Flow cytometry sorting of unlabelled epidermal langerhans cells using forward and orthogonal light scatter properties

Flow cytometry sorting of unlabelled epidermal langerhans cells using forward and orthogonal light scatter properties

Journal oflmmunological Methods, 79 (1985) 79-88 79 Elsevier JIM 03475 Flow Cytometry Sorting of Unlabelled Epidermal Langerhans Cells Using Forwa...

2MB Sizes 0 Downloads 103 Views

Journal oflmmunological Methods,

79 (1985) 79-88

79

Elsevier JIM 03475

Flow Cytometry Sorting of Unlabelled Epidermal Langerhans Cells Using Forward and Orthogonal Light Scatter Properties Genevi6ve Cordier 1, Colette Dezutter-Dambuyant 2, Ray Lefebvre 1 and Daniel Schmitt 2 l Centre de Cytofluorombtrie, Unwersitb Lyon I and 1NSERM U.80, and 2 1 N S E R M U.209, Hbpital E. Herriot, 69374 Lyon Cedex 08, France

(Received 17 September 1984, accepted 15 January 1985)

A variety of techniques based on the presence of specific markers has been proposed to enrich Langerhans cells from epidermal cell suspensions. Computer analysis of multiparameter flow cytometry records involving forward angle and orthogonal scattered light and immunofluorescence of epidermal cells allowed us to determine the scattering properties of Langerhans cells. Unlabelled cells sorted according to these properties were shown to be Langerhans cells by electron microscopy and/or subsequent labelling by anti-HLA-DR monoclonal antibody. The relevance of this method is discussed to sorting viable Langerhans cells which may be used in functional studies and for establishing long-term culture. Key words: Langerhans cells - light scattering - flow cytometry analysis - sorting - electron microscopy

Introduction Flow cytometry analysis offers a convenient solution to the problem of characterizing subpopulations of cells in a mixed population. As reviewed by Shapiro (1983) parameters classified as intrinsic or extrinsic and related to structural or functional properties of cells m a y be used for analysis and sorting. N e w information might appear also from the addition of c o m p u t e r capabilities. With the increasing scope of available cytometers it is now possible to analyse simultaneously several properties of cells. Thus for the determination of lymphocyte cell subpopulations in whole blood the use of light scattering properties for recognition of lymphocytes, monocytes and p o l y m o r p h s and of immunofluorescence for detection of specific markers has been proposed ( H o f f m a n et al., 1980). M a m m a l i a n epidermal cells have been distinguished on the basis of ultrastructural characteristics (Mahrle and Orfanos, 1974; Shelley and Juhlin, 1978), histochemical staining pattern (Wolff and Winkelmann, 1967) and more recently the presence of cell surface antigens. A m o n g them, Langerhans cells are characterised by 0022-1759/85/$03.30 © 1985 Elsevier Science Publishers B.V. (Biomedical Division)

80 the simultaneous expression of T6 (Fithian et al., 1981), HLA-DR antigens (Klareskog et al., 1977; Rowden et al., 1977) and receptors for the Fc portion of IgG and for C3b (Stingl et al., 1977). The proportion of Langerhans cells is usually less than 5% of epidermal cells and several methods have been used to enrich this population for functional studies. These rely on recognition and sorting of ceils by indirect immunofluorescence (Morhenn et al., 1982), immuno-adsorption to coated petri dishes through either anti-Ia antibodies (Scheynius et al., 1982) or Staphylococcus aureus-treated cells (Schuler et al., 1983) or Fc-receptor rosetting (Stingl et al., 1978). It has been assumed that some of these methods allow recovery of cells with functional properties (Morhenn et al., 1982), but the need remains for a procedure giving a good yield of highly purified viable Langerhans cells without previous labelling. We describe here flow cytometry sorting of Langerhans cells based on their light scattering properties. These properties were assessed by computer analysis of results on cells identified by specific labelling. Unlabelled cells sorted according to light scattering properties were shown to be Langerhans cells by electron microscopy a n d / o r subsequent labelling with anti-HLA-DR monoclonal antibody.

Materials and Methods

Preparation of enriched Langerhans cell suspensions Single cell suspensions were prepared from freshly removed normal human skin from patients undergoing plastic surgery of the breast and the abdomen as previously described in detail (Dezutter-Dambuyant et al., 1984). Initial enrichment of Langerhans cells was by Ficoll-Hypaque sedimentation (density 1.77): the cells at the interface were shown to contain a 2-5-fold increase in the number of HLA-DR expressing cells. They were resuspended in phosphate-buffered saline (PBS) with 30 mM EDTA and 1% bovine serum albumin (BSA).

Indirect irnmunofluorescence In preliminary studies cells bearing HLA-DR antigens or T6 antigen were identified in epidermal cell suspensions. Cells were incubated for 60 rain at 20°C in the presence of BL2 antibody (a monomorphic anti-HLA-DR monoclonal antibody produced in our laboratory (Yonish-Rouach et al., 1983), or of OKT6 commercially available monoclonal antibody. They were then washed 3 times and further incubated for 30 rain at 4°C with conjugated goat anti-mouse IgG (GAM-FITC), (TAGO Inc., Burkingame, CA) at the optimal dilution. The cells were again washed twice prior to flow cytometry analysis. Cells treated with the second antibody alone were used for analysis of background immunofluorescence. Cells were processed either for flow cytometry or microscopy.

