A new membrane permeabilization method for the detection of intracellular antigens by flow cytometry

Journal of Immunological Methods, 124 (1989) 103-109

103

Elsevier JIM05341

A new membrane permeabilization method for the detection of intracellular antigens by flow cytometry G. H a l l d r n 1, U. A n d e r s s o n 2, j. H e d 1 a n d S.G.O. J o h a n s s o n 1 1Department of Clinical Immunology, Karolinska Hospital, Stockholm, Sweden, and 2 Department of Pediatrics, St. G~ran's Children's Hospital, Stockholm, Sweden

(Received 3 October 1988, revisedreceived12 June 1989, accepted22 June 1989)

This article desrcibes a new cell membrane permeabilization method for the detection of intracellular antigens by immunofluorescence staining and flow cytometry. The number of cells remained unaltered and no cell aggregation or loss of intracellular antigenicity was observed after this permeabilization treatment. A mixed leukocyte population from human peripheral blood was used in this study and the leukocytes were fixed and permeabilized, which permitted monoclonal anti-vimentin antibodies to reach intracellular antigens. The stabilization of cell membranes and preservation of intracellular antigenicity was achieved with paraformaldehyde fixation. This pretreatment prevents cell destruction and subsequent treatment with the detergent n-octyl-fl-D-glucopyranoside results in permeabilization of the cell membrane. The procedure does not alter the expression of cell surface antigens, which is of importance if phenotypic characterization of intracellularly stained cells is to be undertaken. Furthermore, this simple, rapid and reproducible technique makes it possible to detect and distinguish between different human peripheral blood leukocytes without prior purification steps. The leukocyte subpopulations remain well-separated and easily detectable by flow cytometry. Key words: Flow cytometry;Membranepermeabilization; Leukocyte;Intracellularimmunofluorescence;Paraformaldehyde;n-octyl-

fl-D-glucopyranoside

Introduction

Immunoassays using fluorochrome-conjugated monoclonal antibodies to characterize cells have been facilitated by the introduction of flow cy-

Correspondence to: G. Halldrn, Department of Clinical Immunology, Karolinska Hospital, S-104 01 Stockholm, Sweden. Abbreviations: OG, n-octyl-fl-D-glucopyranoside; CFDA, carboxyfluoresceindiacetate; CMC, critical micelleconcentration; EDTA, ethylenediaminetetraaceticacid; FITC, fluorescein isothiocyanate;MFI, mean fluorescenceintensity; PBS, phosphate-buffered saline; PE, phycoerythrin;PFA, paraformaldehyde.

tometry. However, flow cytometry has been limited to the analysis of cell surface antigens, since the methods of cell membrane permeabilization which should permit the detection of intracellular structures and products have usually been unsatisfactory. Treatment of cells with various fixatives and detergents has led to morphological damage, cell aggregation or loss of intracellular antigenicity (Andersson et al. 1988). Schroff et al. (1984) used lysolecithin to permeabilize cell membranes of separated human mononuclear cells and in this study we have attempted to overcome the problems mentioned by using the mild non-ionic detergent n-octyl-fl-oglucopyranoside (OG) (Womack et al., 1983) on

0022-1759/89/$03.50 © 1989 Elsevier SciencePublishers B.V. (BiomedicalDivision)

104

paraformaldehyde (PFA) pre-fixed unseparated human leukocytes. PFA has earlier been shown to stabilize cell membranes and preserve the cell morphology (Williams and Chase, 1976). This new method retained well-separated leukocyte subpopulations and permitted us to detect the presence of intermediate filaments in human peripheral blood leukocytes with a monoclonal antibody against vimentin. Cell membrane permeability was also demonstrated by analysing the leakage of carboxyfluorescein diacetate (CFDA) from preloaded cells.

Materials and methods

Monoclonal antibodies Monoclonal mouse anti-vimentin (IgG1) (Dakopatts, Glostrup, Denmark) antibodies were used as a probe to test for accessibility of intracellular antigens and were tested in various concentrations (0.18-14.0/~g/ml). The same concentration range of mouse IgG1 (Coulter Electronics, England) was used to detect non-specific background fluorescence. Fluorescein isothiocyanate (FITC)-conjugated rabbit F(ab')2 anti-mouse IgG (Dakopatts, Glostrup, Denmark) was used as a secondary antibody (2.5/xg/ml). Cell surface antigens were detected using the following FITC- or phycoerythrin(PE)-conjugated monoclonal antibodies against: CD3 (5 /~g/ml), CD8 (2.5 Izg/ml), CD4 (5/~g/ml), CD36 (OKM5) (10 /~g/ml) (Ortho Diagnostic Systems, Raritan, N J, U.S.A.), CD20 (50/~g/ml) (Coulter Electronics, England) CD16 (12.5 /zg/ml), HLA-DR (2.5 /~g/ml) and CR3 (5 /~g/ml) (Becton Dickinson, Mountain View, CA, U.S.A.). FITC- and PEconjugated mouse IgG1 and IgG2 (Becton Dickinson, Mountain View, CA, U.S.A.) used at comparable concentrations served as background controls.

