|Home | About | Journals | Submit | Contact Us | Français|
E.F. contributed to the conception, design, performance and analysis of the protocols and wrote the manuscript. L.E. contributed to the configuration and establishment of the flow cytometer in the laboratory and to the writing of the manuscript. J.A.T. participated in the conception of the protocols and the writing of the manuscript. L.S.W. contributed to the conception, design and analysis of the protocols and the writing of the manuscript.
We describe two modular protocols for immunostaining and multiparameter flow cytometric analysis of major human antigen-presenting cells (dendritic cells, monocytes, B lymphocytes) in minimally manipulated whole blood. Simultaneous detection of up to eight colors is enabled by careful selection and testing of cell-subset-defining monoclonal antibodies (anchor markers) in the appropriate fluorochrome combinations, to demonstrate the quantification of surface expression levels of molecules involved in chemotaxis (e.g. CX3CR1, CCR2), adhesion (e.g. CD11b, CD62L), antigen presentation (e.g. CD83, CD86, CD209) and immune regulation (e.g. CD101) on circulating antigen-presenting cells. Each immunostaining reaction requires as little as 50–100 μl of peripheral whole blood, no density-gradient separation, and the entire procedure from preparation of reagents to flow cytometry can be completed in <5 h.
Antigen presentation is an innate immune function that is crucial in self/non-self discrimination and in the pathogenesis of most immune-mediated diseases. The chief immune cells endowed with this function, known also as professional antigen-presenting cells (APCs), include dendritic cells (DCs), macrophages, monocytes and B lymphocytes. DCs, macrophages and monocytes, including immature monocytes, are particularly competent in antigen capturing and presentation and in activating antigen-specific T-cell responses through major histocompatibility complex class II (MHC II) molecules1-7. MHC II molecules are encoded in the highly genetically polymorphic human leukocyte antigen (HLA) region that produces multiple molecules required for antigen presentation, including HLA-DR, a MHC II molecule predominantly and abundantly expressed by APCs. Because APCs contribute crucially to the pathogenesis of a range of diseases, detailed characterization of their functional status would be valuable in devising strategies for diagnosis, immunologic monitoring and therapy and would furnish important insight into mechanisms of APC mobilization. More generally, the human immune system is poorly understood8,9, and there is an urgent need to improve our knowledge using the latest technologies.
Phenotypic characterization of antigen-presenting cells has been hampered by the large number of surface markers required to identify a subpopulation or subset, and furthermore limitations of conventional three- to four-color flow cytometers are major drawbacks1,10,11. Advantages of single-cell immunophenotyping of blood antigen-presenting cells by flow cytometry have been demonstrated and proven feasible12, offering relatively high sensitivity and statistical power compared with non-flow-based technologies. Polychromatic (>6 colors) flow cytometry increases the data-capturing throughput and enables concurrent, multiplexed analysis of cell subsets and populations in the same reaction tube (e.g. calculation of percentages or ratios of cell subsets, parallel comparison of surface expression density of a molecule between two cell populations), as well as a qualitative assessment and characterization of the functional status of cells13,14. In conjunction with the use of minimally manipulated leukocytes derived from erythrocyte-lysed fresh blood, instead of density-gradient-enriched mononuclear cells, immunophenotypic data can be captured with minimal introduction of artefacts or factors that affect immunostaining15.
Selection and rigorous testing of monoclonal antibodies and fluorochromes are crucial to successful polychromatic flow cytometry. Recently, Jansen et al. reported an eight-color immunostaining protocol for flow cytometry that uses MHC II, CD14, CD11c and CD123 with/without CD19 as cell-subset-defining markers16. Our early experience with eight-color flow cytometry using six subset-defining monoclonal antibodies concomitantly (targeting CD14, CD16, CD19, CD11c, CD123 and HLA-DR) indicated that an accurate definition of DC subsets is better achieved using the stringent Lineage cocktail 1 (Lin1)–FITC negative selection markers or antibodies (BD Biosciences) that target CD3, CD14, CD16, CD19, CD20 and CD56; these lineage markers are all undetectable on the two major DC subsets. In other words, a lack of expression of Lin1 effectively excludes T lymphocytes, CD14+ monocytes, CD14+CD16+ monocytes, CD16+ monocytes, polymorphonuclear neutrophils (PMNs), B lymphocytes and natural killer cells. Moreover, CD11c (integrin αX) and CD123 (IL-3 receptor-α), without the exclusion of lineage markers, are not DC-specific as they are expressed by a variety of myeloid and lymphoid cells.
