two-color immunofluorescence - Semantic Scholar

14 downloads 0 Views 1MB Size Report
PAUL KARL HORAN*, SUE E. SLEZAK, AND GEORGE POSTE. Department of Cell Biology, SmithKline and French Laboratories, 1500 Spring Garden Street, ...
Proc. Natl. Acad. Sci. USA Vol. 83, pp. 8361-8365, November 1986 Medical Sciences

Improved flow cytometric analysis of leukocyte subsets: Simultaneous identification of five cell subsets using two-color immunofluorescence (automated differentials/monoclonal antibodies)

PAUL KARL HORAN*, SUE E. SLEZAK, AND GEORGE POSTE Department of Cell Biology, SmithKline and French Laboratories, 1500 Spring Garden Street, Philadelphia, PA 19101

Communicated by Leopold Pospisil, July 21, 1986

Flow cytometric analysis of human peripherABSTRACT al blood leukocytes has typically been achieved by staining multiple aliquots of the same sample with fluorescent reagents specific for cell subsets of interest. Spectrally discrete fluorochrome tags have been developed for applications in which identification of multiple subsets (e.g., T and B cells) or of subsets not uniquely identified by a single reagent (e.g., activated T cells) requires use of multiple reagents per aliquot. Extension of this approach to more than two reagents per aliquot has led to multicolor methods requiring dual laser excitation and complex instrumentation. We describe an alternative two-color method using commercially available reagents that allows simultaneous identification of five discrete immune cell subsets using only a single excitation source. The technique uses dilution of commercial fluorochrome-labeled reagents with competing unlabeled reagents to selectively produce discrete fluorescence intensity profiles for cell subsets that would otherwise display overlapping or indistinguishable profiles when stained with reagents bearing the same fluorochrome. For example, the fluorescence intensity of phycoerythrin-labeled helper T (Th) cells can be adjusted to be distinct from that of phycoerythrin-labeled suppressor T (T) cells. Extending this technique to two colors, we have used a combination of seven different monoclonal antibodies to simultaneously quantify Th, T5, B cells, natural killer cells, and monocytes in a single aliquot. An additional advantage of this approach is the ability to more accurately quantify "null" cells. Adjustment of fluorescence intensity profiles of different cell subsets by this method is applicable to flow cytometric analysis of a wide variety of cell types. The technique significantly extends the analytical capacity of flow cytometry without significantly increasing the complexity of the instrumentation required.

Flow cytometry is used widely as a powerful tool for the identification and quantification of leukocyte subsets in peripheral blood in experimental and clinical immunology (1-4). Specific cell subsets bearing characteristic cell-surface antigens are identified using antibodies labeled with fluorochromes such as fluorescein (5), rhodamine (6), Texas red (7), or phycoerythrin (PE) (8). These reagents bind to antigen-positive cells, which are then detected as a fluorescent population(s), while nonlabeled cells fail to fluoresce. Mixed populations can be analyzed at rates of 300,000 cells per min, in modem instruments, affording statistical precision because of the large sample size. Although a large panel of monoclonal antibodies is available that permits accurate identification of immature and mature leukocytes, and also various neoantigens expressed by neoplastic leukocytes, the number of leukocyte subsets that can be identified in a single

blood sample is at present limited by the lack of spectrally distinguishable fluorochromes. To use a single excitation source, the chromophores used to label different antibodies must exhibit similar excitation spectra but differ sufficiently in their emission spectra to permit detection at different regions of the spectrum. By using a mixture of antibodies labeled with two fluorochromes that display distinct emission spectra, such as fluorescein and PE, simultaneous identification of only three cell subsets can be achieved using a single laser for excitation. To detect additional cell subsets within a single preparation it is necessary to use a third chromophore, requiring either a second laser for excitation or complex filter windows and color compensation systems to correct for spectral overlap. We report a method that allows simultaneous detection of two subsets of leukocytes for each chromophore used and up to five cell subsets using two-color fluorescence and a single laser. This method permits simultaneous quantification of T-helper (Th) cells, T-suppressor (TJ) cells, B cells, monocytes, and natural killer (NK) cells from the same blood sample. Identification of these, and other cell types, offers a significant expansion in the capacity of flow cytometric analysis without the need for new reagents or expensive modifications in instrument design.