Detection of carboxyfluorescein diacetate (CFDA) leakage The efficiency of cell membrane permeabilization was also estimated by detecting CFDA leakage from cells preloaded with CFDA and then permeabilized. Leukocyte pellets, prepared as described below, were resuspended in 100/~1 CFDA

solution (10 /zg/ml) in Tris-buffered Hanks' salt solution (Hansson et al., 1987), incubated for 15 rain at room temperature and washed in ice-cold PBS-EDTA. The cells were then treated with PFA or OG (6.0 mg/ml) alone or in combination as described in the method of fixation and permeabilization.

Cell preparation Leukocytes were prepared from EDTA blood collected from healthy adult blood donors. Erythrocytes were specifically lysed by adding 2 ml NH4C1-EDTA ('lysing reagent') (Ortho Diagnostic Systems, Raritan, NJ, U.S.A.) to 100 /~1 blood. The suspensions were incubated for 2-5 min at 15 ° C and then centrifuged at 300 × g for 5 min at 4 ° C. The leukocyte pellets were finally washed in 2 ml ice-cold 0.15 M PBS, pH 7.4, supplemented with 10 mM EDTA and 0.02% NaN 3 (PBS-EDTA) (Hed et al., 1987) and then kept on ice until used.

Membrane fixation and permeabilization Cell membrane stabilization (fixation) was performed by paraformaldehyde (PFA) treatment according to Williams and Chase (1976). Leukocyte pellets were suspended in 100/~1 4% PFA in 0.1 M PBS, pH 7.2, incubated for 5-10 min at room temperature and washed once in PBS-EDTA by centrifugation at 400 x g for 5 min at 4 ° C. Cell membrane permeabilization was achieved by exposure to n-octyl-fl-D-glucopyranoside (OG) (Sigma, St. Louis, MO, U.S.A.), a mild non-ionic detergent with limited protein denaturation effects and a critical micelle concentration (CMC) at 0.74%. The fixed leukocytes were exposed to OG (0.37-8.0 mg/ml) dissolved in 0.15 M PBS for 5 min at room temperature then washed twice in ice-cold PBS-EDTA by centrifugation at 400 x g for 5 min at 4 ° C. The fixed and permeabilized leukocyte pellets were kept on ice until used. Lysolecithin treatment was preformed according to Schroff et al. (1984) by suspending untreated leukocyte pellets in 200 /~1 cold lysolecitin (50 /1g/m1) (lysophosphatidylcholine; Sigma, St. Louis, MO, U.S.A.). The resuspended cells were incubated at 4 ° C for 5 min and then washed twice.

105

Immunofluorescent staining of intracellular antigen

Flow cytometry

To detect cytoplasmic antigen cells were fixed (PFA) and permeabilized (OG) and then resuspended in 100/~1 anti-vimentin solution (0.18-14.0 /~g/ml) in PBS-EDTA and incubated for 30 min in an ice bath. Leukocytes resuspended in mouse IgG1, treated in the same way, served as controls. After two washings in ice-cold PBS-EDTA the leukocytes were incubated in 100 /al FITC-conjugated rabbit F(ab')2 anti-mouse IgG solution (2.5 # g / m l ) for 30 min in an ice-bath. The cells were washed twice and resuspended in 1 ml PBSE D T A before being tested by flow cytometry. To detect the effect of the fixation and permeabilization procedures on cell surface markers 100 ttl EDTA-blood were hemolysed and washed as described. Untreated leukocytes were then incubated with different FITC- or PE-conjugated monoclonal antibodies for 30 rain at 4 ° C at the same concentrations as described above. Directly conjugated mouse IgG1 and IgG2 were used as controls. Each monoclonal antibody was tested in duplicate. After two washings in PBS-EDTA, the leukocytes in half of the tubes were fixed and permeabilized while the rest was incubated in PBS-EDTA alone and served as controls. Finally the cell suspensions were diluted in 1 ml PBSEDTA and tested by flow cytometry.