Therefore, to maximise the capabilities of an eight-color flow cytometer (BD LSR II with blue, red and violet lasers) while preserving the accurate identification of monocyte subsets, B lymphocytes and DC subsets in addition to retaining the capacity to analyse three to four target molecules of interest, we have devised two protocols with similar workflow, Protocols A and B (Fig. 1), that have undergone testing and ascertainment. The two immunostaining protocols are modular and can be easily performed in parallel, or independently, to define the major circulating antigen-presenting cell subsets in whole blood: CD14+, CD14+CD16+ and CD16+ monocytes; CD19+HLA-DR+ B cells; and myeloid (MDCs or preDC1) and plasmacytoid (PDCs or preDC2) DCs. Performing both protocols concurrently can reduce the elapsed time from phlebotomy to immunostaining (‘arm-to-antibody’ time) and minimize any potential time-dependent in vitro confounding effects when studying innate cellular attributes that are intrinsically volatile. The use of predeveloped, standardized fluorescent beads for normalization of fluorescence intensity data as well as for periodic fine-tuning of the flow cytometer17,18 can reduce non-biological and/or instrumental variation arising from day-to-day operation of the flow cytometer, especially for multicenter and/or longitudinal immunophenotyping studies.
Monocytes are progenitors for macrophages and certain DCs, and two principal subpopulations have been described: CD14+HLA-DR+ (‘inflammatory’) and CD16+HLA-DR+ (‘resident’) monocytes19,20. Further characterization based on their differential expression of chemokine receptors and adhesion molecules identified inflammatory CD14+ monocytes as CX3CR1lowCCR2+CD62L+CD11bhigh, and resident CD16+ monocytes as CX3CR1highCCR2−CD62L−CD11bint 2,20. CD14+ and CD16+ monocytes differ in size [~18 μm or intermediate-high forward scatter channel (FSC), and ~14 um or low-intermediate FSC, respectively], cytoplasmic granularity [relatively higher side scatter channel (SSC) and lower SSC, respectively] and in their functional capacities1,2,19,20 (see also Fig. 2a for FSC and SSC comparisons). Although less well understood, CD14+CD16+ monocytes are another subset, possibly in a transitional or intermediary state, that is expanded and active in inflammatory disorders6,21-23. Even in the absence of an exogenous stimulus, the CD14+CD16+ subset differentiates within 12 h from CD14+ monocytes, and therefore we recommend proceeding with the immunostaining within 4 h of blood donation and to rotate the blood collection tubes if possible to avoid subsequent miscalculation of monocyte subset ratios and changes in the expression levels of certain surface molecules (Fig. 3).
MDCs and PDCs derive from different hematopoietic lineages and have divergent physiological functions7,24-29. In conjunction with positive expression of MHC II molecules such as HLA-DR and the absence of Lin1 markers, MDCs and PDCs can be further classified as Lin1−HLA-DR+CD11c+ and Lin1−HLA-DR+CD123+, respectively. CD303 (BDCA-2) can be used alternatively to identify PDCs, whereas CD1c+ (BDCA-1) and CD141+ (BDCA-3 or thrombomodulin) are markers that identify additional MDC subsets30-32. CD303 is a PDC-specific C-type lectin involved in antigen uptake, processing and presentation to T cells and regulates type I interferon release, a cell-defining attribute of PDCs33.