MATERIALS AND METHODS Isolation of Mononuclear Cells. Human peripheral blood was collected from four normal donors in Vacutainers containing sodium heparin (Vacutainer Systems, Rutherford, NJ). Mononuclear cells were isolated by layering 8 ml of whole blood on 5 ml of Lymphoprep sodium metrizoate/ Ficoll separation medium (Nyegaard, Oslo) and centrifuging at 400 x g for 40 min at 20'C. The interface layer containing lymphocytes and monocytes was aspirated and washed twice in a phosphate-buffered saline solution (PBS/BSA/Az) containing 1% bovine serum albumin (Sigma) and 0.05% sodium azide (Fisher). The cells were resuspended in the PBS/BSA/ Az buffer, counted, and adjusted to a final concentration of 2 x 107 cells per ml. Labeling of Leukocytes with Monoclonal Antibodies. Fluorescein isothiocyanate (FITC)-conjugated anti-human B1 (B-lymphocyte marker; FITC-B1) and FITC-conjugated antihuman Mo2 (monocyte marker; FITC-Mo2) were obtained from Coulter. Unconjugated and FITC-conjugated antihuman Leu-4 (T-lymphocyte marker; FITC-Leu-4), PE-conjugated anti-human M3 (monocyte marker; PE-M3), PEAbbreviations: Th, T-helper; T, T-suppressor; PE, phycoerythrin; NK, natural killer; FITC, fluorescein isothiocyanate; BSA, bovine serum albumin. *To whom reprint requests should be addressed at: Associate Director of Cell Biology (L102), SmithKline and French Laboratories, 709 Swedeland Road, Swedeland, PA 19406.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

8361

8362

Medical Sciences: Horan et al.

Proc. Natl. Acad. Sci. USA 83

(1986)