The final preparation of leukocyte pellets was resuspended in I ml cold PBS-EDTA and analysed in an Ortho Spectrum III flow cytometer (Ortho Diagnostic Systems, Westwood, N J, U.S.A.). Based on light-scattering properties each cell was represented by a point in a rectangular co-ordinate system where lymphocytes, granulocytes and monocytes were represented by well-separated clusters. A discriminating frame set around the different cell clusters permitted analysis of each cell population one by one. In addition to detecting the light scatter properties from the cells, fluorescent signals from the fluorochromes could also be detected. The instrument gave the absolute number of cells, the percentage and actual number of fluorescent cells and the mean fluorescent intensity (MFI) of the cell population within the analysed field, which is a measure of receptor density.

Results

The effect of different permeabilization treatments on the cytogram distribution of human peripheral leukocytes The light scatter cytogram of untreated leukocytes was compared to cytograms of leukocytes

Fu_S¢=

;~0~~

Nonocyt.. Granulocyt...

". • " " :"

1~0

~• ~..-=~=:."'"'*' ~ : ~ : : ~ : ' "~ . : : ::...'~m..:;:. : :. ";"

"'i'"

¢¢

i :-:::

:': 5O

A

B

C

Fig. 1. Light-scatteringproperties of permeabilized human blood leukocytes. Cytograms detecting forward and right-angle light scatter where each cell is represented by a point in a rectangular co-ordinate system. The resolution of lymphocyte,monocyte and granulocyte dusters are compared in (A) untreated leukocytes,(B) PFA+OG permeabilized, and (C) lysolccithin permeabilized. Gating frames for lymphocytesare indicated. Notice the elimination of debris in PFA + OG treated samples.

106

permeabilized with lysolecithin or O G (0.6%) on PFA pre-fixed cells (Fig. 1). In repeated experiments ( n - - 5 ) the cytogram was heavily altered when the leukocytes were permeabilized with lysolecithin. It was not possible to gate the monocyte cell population and a conspicuous decrease in the number of cells was obtained within the lymphocyte cell population (56.1-76.6% loss of cells) (n = 5) whereas the number of granulocytes remained unaltered. Only a slight change in fight scatter properties could be seen when the leukocytes were treated with PFA and OG, but the different cell populations were still well-separated and were readily discriminated by setting a frame around each leukocyte field. No significant decrease in cell numbers was obtained. There were no changes in the percentage of CD3-, CD20 (lymphocyte field)-, CD16 (granulocyte field)- or CD36 (OKM5) (monocyte field)-positive cells within the different cell clusters after the permeabilization with lysolecithin or PFA + OG. The combination of PFA fixation and O G treatment was a better method of cell membrane permeabilization for flow cytometry and was selected for use in further studies.

Membrane permeability

Effect of OG concentration

In order to select the optimal concentration of anti-vimentin to assay the efficiency of cell mem-

The effect of O G concentrations on antivimentin binding by human peripheral blood lymphocytes treated in different ways is shown in Fig. 2. No anti-vimentin binding could be detected in cells treated with O G alone and the number of cells decreased. This result was probably due to cell lysis and loss of cells caused by micelle formation and release of antigen (vimentin) from permeabilized cells or antigen destruction of non-fixed vimentin. Fixation in PFA and exposure to 6.0 m g / m l O G led to a high percentage of permeabilized cells (87.2 5: 5.9) as judged by anti-vimentin binding. Higher O G concentrations raised this percentage to a maximum of 9 8 . 4 _ 0.4% at an O G concentration of 8.0 mg/ml. This result highlights the importance of membrane stabilization and antigen fixation before adding the detergent. At concentrations lower than 3.0 m g / m l no cell anti•¢imentin binding could be detected. The cell recoveries after both fixation and permeabilization procedures were 100%.

100 Q e~

~' 50 °~

I

0.37 0.75 1.5

3.0

6.0

12.0

/~-octyl-glucoside (mg/ml) Fig. 2. Effect of n-octyl-fl-D-glucopyranosideconcentration. Untreated (<3) and paraforrnaldehyde fixed (e) leukocytes incubated with various concentrations of OG. Permeabilized lymphocytes were detected by flow cytometry after staining with antivimentin (2.8 #g/ml) and FITC-conjugated rabbit F(ab')2 anti-mouse IgG (2.5 /~g/ml). Mean values+SD (n = 3).

100 ~

2[.