Neutrophils can be identified and distinguished from monocytes, macrophages and DCs by virtue of their high FSC and SSC11 – thus, the use of a different gating strategy (see also Fig. 2a) and immunophenotypic profile: CD16highCX3CR1−CD62L+CD11bhigh. In line with recent findings, neutrophils differentiate from the CX3CR1-negative lineage of myeloid progenitors that diverge from the macrophage-DC precursors early in hematopoiesis29. Investigators interested in this cell population can maximize the yield of immunophenotyping data by gating on the granulocyte cluster (Fig. 2a).
Protocol A details the use of eight fluorochrome-conjugated monoclonal antibodies (or ten parameters) to characterize in whole blood CD14+, CD14+CD16+ and CD16+ monocyte and B-cell surface expression of chemokine receptors (e.g. CX3CR1 or fractalkine receptor, CCR2), adhesion molecules (e.g. CD11b or integrin-αM, CD62L or L-selectin) and costimulatory/accessory antigen presentation molecules (e.g. CD83, CD86, CD209 or DC-SIGN).Determination of monocyte subset ratios is also shown. Protocol B takes advantage of the Lin1 immunostaining strategy focusing on the identification of CD11c+ MDCs and CD123+ PDCs, with demonstration of the alternative and/or concomitant use of the DC markers CD1c and CD303 in combination with Lin1−HLA-DR+ to ascertain the identity of MDCs and PDCs, respectively. Specific to this protocol is the use of a violet laser-excitable fluorochrome, particularly Pacific Blue (peak excitation/emission of ~410/455 nm) or an equivalent such as eFluor 450 (peak excitation/emission of ~405/450 nm), conjugated with anti-CD123 antibody that opens up other commonly used channels (e.g. PE, Alexa Fluor 700, APC-Alexa Fluor 750) for use.
Isotype control immunoglobulins are not needed for quantitative immunophenotyping18. In a recent large-scale, gene-immunophenotyping study of nearly 200 individuals that we carried out14, we eliminated the use of the isotype controls after ascertaining beforehand that the monoclonal antibodies are specific. Isotype control immunoglobulins are liable to fluctuate over time, and the use of fluorescent microspheres is a solution to the time-dependent effects or variability observed in longitudinal studies18. We recommend only the initial use of isotype-matched control immunoglobulins to establish that a target antibody is specific, then proceed to longitudinal immunophenotyping using fluorescent microspheres to normalize the acquired data.
When performing large-scale, quantitative immunophenotyping by polychromatic flow cytometry over a prolonged period (i.e. on the order of months), the investigator should have a precise hypothesis a priori, determine the power of the study and calculate the sample size required, to enable estimation of the amount of antibodies and reagents required. Apart from instrumental variation that can be remedied by using standardized fluorescent microspheres18, a batch-to-batch effect of reagents is probably another major source of variability. In a carefully planned study with a known number of immunostaining reactions, the investigator can then order an oversized batch and pool the antibodies to circumvent batch variation.
Each immunostaining reaction requires as little as 50–100 μl of fresh whole blood without prior density-gradient separation. This method provides a cost-effective, practical approach to high-throughput polychromatic immunophenotyping of the major APC subsets in whole blood and serves as a framework upon which to incorporate additional analysis parameters, depending on the flow cytometer setup and fluorochromes used.
Our BD LSR II is equipped with three lasers: violet (405 nm excitation by Coherent VioFlame, 25 mW solid-state laser), blue (488 nm excitation by Coherent Sapphire, 20 mW solid-state laser) and red (633 nm excitation by JDS Uniphase 1344P, 17 mW HeNe laser). Bandpass filters 440/40 and 525/50 are configured with the violet laser; 530/30, 575/26, 695/40 and 780/60 with the blue laser; 660/20 and 730/45 and 780/60 with the red laser. The specifications of long-pass dichroic mirrors used were 505 nm, 550 nm, 685 nm, 710 nm and 755 nm (see Table 1). When frequent, in-house calibration and fine-tuning of flow cytometer as detailed by Perfetto et al.17 is not possible, regular (quarterly) evaluation and servicing of the flow cytometer by the manufacturer’s engineer should be scheduled. ! CAUTION Internal components of a flow cytometer pose serious health hazards and should not be tampered with by unqualified personnel.