conjugated anti-human Leu-2a [suppressor T-lymphocyte A (TJ) marker; PE-Leu-2a], unconjugated and PE-conjugated anti-human Leu-3a [helper T-lymphocyte (Th) marker; PELeu-3a], and PE-conjugated anti-human Leu-llc (NK marker; PE-Leu-llc) antibodies were purchased from BectonUnstained Dickinson. To obtain the desired separation of the five cell subsets identified by these antibodies, the Leu-4 and the Leu-3a antibodies were diluted with unconjugated antibodies with Ts cells identical specificity (same clone). The Leu-4 reagent was diluted by mixing 3 1.l offluorescein-labeled reagent with 1 Al of unconjugated antibody and the Leu-3a was diluted at a ratio of 13.0 A.l of PE-labeled reagent to 1.5 ,u1 of unconjugated antibody. Cells (1 x 106 leukocytes per 50 ILI) were stained with B labeled antibodies in 96-well V-bottom plates at 40C for 30 min. The final preparation for simultaneous labeling of all five leukocyte subsets contained the following antibodies: S A.l of U FITC-B1; 5 ,.l of FITC-Mo2; 20 Al of PE-M3; 20 1ul of Unstained .0 _ h E~~~ PE-Leu-11c; 20 /4 of PE-Leu-2a; 15 Al of diluted Leu-3a;~cells and Tcells C 5 ,ul of diluted FITC-Leu-4. Control wells were prepared by 100 using the appropriate individual antibodies or antibody comU binations. PBS/BSA/Az buffer was added to test and control Tcells wells to bring the total volume in each well to 90 ul. After incubation, 50 Al of PBS and 20 ,4 of fetal bovine 0.~~~~~C serum were added sequentially to each well. Plates were centrifuged at 400 x g for 10 min at 4°C, supernatants were aspirated, and the cell pellets were resuspended in 200 ,ul of C PBS/BSA/Az for flow cytometric analysis. Flow Cytometric Analysis. Leukocyte cell subsets were 200 quantified by using an EPICS 753 flow cytometer (Coulter) with an exciting wavelength of 488 nm at 500-mW power. A 4- Th cells 488-nm dichroic mirror and 488-nm bandpass filter were used to measure the right angle light scatter (RALS) signal. A 515-nm interference filter and a 515-nm long-pass absorbance 100 Non-T cells T. cells filter were used to block the excitation light from the fluorescence detectors. A 560-nm short-pass dichroic mirror was used to split the FITC/PE signal. A 575-nm bandpass filter was used in front of the PE photomultiplier. A 525-nm bandpass filter was used in front of the fluorescein 00 250 200 150 100 50 photomultiplier and a 1.5 OD filter was used in front of the forward angle light scatter (FALS) detector. Gates were set log fluorescence intensity around the mononuclear cells using RALS and FALS to FIG. 1. Flow cytometric analysis of human peripheral blood exclude cell clumps or debris (9-11). Cell viabilities, detercells. (A) Overlay of histograms of T, cells stained with PEmined by flow cytometry (1, 12) using propidium iodide conjugated mouse monoclonal Leu-2a (PE-Leu-2a) antibodies (-) staining of dead cells, were consistently >95%. Color comand a histogram of Th cells stained with undiluted PE-conjugated pensation was performed on all samples (13). mouse monoclonal Leu-3a antibodies (PE-Leu-3a) (-------). (B) Overlay of histograms of T, cells stained with PE-conjugated mouse monoclonal Leu-2a (PE-Leu-2a) antibodies (-) and a histogram of RESULTS Th cells stained with PE-Leu-3a antibodies diluted (1:1.2) with Limititions of Current Methods. In Fig. 1A are fluoresunconjugated Leu-3a antibodies (-------). (C) Simultaneous staining of cence histograms obtained by flow cytometric analysis of two T, cells stained with undiluted PE-Leu-2a antibodies and Th cells stained with PE-Leu-3a antibodies diluted (1:1.2) with unconjugated replicate aliquots of lymphocytes from the same individual, Leu-3a antibodies. stained with antibodies to different cell subsets. Both aliquots were stained with antibodies tagged with the same fluorescent chromophore, PE. Since the instrument settings were reduced significantly, Th and T, cells should be distinguishidentical for both samples, it is clear that the fluorescence able from each other and from unstained cells, even though intensities for the helper cells (Leu-3a positive) and suppresthey are of the same color. sor cells (Leu-2a positive) are nearly identical. If these two To test this possibility, Leu-3a antibodies to Th cells were cell populations were mixed together after being stained diluted 1:1.2 with unconjugated antibodies to maintain a independently, it would be impossible to distinguish the two saturating antibody concentration. Flow cytometric analysis subsets since each displays an identical color and similar of cells stained with the mixture of conjugated and unconjufluorescence intensity. gated antibodies revealed that Th cells could be detected at an Establishment of Distinct Fluorescence Prorfles for Cell intermediate fluorescence intensity between the unstained Subsets Using Antibody Dilution. The fluorescence intensity cell fraction and the highly fluorescent T, cells (Fig. 1B). values plotted on the abscissa in Fig. 1A represent logarithFurther experiments established that dilution techniques mic scales. The stained Ts cells in Fig. LA are extremely could be used to achieve simultaneous identification of Th bright (6x) relative to the unstained cells in the preparation. and T, subsets in the same sample even though they are Comparison of the histograms in Fig. 1A indicates that if the stained with the same color chromophore (Fig. 1C). This intensity of the fluorescence profile of Th cells could be overcomes the shortcomings of current methods in which

