panel A, the data are presented as the MFI. For panels B and C, the spores were incubated in 0.5% trypan blue only. The statistical significance in panels B and C was calculated for the differences between data collected using Alexa Fluor 488-labeled spores versus unlabeled spores. The error bars indicate standard deviations.

To study internalization, RAW264.7 cells were incubated with labeled spores at 37°C for 1 h to allow the uptake of B. anthracis spores and then analyzed by flow cytometry in the absence and presence of trypan blue (0.5%). The presence of trypan blue quenched RAW264.7 cell-associated fluorescence by approximately 11% (Fig. 6A), indicating that the majority of cell-associated spores had been internalized. Three independent experiments demonstrated that under these conditions, RAW264.7 cell-associated fluorescence was quenched by an average of 13% (± 3%), indicating that uptake was highly reproducible between experiments. DIC/fluorescence microscopy confirmed that while most cell-associated spores were not susceptible to trypan blue fluorescence quenching, a small number of spores that were susceptible to fluorescence quenching were routinely detected (Fig. 6B, top). Analysis of optical planes throughout individual cells validated that in the presence of trypan blue, spores that remained fluorescent were intracellular, whereas those that were no longer fluorescent were on the surfaces of cells (or not associated with cells).

FIG. 6.

Discrimination between extracellular-membrane-associated spores and phagocytosed spores. RAW264.7 cells were incubated for 1 h with labeled B. anthracis Sterne 7702 spores (MOI = 10) in the presence or absence of cytochalasin D (10 μM) within a humidified environment at 37°C and 5% CO₂ and were then analyzed by either flow cytometry (A) or DIC/fluorescence microscopy (B) in the presence or absence of trypan blue. In panel A, the data are presented as the MFI. (B) Cells were first analyzed by DIC/fluorescence microscopy in the absence of trypan blue (left). Trypan blue was then directly added to the slides, and the same cells were immediately reevaluated (right). In these experiments, some of the spores were unavoidably displaced when trypan blue was added to the slide. For panel A, the error bars indicate standard deviations. *, P < 0.05.

To measure binding, we exposed RAW264.7 cells to labeled spores in the presence of cytochalasin D, which inhibits actin-dependent uptake pathways, such as phagocytosis. Flow cytometry revealed that, compared to uptake experiments, the number of associated spores was decreased by greater than 80% in the presence of cytochalasin D (Fig. 6A), indicating that many fewer spores had become associated with the cells. In the presence of trypan blue, macrophage-associated fluorescence was quenched to background levels (Fig. 6A), indicating that almost all the spores were extracellular. DIC/fluorescence microscopy confirmed that the fluorescence of all visible spores was quenched by trypan blue (Fig. 6B, bottom), and analysis of optical planes within individual cells validated that all the spores either were located on the surface of the cell or were not directly associated with cells incubated with cytochalasin D.

Quantitative evaluation of the effects of FBS on the interaction of RAW264.7 cells with spores.

Having demonstrated the use of flow cytometry to detect labeled spores bound to or taken up into mammalian cells, we next applied this method to directly evaluate two variables that may influence the uptake of spores within in vitro infection models. First, although FBS is sometimes used in cell culture medium (2, 8, 13), this reagent can rapidly promote germination, thus potentially complicating experiments designed to evaluate macrophage interactions with dormant spores (20, 21). Although germination may be delayed in the presence of alternative sera from other animal species (24, 47, 51), it would be potentially advantageous to expose mammalian cells to spores within highly defined, serum-free medium.

To evaluate the effects of FBS on the extracellular spore association or uptake of B. anthracis, monolayers of RAW 264.7 cells were incubated with labeled spores (MOI = 10) in DMEM in the presence or absence of 10% FBS and in the absence or presence of cytochalasin D. These experiments revealed that the extracellular association (Fig. 7A) and uptake (Fig. 7B) of B. anthracis by RAW264.7 cells were not significantly different in either the presence or absence of FBS. During the development of these experiments, we determined that the absence of FBS for up to 4 h had no detectable effect on RAW264.7 viability or the capacity of these cells to take up and kill labeled spores (data not shown). Additional experiments revealed that the levels of spore interactions or uptake into RAW264.7 cells were essentially identical in reactions carried out in HBSS, DMEM, or RPMI medium (data not shown), indicating that differences in the compositions of these media did not detectably alter spore interactions with or uptake into these cells. These data provide direct experimental evidence that some in vitro experiments may be conducted in highly defined medium lacking FBS to overcome potential complications arising from the rapid germination of spores.