80 60

2o

~

.._.----4

0

I

0.18 0.35 0.7

1.4

2.8

14.0

Anti-vimentin conc. (pg / ml ) Fig. 3. Cell m e m b r a n e permeability. Lymphocytes treated with O G (6.0 rng/rnl) (o), P F A (4%) (A) or P F A followed by O G (e) and stained with various concentrations of anti-vimentin monoclonal antibody for flow cytometry. Mean v a l u e s + SD

(n = 3).

107 TABLE I L E A K A G E OF C A R B O X Y F L U O R E S C E I N D I A C E T A T E ( CFDA) F R O M P E R M E A B I L I Z E D L Y M P H O C Y T E S Leukocyte treatment

% Permeabilized lymphocytes a.b

% C F D A stained lymphocytes a

Untreated PFA flOG c P F A + flOG

1.0 + 0.4 8.5 + 2.3 0.6+0.8 97.45:0.7

97.2 + 2.2 81.7 + 1.8 2.0_+0.5 0.3-+0.3

a Results are expressed as mean + SD (n = 3). b Permeabilization was assessed by the percentage of antivimentin positive cells. c n - o c t y l - f l - D - g l u c o p y r a n o s i d e ( f l O G ) concentration was 0.6%.

brane permeabilization, various concentrations (0.18-14.0/~g/ml) of the antibody were used with human peripheral blood leukocytes treated in different ways (Fig. 3). Lymphocytes treated optimally both with PFA and OG as described above bound anti-vimentin in a dose-related manner up to a plateau at 2.8 /~g/ml of the monoclonal antibody. Neither untreated lymphocytes, nor lymphocytes treated with OG alone, bound anti-vimentin. On PFAfixed lymphocytes only a slight increase of 15.3 _+ 1.8% in lymphocyte-bound anti-vimentin could be seen, at the highest antibody concentration (18.0 /~g/ml) tested in this study. The binding of anti-vimentin to PFA- and OG-treated granulocytes was also tested. After treatment, 98.0 _+ 2.0% (n = 3) of the granulocytes were shown to bind anti-vimentin without any

decrease in cell number as counted by flow cytometry. Mouse IgG1 binding to fixed and permeabilized lymphocytes reached 2.3 + 1.5% positive cells when used at the high concentration of 14.0 #g/ml.

CFDA leakage High fluorescence intensity, reflecting the CFDA hydrolysis in the cell cytoplasm, was observed in the majority (97.2 + 2.2%) of the untreated lymphocytes incubated in the presence of CFDA (Table I). Cell membrane permeabilization procedure with PFA plus OG resulted in 97.4 + 0.7% anti-vimentin binding cells and only 0.3% of the lymphocytes retained CFDA staining indicating efficient permeabilization. PFA treatment alone gave 81.7 + 1.8% preserved CFDAstained lymphocytes and limited anti-vimentin binding (8.5 + 2.3%), indicating a low percentage of permeable cells. Treatment with OG alone gave almost complete CFDA leakage, only 2.0 + 0.5% lymphocytes remained CFDA-stained, the cell recovery was low and the lymphocytes failed to bind anti-vimentin. These results were probably due to the OG concentration (0.6%) which was close to CMC and caused cell lysis. Granulocytes were tested in the same way for CFDA leakage with results which were similar to those obtained on lymphocytes. Effect of permeabilization treatment on the detection of cell surface antigens The antibodies to cell surface antigens were added before the fixation and permeabilization

TABLE II T H E E F F E C T OF P E R M E A B I L I Z A T I O N T R E A T M E N T ON L E U K O C Y T E S U R F A C E A N T I G E N S Leukocyte type

Surface antigen

Lymphocytes Lymphocytes Lymphocytes Monocytes Granulocytes Granulocytes

CD3 CD4 CD8 HLA-DR CD16 CR3

Untreated a

Treated a.b

% Positive

MFI c

% Positive

MF I c

78.1 + 51.7 + 34.9 + 89.4 + 96.1 + 100 +

139 + 133 + 139 + 121 + 227 + 72 +

76.2 + 51.9 + 35.1 + 84.0 + 95.0 + 100 +

141 _ 30 127 _ 27 147 _ 39 121 + 22 196 + 60 81+19

5.5 6.3 6.3 5.4 0.3 0

a Results are expressed as m e a n + SD (n = 3). b Treated with PFA + f l O G after binding of monoclonal antibodies. c M F I = mean fluorescence intensity.

34 27 36 29 66 8

5.1 4.8 5.8 1.7 1.8 0

108 steps in order to achieve optimal binding to (unaltered) cell surface antigens and to protect the antigen-antibody reaction from the denaturing effect of the treatment with PFA and OG. This sequence also avoided antibody binding to intracellular structures. As shown in Table II PFA and OG treatment had no effect either on the percentage of positive leukocytes or on the MFI of positive cell populations.