1 Label 2.0-ml microcentrifuge tubes. ▲ CRITICAL STEP 1.5-ml microcentrifuge tubes can be used if reaction volumes are scaled down proportionately. However, the yield of rare cell populations might be low. When working with large numbers of staining reactions, 2-ml 96-well Polypropylene Deep-Well Plates in conjunction with adhesive sealing film are useful; avoid contact of blood with adhesive surface. We do not recommend the use of polystyrene FACS tubes for immunostaining APCs as these cells might become activated by this surface substrate. When polypropylene FACS tubes are used, exercise care when discarding the supernatant as the pellet can be difficult to visualize. Use sterile plasticware.
2 Prepare fresh 1x BD FACS Lysing Solution and 1x BD CellFIX solution by diluting 1:10 the respective 10x solutions with sterile, endotoxin-free H2O prior to use. Keep 1x FACS Lysing Solution at 15–20°C, and CellFIX solution and 1x PBS on ice. ▲ CRITICAL STEP Keep all solutions and antibody mixes sterile.
3 Prepare the antibody master mix for immunostaining of monocyte subsets and B lymphocytes (see Protocol A) and/or DC subsets (see Protocol B).
4 Pipet 100 μl of CPDA-treated fresh whole blood into each tube. If blood droplet is caught on the tube sidewalls, samples can be spun down gently by pressing the ‘Short’ button on the Eppendorf Centrifuge 5417R for 2 s (≤350 g).
5 Add antibodies to whole blood. For immunostaining of monocyte subsets and B lymphocytes see Protocol A. For immunostaining DC subsets see Protocol B.
6 While Protocol A and Protocol B antibodies are incubating with blood samples, perform single-color staining for later fluorochrome spectral overlap compensation by applying the same, single antibody that was used in the master mix for each individual fluorochrome. Incubate samples in the dark for 30 min at 15–20°C. Omit this step when using standardized fluorescence-labeled microspheres for setting spectral compensation (see also PROCEDURE – Acquisition). ▲ CRITICAL STEP Performing single-color staining for all fluorochromes each time when immunophenotyping is performed can also serve as a quality control and help to troubleshoot and gauge the quality of the antibody and the staining.
7 Following incubation of the antibody mix with blood, lyse erythrocytes by adding 2.0 ml of 1x FACS Lysing Solution per 100 μl of blood used. Incubate for 10 min at 15–20°C with inversion of tubes every 2–3 min. ▲ CRITICAL STEP Ensure cells are well resuspended for proper lysis.
8 After incubation, centrifuge at 350 g for 7 min at 15–20°C.
9 Carefully pipet off and discard supernatant. Resuspend cell pellet in 1.0 ml of cold, endotoxin-free 1x PBS for each 100 μl whole blood used and mix well.
CRITICAL STEP Perform all subsequent steps at 4°C.
10 Centrifuge at 350 g for 7 min at 4°C.
11 Carefully aspirate off and discard supernatant. Resuspend in 300 μl of 1x BD CellFIX. ■ PAUSE POINT Fixed cells can be kept at 4°C for up to 24 h prior to analysis by flow cytometry.
12 Transfer fixed samples from microcentrifuge tubes to 10-ml FACS tubes.
13 Switch on flow cytometer and allow it to warm up for 30 min. Open FACSDiva software program.
14 Calibrate PMT voltages by first creating a FSC vs SSC bivariate plot in FACSDiva to visualize cell clusters, then create one-dimensional histograms for each fluorochrome used in the experiment, and set voltages by acquiring blank and fluorescent beads (DakoCytomation FluoroSpheres) gauged on histograms according to the manufacturer’s instructions. In addition, if desired, use an unstained sample of fixed cells (see Step 3 above) to establish the PMT voltages so that the majority of events (cells) contained within a gate encompassing monocyte or lymphocyte clusters (i.e. clusters that are not granulocytes, dead cells or cell debris) are in the first decade (log) of fluorescence. Benchmark against blank beads (Dako Cytomation) to monitor the autofluorescence of the target cell population on a day-to-day basis.