FIG. 7.

B. anthracis cell interactions and uptake are independent of FBS. RAW264.7 cells were incubated with labeled spores (MOI = 10) in the absence (A) or presence (B) of cytochalasin D (10 μM) and in the absence (diamonds) or presence (squares) of 10% FBS. The cells were analyzed by flow cytometry in the absence (B) or presence (A) of trypan blue (0.5%). The data are presented as the MFI at 0 (in the absence of spores), 15, 30, or 60 min. Statistical significance was calculated for differences between data collected in the absence or presence of serum at each time point. Statistical significance was not found at any of the time points tested.

Cell line dependence of spore interactions and uptake.

We next applied flow cytometry analysis of infections using prelabeled spores to evaluate the binding and uptake of B. anthracis spores by two different but commonly used macrophage cell lines. RAW264.7 and J774A.1 cells are both murine-derived, macrophage-like cell lines, and each has been used as an in vitro cell model to investigate the manner in which macrophages interact with B. anthracis spores (2, 8, 13-15, 21, 51). However, the extracellular spore association and uptake of spores by these two cell lines have not been quantitatively compared. Monolayers of RAW264.7 or J774A.1 cells were incubated with labeled spores (MOI = 10) in the absence or presence of cytochalasin D. At 30, 45, or 60 min, the cells were analyzed by flow cytometry in the absence or presence of trypan blue.

These experiments revealed that RAW264.7 cells internalized B. anthracis to a greater extent than J774A.1 cells (Fig. 8A). In contrast, the two cell lines demonstrated equal numbers of associated spores in the presence of cytochalasin D (Fig. 8B). The same results were obtained regardless of whether the experiments were carried out in HBSS, DMEM, or RPMI medium or in the absence or presence of serum (data not shown). These results, which were confirmed by microscopic observation (data not shown), suggest that J774A.1 and RAW264.7 cells are fundamentally different in the manner in which they internalize B. anthracis. Although the reasons underlying the differences between internalization of spores by J774A.1 and RAW264.7 cells have not yet been identified, these results underscore the importance of quantitatively considering cell line differences when evaluating in vitro cell models.

FIG. 8.

Flow cytometry detects differences in spore uptake between RAW264.7 and J774A.1 cells. RAW264.7 or J774A.1 cells were incubated with labeled spores (MOI = 10) in the absence (A) or presence (B) of cytochalasin D (10 μM). The cells were analyzed by flow cytometry in the absence (B) or presence (A) of trypan blue (0.5%). The data are presented as the MFI at 0, 15, 30, or 60 min. Statistical significance was calculated for differences between data collected with RAW264.7 and J774A.1 cells at each time point (C and D). *, statistical significance.

DISCUSSION

Here, we demonstrated the efficacy and power of using flow cytometry to evaluate the binding and uptake of prelabeled B. anthracis spores by mammalian cells. Critical to this work was our demonstration that the labeling reaction did not alter the germination, outgrowth, or heat resistance properties of the spores. Using flow cytometry, large numbers of cells exposed to prelabeled spores were screened to concurrently discriminate infected from uninfected cells and surface-associated from internalized spores. The applicability of this method was demonstrated, as discussed below, to provide new insights into several factors that are likely to be important when studying B. anthracis-macrophage interactions in vitro.