,.;-'* :.~

Medical Sciences: Horan et A

Proc. Natl. Acad. Sci. USA 83 (1986)

these subsets can be identified only by using different aliquots stained with antibodies to Th or T, cells (Fig. lA) or by the use of two-color fluorescence methods to quantify both subsets in the same sample. Increased Analytical Sensitivity by Means of Two-Parameter Analysis. A potential complication in the identification of Th cells using the antibody dilution method in Fig. 1C could arise in samples containing a high percentage of non-T cells and a low Th fraction. Such a situation would result in overlapping peaks for these cell populations. However, the problem can be circumvented by using two-color analysis. A more definitive separation can be achieved by staining with a third anti-T-cell antibody (Leu4) labeled with fluorescein. The latter was diluted 3:4 with unconjugated Leu4 antibodies. Analysis ofred versus green fluorescence profiles from preparations incubated with the three antibody preparations revealed clear separation of Th and T, cells from non-T cells (Fig. 2). Note that the green fluorescence intensity in this analysis was selected so that the Th and T, subsets occupy the middle range of the abscissa. Detection of an Increased Number of Leukocyte Cell Subsets Using One- and Two-Color Fluorescence. The number of mononuclear cell types in normal peripheral blood is limited. The unstained population of non-T cells identified in Fig. 2 comprises B cells, NK cells, and monocytes. These cell types can be identified and quantified by flow cytometry using a similar strategy ofantibody dilution. For example, B cells can be identified by using undiluted FITC-B1 antibody and two-parameter fluorescence (red versus green; Fig. 3). Since identical laser power and gain settings are used as in Fig. 2, it is clear that the B-cell fraction occupies a different position on the abscissa than the T-cell populations. NK cells were identified by using undiluted PE-conjugated Leu-llc antibodies and can be separated reliably from the Tand B-cell populations (Fig. 3). Finally, monocytes can be identified by staining samples with undiluted PE-Leu-M3 and FITC-Mo2 antibodies to human monocytes to attain maximal intensity positions on 60 +:

501

T, cells

CA c *_

40

.

c) .

+

_ ~~~~Thcells

30

a) 0 20 .-

~

Non-T cells

20

10

log

30

NK cells 50 4-

ce ._ (U

T, cells

40

uz 0

30

Th cells

a)

0 0

0

..

L~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~P

20

B cells

10 ~.

10

30 20 40 50 log green fluorescence intensity

60

FIG. 3. Two-parameter flow cytometric analysis (red versus green) of human peripheral blood cells stained with a panel of different antibodies showing simultaneous detection of T, cells (PE-Leu-2a antibodies), Th cells [PE-Leu-3a antibodies diluted (1:1.2) with unconjugated Leu-3a antibodies], T cells [FITC-Leu-4 antibodies diluted (3:4) with unconjugated Leu-4 antibodies], B cells (FITC-B1 antibodies), NK cells (PE-Leu-11c antibodies), and monocytes (MO) (PE-Leu-M3 and FITC-Mo2 antibodies).

red and green axes (Fig. 3). The light scatter gates are set to include monocytes and lymphocytes. The feasibility of using a mixture of seven antibodies to individual leukocyte cell subsets to achieve simultaneous identification of five cell types using undiluted antibodies (PE-Leu-2a; FITC-B1; PE-Leu-llc; PE-Leu-M3; FITCMo2) and diluted antibodies (PE-Leu-3a; FITC-Leu-4) is obvious in Fig. 3. To determine the percentage of Th cells, the number of cells located in the Th region can be integrated and expressed as a percentage of the total lymphocyte cell number (Th, Ts, NK, and B cells); these percentages are presented in Table 1. This method would be limited, however, if there were significant variations in the position of any one subset from sample to sample. Such variability would preclude comparison of different patients or sequential studies of immune cell populations in individuals. Representative results comparing Table 1. Percentage of cell subsets % of population NK T5 Th Total T B1 cells cells cells cells cells Monocytes Donor 1 (SS) 4.9 26 40 66 21 11 Donor2 (CG) 5.5 28 41 69 21 8 Donor 3 (MS) 4.5 20 62 79 14 11 Donor4 (TCS) 5.8 35 41 74 15 7

0

0

60

8363

40

green fluorescence

50

60

intensity

FIG. 2. Two-parameter flow cytometric analysis (red versus green) ofmononuclear cells stained with PE-Leu-2a (T, cells), diluted PE-Leu-3a (Th cells), and FITC-Leu4 (non-T cells) antibodies diluted (3:4) and gated (light scatter) on lymphocytes. Fluorescence color compensation was achieved electronically.