15 Determine fluorochrome pair compensations by first acquiring samples from single fluorochrome–antibody staining. Use bi-exponential (or ‘Logicle’) display of x-axis and y-axis when determining amount of compensation35. Set compensation now or after collection of all samples, depending on capabilities of flow cytometer and software capabilities; automatic compensation might be available on some platforms. An orderly workflow including the systematic organization of bivariate plots displaying antibody–fluorochrome pairs is essential to successful polychromatic flow cytometry (Fig. 4). For literature on flow cytometry instrument setup and fluorochrome compensation, investigators are referred to excellent reviews published by others10,33-39. ▲ CRITICAL STEP Each target antigen can have a drastically different cell-surface distribution and biophysical properties, requiring configuration of a unique set of fluorochrome compensation values. Use an abundantly expressed surface epitope when setting compensation.
16 Begin acquisition of data.
17 After acquisition, export data in the *.fcs format and analyze using the FlowJo software (see also ANTICIPATED RESULTS and Figures; for latest information on cross-platform compatibility of FlowJo files, check the FlowJo website.). For monocytes follow option A. For B lymphocytes follow option B. For DC subsets, follow option C.
The gating of monocytes on a FSC vs SSC bivariate plot is important for identifying the monocyte subsets that have different sizes and granularity1,2. CD16+ monocytes have lower FSC and SSC, and are situated close to the lymphocyte cluster (Fig. 2a, green oval), whereas CD14+ monocytes, owing to their larger numbers, size (higher FSC) and higher cytoplasmic granularity (higher SSC), cluster near granulocytes (Fig. 2a, red oval). The former subset can be missed or its percentages underestimated if the gating is too conservatively placed on only the latter subpopulation. Within the monocyte gate, HLA-DR+ cells (Fig. 2b) are separated based on expression of CD14 and CD16 and are displayed on a quantile contour (zebra) plot that can also be used to calculate relative ratios or percentages of CD16+ (Fig. 2c, green box), CD14+CD16+ (Fig. 2c, yellow box) and CD14+ (Fig. 2c, red box) subsets. We have found the zebra display useful in guiding the gating of monocyte subsets. Histograms showing expression profiles of the monocyte subsets are consistent with those reported previously: CD16+HLA-DR+ monocytes are CX3CR1highCCR2− CD11blowCD62L− (Fig. 2d, green tracings), and CD14+HLA-DR+ monocytes are CX3CR1lowCCR2+CD11bhighCD62L+ (Fig. 2d, red tracings)20. CD19+HLA-DR+ B cells do not express CX3CR1 or CCR2, but the majority do express CD62L. A recent study reported that ~34% (19–44%) of B cells in peripheral blood of healthy human donors express CD11b on the cell surface, and that CD11b mediates homing of memory B cell subsets, in contrast to naïve cells40. Accumulating evidence supports the divergent lineage and functions of CD11b+ and CD11b−/low B cells40-42. Another analysis of monocyte subsets and CD19+HLA-DR+ B cells is demonstrated by CD86–PE, CD83–APC, CD209–PerCP-Cy5.5 through use of HLA-DR–FITC instead of HLA-DR–PerCP (Fig. 2). Although the expression of CD209 is negligible or undetectable on resting, circulating monocytes and HLA-DR+CD19+ B cells from a healthy blood donor (Fig. 2h and 2i), the functionality of this fluorochrome-labeled antibody is demonstrated by its increased expression on CD11c+HLA-DR+ cells in the monocyte gate following cytokine activation (Fig. 2j). We advise immunostaining within 4 h of blood donation, as changes in the monocyte subset dynamics occur with time (Fig. 3). Protocol A utilizes ten parameters or eight fluorochromes (FITC, PE, PerCP or PerCP-Cy5.5, Alexa Fluor 700, APC, APC-Alexa Fluor 750, Pacific Orange and Pacific Blue).
