Flow cytometry of sputum: assessing inflammation and immune response elements in the bronchial airways

Abstract

Background: The evaluation of sputum leukocytes by flow cytometry (FCM) is an opportunity to assess characteristics of cells residing in the central airways, yet it is hampered by certain inherent properties of sputum including mucus and large amounts of contaminating cells and debris.

Objective: To develop a gating strategy based on specific antibody panels in combination with light scatter properties for flow cytometric evaluation of sputum cells.

Methods: Healthy and mild asthmatic volunteers underwent sputum induction. Manually selected mucus “plug” material was treated with dithiothreitol, filtered and total leukocytes acquired. Multicolor FCM was performed using specific gating strategies based on light scatter properties, differential expression of CD45 and cell lineage markers to discriminate leukocytes from squamous epithelial cells and debris.

Results: The combination of forward scatter and CD45 expression reliably segregated sputum leukocytes from contaminating squamous epithelial cells and debris. Overlap of major leukocyte populations (neutrophils, macrophages/monocytes) required the use of specific antibodies (e.g. CD16, CD64, CD14, HLA-DR) that differentiated granulocytes from monocytes and macrophages. These gating strategies allowed identification of small populations of eosinophils, CD11c+ myeloid dendritic cells, B-cells and natural killer cells.

Conclusions: Multicolor FCM can be successfully applied to sputum samples to identify and characterize leukocyte populations residing on the surfaces of the central airways.

Clinical relevance: This research describes detailed methods to overcome difficulties associated with FCM of sputum samples, which previously has been lacking in the literature. FCM of sputum samples can provide valuable information on inflammation and immunological response elements in the bronchial airways for both clinical diagnostic and research applications and can be a useful tool in inhalation toxicology for assessing health effects of inhaled environmental pollutants.

Keywords::

• Induced sputum
• flow cytometry
• immunophenotype
• methods
• human

Acknowledgments

The authors wish to acknowledge and thank our highly skilled clinical and technical staff. These include Lynne Newlin-Clapp, Martha Almond, Carole Robinette, Margaret Herbst-Saunders, Aline Kala and Sally Ivins who performed sputum inductions, as well as Heather Wells, Fernando Dimeo, Danuta Sujkowski, Nolan Sweeney and Katherine Mills who processed the sputum samples and acquired data on the flow cytometer.

Declaration of interest

This research was funded in part by grants from the National Institutes of Health U19AI077437, R01-ES012706, RC1ES018417 and P01AT002620, as well as cooperative agreement CR 83346301 from the US Environmental Protection Agency. Although the research described in this article has been funded wholly or in part by the US Environmental Protection Agency through cooperative agreement CR 83346301 with the Center for Environmental Medicine and Lung Biology at the University of North Carolina at Chapel Hill, it has not been subjected to the Agency’s required peer and policy review and therefore does not necessarily reflect the views of the Agency, and no official endorsement should be inferred.

Appendix: extended methods and comments

Sputum sample collection, processing and factors affecting sample quality

Sputum induction and processing techniques vary and are described in detail elsewhere (Pizzichini et al., 1996; Hargreave et al., 1998; Kips et al., 1998a, b; Spanevello et al., 1998; Ronchi et al., 2002; Efthimiadis et al., 2002b). Methods used here are described in detail elsewhere (Alexis et al., 2000) and are based on the methods of Hargreave et al. (1998). The quality of the sputum sample has major impact on the quality of the flow cytometric assessment and can be influenced both by the induction process and subsequent processing of the sample. Sputum samples should immediately be placed on ice and processed within 2 h (Efthimiadis et al., 2002a). Careful processing of the sample is crucial and it is important that both sputum induction and processing be performed by well-trained and experienced personnel to maximize cell yield and minimize squamous epithelial cell and salivary contamination. Excessive or prolonged exposure to DTT during processing adversely affects both cell viability and surface marker expression (Efthimiadis et al., 1997; Loppow et al., 2000) and should be avoided. It is important to note that our laboratory employs the “plug selection” method where cell-enriched mucus “plugs” are manually selected from the surrounding clear saliva fluid and the total selected material is then processed, rather than the “whole sample” method, where the entire raw expectorated sample is processed. The main advantage of the plug selection method is that it minimizes both squamous cell contamination and dilution from saliva (Pizzichini et al., 1996; Spanevello et al., 1998; Alexis et al., 2006).

Construction of antibody panels

When constructing antibody panels for sputum analysis, it is essential to include anti-CD45 (leukocyte common antigen, pan leukocyte marker) in all sample tubes to help differentiate leukocytes from debris and facilitate identification of leukocyte populations. We take advantage of major lineage markers such as CD16, CD14, HLA-DR, CD3, CD19, CD56, as well as differential expression of other population-specific markers as an aid in isolating specific populations and subpopulations. Certain specific antibodies are included in multiple sample tubes to aid in identifying specific populations for gating purposes. Using these principles, we have developed “standard” panels of fluorescent antibodies (see manuscript Table 2) for identifying leukocyte populations and assessing expression of select cell surface proteins associated with innate and adaptive immunity, antigen presentation and inflammation. Many of the innate immune proteins also interact with the adaptive immune system.

Cytometers equipped with multiple lasers are capable of simultaneously using many more than five fluorochromes incorporated in our standard 5-color panel. It would be desirable to incorporate more fluorochromes in order to reduce duplication of certain markers, consolidate tubes and conserve cells. We have done this for certain of our more specific protocols, however, it is often difficult (short of labeling antibodies in-house) to procure the desired antibody labeled with the necessary fluorochrome to accomplish this. Individual investigators can tailor antibody panels to incorporate additional fluorochromes and to fit their specific research objectives.