Normal range

2-22 13-38 32-50 56-79 6-22 NA Integration was accomplished on the four histograms displayed in Fig. 4. Normal ranges were determined on 1500 normal donors and represent the normal ranges for our (SmithKline Biosciences Laboratories) clinical service. Subset data are reported as percentage of total lymphocyte population except for the monocyte population, which is reported as percentage of total mononuclear population. NA, not applicable.

Medical Sciences: Horan et al.

8364

Proc. Nadl. Acad. Sci. USA 83 (1986)

the subsets from four different individuals (two male; two female) in samples stained with the full panel of seven different monoclonal antibodies are shown in Fig. 4. These results suggest that there is little interindividual variation in the position of the five cell subsets for normal subjects. Although the individuals differ in the relative number of each subset (see Table 1), the fluorescence intensities for each specific subset are almost identical.

require more than two fluorochromes with distinct excitation and emission spectra and expensive instruments with two lasers (14). The present method thus offers a significant expansion in the capacity of conventional flow cytometric analysis without the need for new fluorophors or costly modifications in instrument design. The present data illustrate the feasibility of using a cocktail of diluted and undiluted commercially available antibodies to achieve simultaneous detection of T cells, B cells, NK cells, and monocytes in a single aliquot. Although these particular cell subsets represent the major leukocyte classes examined in routine flow cytometric analyses for clinical investigation (15-20), the same approach could be used to quantify lymphocytes, monocytes, granulocytes, eosinophils, and basophils if appropriate fluorescent-labeled antibodies to each subset were available. A further attractive feature is that the method need not be limited to measurements with fluorophore-conjugated antibodies but is equally applicable to lectins or other ligands that can be coupled to fluorophores and bind with sufficient specificity to achieve differential fluorescence intensities between stained and unstained cells. The ability to quantify immune cell subsets in the same blood sample by using antibodies labeled with a single fluorochrome is particularly useful in clinical laboratory analysis. Immunotaxonomic measurements of the relative quantities of immune phenotypes are being used increasingly as an important index of specific immune subsets in patients and to detect and monitor infections and immune dysfunction syndromes such as acquired immunodeficiency syndrome (21). The method described in this study eliminates the need to use more complex two-color staining methods to detect Th cells and Ts cells in the same sample or the assay of individual

DISCUSSION This paper describes a simple and inexpensive method to expand the analytical capacity of flow cytometric techniques for characterizing leukocyte cell subsets in human peripheral blood. By diluting conventional commercial preparations of fluorophore-conjugated antibodies with unconjugated antibodies and staining cells, the fluorescence intensity profiles of different cell subsets can be adjusted to display distinguishable fluorescence profiles and permit simultaneous enumeration of multiple cell types. One might be tempted to reduce the fluorescence intensity of the cells by merely lowering the antibody concentration. This approach is not recommended because it is important to stain cells in saturation amounts of reagent, avoiding fluctuations in stain intensity due to cell number or antigen affinity differences. For this reason, the fluorochrome-conjugated reagent is diluted with unlabeled antibody of the same specificity. In addition to allowing simultaneous detection of two cell subsets using antibodies bearing the same fluorochrome, this method permits identification of up to five cell subsets by using two-color fluorescence and a single laser beam, an accomplishment hitherto limited to complex methods that 60 _ Donor I

60

50

50

*

40

w

40

;X~vA

30

i

x0

1 ~~4.IW~~~4A; L

30

K7

_

20 *~~ ~~~~~~~~~~~~~~~~~~~~ .4 -

*._ r.

Donor 2

1010

20

10

r. 0 I4) 0

4)

- Donor 3

D~onor 4

60 50

40

,.

*

eC

.

.,8

40

-

30

30

20

20

10

10 10

20

3()

* 3, ; ^S ' .

40

.

*

50 60 log green fluorescence

3

'

' '.7.

'*'