This protocol is a modification of the four-color staining reported by Willmann et al.12 and others. Because four fluorochrome channels (typically, Lin1–FITC, HLA-DR–PerCP, CD11c–APC and CD123–PE) are required for separating MDCs and PDCs, this limitation precludes the use of antibodies of interest that are commonly available in PE and APC. Using a three-laser flow cytometer (BD LSR II) and a violet-laser-excitable fluorochrome (e.g. anti-CD123 conjugated with Pacific Blue), we were able to maximize immunostaining combinations and attain seven-color multiplexed analysis of two major blood DC subsets; there exist future opportunities for robust eight-color analyses with the increased availability of monoclonal antibodies coupled with new fluorochromes taking advantage of the violet laser such as eFluor 605 and eFluor 650 monoclonal antibodies (eBioscience). In this protocol, Lin1 and HLA-DR are indispensable for identifying DCs (Fig. 5a and 5b) that can be subclassified into MDCs (Lin1−HLA-DR+CD11c+) and PDCs (Lin1− HLA-DR+CD123+) (Fig. 5d). CD1c and CD303 are alternative markers that can identify Lin1− HLA-DR+ MDC and PDC, respectively, (Fig. 6a and 6b) but should not be regarded as equivalent to CD11c and CD123 in view of the tremendous DC heterogeneity26,27,30-32,44. MDC:PDC ratios or percentages (Fig. 5c) have been reported to alter in disease states; however, gating strategies and the use of DC-defining markers can vary between studies45,46. The CD11c–CD123 bivariate plot of Lin1−HLA-DR+ cells (Fig. 5c and Fig. 5a, top row) is a common strategy used to simultaneously display ratios between the two DC subsets44,45. For quantification of expression levels (i.e. in mean or median fluorescence intensity units), we find that the CD11c+HLA-DR+ and CD123+HLA-DR+ plots (Fig. 5d) generated from Lin1− cells (Fig. 5b, pink box) are preferable as they provide a clear visualization of DC clusters (Fig. 5d, orange and purple boxes) and other cells or events; this is also the method demonstrated in the product sheet provided by the manufacturer of the Lin1 mixture antibodies (BD Biosciences). Protocol B has the capacity for at least nine parameters or seven fluorochromes (FITC, PE, PerCP, APC, APC-Alexa Fluor 750, Alexa Fluor 700, Pacific Blue). Using CD101 expression profiling of circulating DCs as an example (Fig. 5f), we have concurrently identified positive CD101 surface expression on MDCs but non-expression on PDCs as a novel strategy for differentiating the two DC subsets, findings that might have implications also in their differential immunomodulatory functions31,47-50,52.
Surface markers useful for the identification of circulating APC subsets are summarized in Table 3. As the complexity of polychromatic flow cytometry increases, multidimensional analyses of the data by computational modelling might one day replace manual gating and increase the throughput and objectivity of analyses53. Comprehensive information on flow cytometer setup, fluorochromes, spectral overlap compensations, fluorescent bead normalization for longitudinal immunophenotyping studies, operational pitfalls and other useful technical information are detailed elsewhere10,18,34-39. Further links to major manufactures of flow cytometers, accessories and fluorochrome-conjugated antibodies can be found at the end of an article by Perfetto et al.11.
We thank the Wellcome Trust, the Juvenile Diabetes Research Foundation (JDRF) International, and the National Institute for Health Research Cambridge Biomedical Research Centre for funding. The Cambridge Institute for Medical Research is a recipient of the Wellcome Trust Strategic Award (079895). E.F. was funded by the JDRF and the Prince Philip Graduate Exhibition from the Cambridge Overseas Trust. L.S.W. is a Wellcome Trust Principal Research Fellow.
COMPETING INTERESTS STATEMENT
The authors declare that they have no competing financial interests.