Staining procedures

Cells should be labeled (stained) in a timely manner (as soon as possible) following sample processing and kept at 4°C in the dark until acquired on the flow cytometer. Data acquisition should be done within 24 h for best results (Stewart et al., 2007). The specific antibody staining procedures have been described elsewhere (Alexis et al., 2000). We routinely label 100 µl of cell suspension (1 × 10⁶ cells/ml) based on total non-squamous nucleated cells (i.e. leukocytes and bronchial epithelial cells). Following staining, cells were fixed in 1% paraformaldehyde in DPBS. Sources for specific antibodies used in these panels are listed in Table A1.

Specific antibodies were titrated to determine an appropriate amount for staining. The majority of the antibodies were still on the plateau of the titration curve when used at half the manufactures suggested amount (i.e. 10 µl vs. 20 µl). A few (e.g., CD16, CD3) required further dilution. Individual investigators should titrate their antibodies to determine appropriate concentrations for their specific applications. Isotype antibodies also were titrated; however, when we attempted to match protein concentrations and fluorochrome/protein concentrations, we found that the isotype controls had a much higher MFI signal than did certain of the specific antibodies, particularly for the IgG2 isotypes. We therefore found it necessary to dilute isotype control antibodies up to 1:5 (or more). The cause of this disparity is not entirely clear (Hulspas et al., 2009). Some labs have advocated doing away with isotype controls for this and other reasons (Keeney et al., 1998) whereas others do not (O’Gorman & Thomas, 1999). It may also be useful to employ non-specific unlabeled blocking antibody to help minimize non-specific staining.

Alternatively, a “cleaner” method for determination of background fluorescence might be the use of FMO analysis (Tung et al., 2004); however, to be done properly, this requires large numbers of cells when multiple populations, multiple specific antibodies, multiple fluorochromes and two different isotypes (IgG1 and IgG2) are used. Because the number of leukocytes obtained from sputum samples is often limited, the FMO approach is usually not feasible for the typical sputum sample. Substitution of blood leukocytes for estimating background fluorescence (Dua et al., 2010) may be adequate for sputum lymphocytes, granulocytes and monocytes; however, the absence of macrophages in blood is an obvious problem; although this might be overcome by “spiking” sputum isotype control tubes with blood leukocytes. In our experience, however, sputum monocytes tend to be more granular (higher SSC) with moderately higher autofluorescence (especially in the FITC channel) than peripheral blood monocytes.

Determination of background fluorescence using isotype controls

Gating populations in the isotype control tubes is more complex for sputum than for blood as, unlike blood, major non-lymphocyte populations overlap and small or rare populations are imbedded within larger populations. More precise determination of background fluorescence for these small populations requires inclusion of specific antibodies (e.g. CD9 or CD16 (EOS), CD203c (basophils), or CD14 (DCs)) in addition to CD45 in the isotype control tube.

MFI of isotype controls (Tubes 1 and 2, Table 2) for specific populations was subtracted from the MFI measured for specific markers to control for background autofluorescence and non-specific fluorescence. For individual populations, background fluorescence was determined simultaneously for all fluorescence channels by creating a single histogram and gating 99% of the population (95% at the minimum) in only one fluorochrome channel, usually FITC. As illustrated for the monocyte population (Table 3), background MFIs for the various fluorochromes were relatively constant regardless of which channel was used for gating. This is also true for the other leukocyte populations. This approach avoided the creation of a very large number of histogram gates.

Instrumentation

The majority of our samples are acquired on a BD^™ LSR-II digital flow cytometer equipped with 405 nm solid state, 488 nm argon-ion and 633 nm Helium-Neon lasers, appropriate filters and capability for cross-laser compensation (BD Biosciences Immunocytometry Systems). Compensation for spillover and spectral overlap are set for a particular set of instrument settings using BD^™ CompBeads (BD Biosciences, cat 552843) and an automated compensation algorithm in BD^™ FACSDiva 6.1 software (BD Biosciences). Once established, compensation values are not changed unless the established parameter settings are altered. Our “standard” panel uses five colors exciting off the 488 and 633 lasers; however, additional fluorochromes exciting off the 405 laser are used in some studies. This is determined by the availability of specific antibodies in a particular fluorochrome and the amount of sample available. Although it is true that less sample is required when more fluorochromes are employed in a particular panel, the trade-off is that the complexity of instrument setup, compensation and data analysis also increases.

Table A1.  List of antibodies and cellular expression of target molecules.^a

Instrument setup, data acquisition and special considerations for sputum samples

The goal in establishing appropriate instrument settings is to establish the best signal-to-noise ratio which will provide a measurable signal, while minimizing the coefficient of variation of that signal. Optimal PMT voltages for samples containing mixed populations (e.g. blood) are usually established based on the population having the lowest autofluorescence (i.e. lymphocytes) and can be accomplished using cells or beads with low-range fluorescent intensity (BD Application Note, 2000). The high autofluorescence of sputum macrophages, relative to lymphocytes, requires a compromise to allow application of a single set of PMT voltages for all populations. Our approach has been to optimize instrument settings for monocytes. The monocyte population, labeled only with CD45, is displayed in a separate histogram for each fluorochrome and voltages adjusted to place the entire monocyte population between 200 and 2000, which will place macrophages much higher on the log scale (between 5,000 and 10,000) and lymphocytes somewhat lower, perhaps even straddling zero. In the case of BD^™ FACSDiva 6.1 software, bi-exponential plots allow the display of events with fluorescent intensity below zero. It is best, however, to use a setup optimized for lymphocytes if they are of primary interest.