Flow cytometry, especially when combined with fluorescence quenching, provides an approach highly complementary to currently used assays, such as gentamicin protection assays or those involving visualization of cells using microscopy, for in vitro studies of the binding and engulfment of spores by mammalian cells. Because spores are prelabeled prior to exposure to mammalian cells, samples can be analyzed almost immediately without requiring additional manipulation. Flow cytometry assays are rapid, and we routinely analyze 10,000 cells within minutes, while gentamicin protection assays require overnight growth to enumerate CFU. Moreover, we discriminated intracellular from surface-bound spores by a very simple procedure using the fluorescence-quenching agent trypan blue. In contrast, microscopic analysis of optical z planes within individual cells can require many hours to investigate a relatively small number of cells (<100). An alternative, previously described microscopy-based approach (21) is effective but also time-consuming, requiring the fixing of cells prior to visualization and multiple labeling steps both before and after permeabilization of the cells. Analyzing cells by microscopic observation can be prone to the unavoidable nuances of individual investigators, while flow cytometry is entirely automated. Flow cytometry-based analysis of infections also overcomes the intrinsic limitations of using gentamicin and enumerating CFU, which include the inability to provide information about spores that have been engulfed but subsequently killed or unable to germinate (44). Recent reports suggest that a considerable percentage of B. anthracis spores may be killed soon after uptake into macrophage vacuoles (20, 21). Moreover, reports of gentamicin potentially interfering with the germination of phagocytosed B. anthracis spores (2, 8), possibly by entering cells and intracellular vacuoles like those occupied by B. anthracis (10, 16), suggest that the use of gentamicin to study in vitro infections should be given additional experimental consideration.

An important application of this flow cytometry-based method will be the systematic comparison of existing in vitro cell models. Understanding B. anthracis interactions with macrophages has been complicated by the wide variety of macrophages and established cell lines used as in vitro models, as well as the numerous conditions under which spore-macrophage interactions have been studied. In particular, the manners in which different cell lines interact with and internalize B. anthracis have not been directly compared. Our data indicate that RAW264.7 and J774A.1 cells, two commonly used macrophage-like cell lines, are significantly different in both the rate and amplitude of spore uptake (Fig. 8A). Although RAW264.7 and J774A.1 cells are both murine derived, with peritoneal-like properties (7, 28, 29, 33), our data suggest that these two cell lines are not similar in the manner in which they engulf spores. While the molecular basis of these differences between RAW264.7 and J774A.1 cells is not currently known, these results underscore the importance of carefully considering the choice of cells lines when establishing in vitro models.

Our studies also indicated that the presence or absence of FBS had no discernible effect on either the interactions with or uptake of B. anthracis by RAW264.7 cells at up to 4 h of serum deprivation. These results are especially interesting because they suggest that in vitro experiments may be conducted in highly defined medium, such as DMEM, to overcome potential complications arising from rapid spore germination or, conversely, the inhibition of spore germination that can occur in the presence of various sera (24, 47).

Extensive heterogeneity was revealed in the manner in which RAW264.7 cells interacted with B. anthracis spores (Fig. 3), which to our knowledge had not been documented previously. Spore uptake was not uniform, but instead, highly heterogeneous and characterized by subpopulations of infected and uninfected cells, as well as considerable variation in the numbers of spores associated with individual cells. Macrophage populations, even those derived from a single tissue, are known to be heterogeneous (19, 23, 35, 41). Within a single infection, macrophages exhibit cell-to-cell variation in their responses to pathogens (34). For example, heterogeneity in the interactions between macrophages and Listeria monocytogenes (40) has been interpreted as a direct reflection of the complex relationship between the host and the pathogen. Although poorly understood, several overall factors have been proposed to contribute to macrophage heterogeneity within a population (34), and this is an interesting and important topic that will be pursued in future research. Importantly, our results indicate that within a single infection, flow cytometry can be used to study macrophage responses as a function of the spore load, which can be determined by measuring macrophage-associated fluorescence derived from uptake of labeled spores.

In summary, using high-throughput flow cytometry-based approaches, we evaluated B. anthracis spore interactions with RAW264.7 cells at the single-cell level. We prelabeled spores in a way that did not alter several spore properties, including germination, outgrowth, and heat resistance. This strategy should be broadly applicable for studying the interactions of essentially any host cell with wild-type or mutant strains of B. anthracis. These studies provide the experimental framework for future work to investigate mechanisms of spore phagocytosis, as well as for quantitatively evaluating strategies for altering the extent to which host cells interact with and engulf B. anthracis spores.