Discussion There is a need in flow cytometry to develop rapid and simple methods to permeabilize cell membranes in order to permit detection of intracellular antigens. However, in the application of any such technique it is important to avoid cell destruction, cell aggregation and antigen denaturation. In this study we used a method in which peripheral blood leukocytes were fixed in suspension by PFA. PFA is known to stabilize cell membranes and preserve cell morphology (Williams et al., 1976). Formalin is primarily a 'complex-binding' fixative without denaturative effects in immunohistochemistry (Hed and Enestr~Sm, 1981). The most critical functions of PFA in this study were to stabilize the cell membranes, preserve intracellular antigens and preserve forward (cell size) and right (complexity) scatter properties in the flow cytofluorometric cytogram. Subsequent exposure to increasing concentration of OG made the cell membrane permeable to anti-vimentin monoclonal antibody and the CFDA leakage was almost complete. OG is a mild non-ionic detergent, with high CMC and can be rapidly removed by washing. It is known to be effective in the solubilization of lipid bilayers without altering the structure of proteins and without affecting enzyme activity, if it is used in concentrations at or preferably slightly below its CMC value (Womack et al., 1983). OG concentrations from 6.0 m g / m l to 8.0 m g / m l gave optimal permeabilization of the cell membranes of PFA-fixed lymphocytes in this study. PFA fixation alone had only a small effect on membrane permeabilization. OG could not be used alone since the cells were destroyed, probably as a

result of instability in the membranes and release of soluble intracellular antigens or destruction of the antigenicity of vimentin. At the optimal conditions for cell membrane permeabilization the cell recovery was 100% and the flow cytometric cytogram showed only slight change in light-scattering properties which suggest that cell morphology was retained and that no cell aggregation had occurred. Furthermore, the treatment did not affect the detection of cell surface antigens (DR, CR3, CD16, CD3, CD4, CD8) if the cells were labelled with fluorochrome-conjugated antibodies before PFA fixation. Measured by flow cytometry the forward and right-angle scatter properties were only slightly changed in the permeabilized mixed leukocyte preparations used in these studies, but well separated leukocyte fields were readily identified. Since the fixation and permeabilization procedures described in this paper disrupt the integrity of the cell membrane without affecting the stained cell surface antigens the phenotypic characterization of intracellularly stained cells should be easily determined by double staining. This suggests possible applications. For example, we have already successfully used the method in the enumeration of IFN-y producing cells with excellent agreement between fluorescence microscopy and flow cytometric methods (Andersson et al., 1988) and also showed stability of intracellularly-stained cells in suspension up to 4 weeks at 4 ° C.

Acknowledgements We would like to thank Ms. Maria Grylling for excellent technical assistance. This work was supported by the Swedish Medical Research Council (Grant 16x-105), the Swedish Work Environment Fund (no. 84/1302) and the Swedish Lung and Heart Foundation.

References Andersson, U., Halld6n, G., Persson, U., Hed, J., M611er,G. and DeLey, M. (1988) Enumeration of IFN-y-producing cells by flow cytometry:comparison with fluorescencemicroscopy. J. Immunol. Methods 112, 139.

109 Hansson, Y., Jacobson, E., tgrtlund, J., Paulie, S. and Perlmann, P. (1987) A rapid method for detection of cellular proliferation using carboxyfluorescein. Assay of growth factors (IL-2, IL-1) and growth inhibiting antibodies. J. Immunol. Methods 100, 261. Hed, J. and EnestriSm, S. (1981) Detection of immune deposits in glomeruli: The masking effect on antigenicity of formalin in the presence of proteins. J. Immunol. Methods 41, 57. Hed, J., Halld6n, G., Johansson, S.G.O. and Larsson, P. (1987) The use of fluorescence quenching in flow cytofluorometry to measure the attachment and ingestion phases in phago-

cytosis in peripheral blood without prior cell separation. J. Immunol. Methods 101, 119. Schroff, R.W., Bucana, C.D., Klein, R.A., Farrell, M.M. and Morgan, Jr., A.C. (1984) Detection of intracytoplasmic antigens by flow cytometry. J. Immunol. Methods 70, 167. Williams, C.A. and Chase, M.W. (Eds.), (1976) Methods in immunology and immunochemistry, Vol. V. Academic Press, New York. Womack, M.D., Kendall, D.A. and MacDonald, R.C. (1983) Detergent effects on enzyme activity and solubilization of lipid bilayer membranes. Biochim. Biophys. Acta 733, 210.