Flow cytometry analysis Cell suspensions unlabelled or labelled by indirect immunofluorescence were run in a Cytofluorograf 5OH (Ortho Instruments, Westwood, MA) equipped with a 5 W

81 Argon Ion laser and interfaced with an interactive computer developed in this laboratory (Cordier et al., 1982) which allows simultaneous analysis of up to 4 parameters. The 488 # m laser line was used at 400 mW for excitation. Three parameters were recorded for each cell crossing the laser beam: the forward angle light scatter (FAS), which is mainly related to cell diameter, the right angle light scatter (RAS), which gives information on the internal structural properties of the cell (Brunstig and Mullaney, 1974), and fluorescence (GREEN) related to the presence of membrane antigens recognised by specific antibodies. They were stored in a list mode manner to allow further analysis of a single parameter by histogram distribution and the gating of area of interest by computer facilities. Because enriched Langerhans cell suspensions were obtained in the same way as lymphocytes we choose to analyse these suspensions with the same settings for photomultipliers measuring RAS and FAS as for lymphocytes. For fluorescence either a linear or a log amplification was used to obtain the best signal. Determinations were based on 10 4 cells per sample.

Flow cytometry sorting For cell sorting, suspensions were adjusted to 106/ml and separation carried out at a rate of approximately 1000 cells/s. The fluid was filtered PBS. Gates were chosen as described in Results.

Fluorescence microscopy Labelled cells were examined under a Zeiss fluorescence microscope. The percentage of labelled cells was evaluated from at least 103 cells.

Electron microscopy Unlabelled epidermal cells recovered from sorting were fixed with 2% paraformaldehyde in periodate-lysine buffer. After washing, the cells were incubated with anti-HLA-DR BL2 monoclonal antibody. The murine antibodies bound to the cell surface were identified with goat anti-mouse immunoglobulin labelled with gold granules of 20 nm size (GAM 20, Janssen, Pharmaceutica, Beerse) according to the indirect immunogold technique described by Dezutter-Dambuyant et al. (1984). Labelled cells were then processed according to our standard embedding procedures (Dezutter-Dambuyant et al., 1984). Ultrathin sections were examined with a Philips EM 300 electron microscope. Langerhans cells were counted and identified by established morphologic criteria (Wolff, 1972).

Results

Determination of light scattering properties of Langerhans cells Epidermal cells labelled either with BL2 or OKT6 antibody as described were run in the Cytofluorograf and measurements of RAS, FAS and G R E E N fluorescence were stored in list mode fashion. From this list mode it was possible to display 3

82 w,

•o ~

~



18

Fig. 1. Computer-drawn contour plot from cytofluorometric analysis of epidermal cells. The distribution of cells is shown according to RAS (right angle scattered light, r-axis) and I=AS (forward angle scattered light, )'-axis).

bivariate representations: RAS versus FAS, RAS versus G R E E N fluorescence and FAS versus G R E E N fluorescence. The distribution of epidermal cells according to RAS and FAS is shown in Fig. 1. The cell distribution was heterogeneous; numerous 2A

2B

Z W W . . . .

s~ . . . .

1~

''

'1~6

''

~

""

~

RA3

Fig. 2. Computer drawn dot plot from cytofluorometric analysis of epidermal cells treated with OKT6 antibodies. Distribution of cells is presented according to RAS (x-axis) and GREEN fluorescence (y-axis). A: gating for fluorescent cells. B: the correlate computer-drawn contour plot shows the R A S / F A S distribution of labelled cells.

83 cells had low RAS and FAS properties, other cells had high RAS and low FAS, and a cluster of cells could be defined as low RAS but high FAS. This distribution was not modified by treatment with specific monoclonal antibody as shown by similar distribution in labelled suspensions and controls. Looking at the R A S / G R E E N fluorescence distribution from cells previously labelled with OKT6 monoclonal antibody, we observed a distribution of fluorescent cells restricted to low RAS properties (Fig. 2A). We gated on these cells as indicated and redisplayed the correlated R A S / F A S distribution (Fig. 2B). Labelled cells were characterized by low RAS properties and rather high FAS. The same result was obtained on 8 different suspensions from skins of different individuals. The same analysis was done on cell suspensions from the same donors treated with BL2 monoclonal antibody and again labelled cells were found to be restricted to the region of low RAS and high FAS (Fig, 3A and B). A further analysis was performed to confirm the distribution of cells. The gated R A S / F A S distribution was evaluated from OKT6-1abelled cells. For the same run total cell count, fluorescent cell count and fluorescent cell count in the gated area were determined. Then a second run was performed with BL2-1abelled cells and the number of positive cells was evaluated before and after the same gating as for OKT6. Results from 4 different experiments are shown in Table I. About 70% of OKT6-1abelled cells were present in the gated area. With BL2-1abelled cells, despite a slight difference in the percentage of positive cells the same recovery was obtained indicating that gating from OKT6-positive cells allow the selection of BL2-positive cells. Since Langerhans cells are distinguished among epidermal cells by the simultaneous presence of OKT6 and H L A - D R structures these analyses allowed us to define Langerhans cells as low RAS-high FAS cells•

3B

3A

E I®

RA8

Fig. 3. Computer-drawn dot plot from cytofluorometricanalysis of epidermal cells treated with BL2. Conditions for analysis and displayas in Fig. 2. A: gating for fluorescentcells. B: correlate computer-drawn contour plot of RAS/FAS distribution of labelled cells.

84 TABLE I D E T E R M I N A T I O N OF LIGHT SCATTERING PROPERTIES OF OKT6-LABELLED AND BL2LABELLED CELLS A N D EVALUATION OF FLUORESCENT CELL NUMBER IN THE SELECTED R A S / F A S D E F I N E D W I N D O W Cells pre-treated with OKT6 Cell count " Total cells Fluorescent cells before gating Fluorescent cells after gating Percentage of recovew

BL2 %h

13 679 ± 469

Cell count

9i

13 287 ± 450

2 754 ± 383

20,1

2 415 + 603

18.2

2 001 ± 268

14,6

1 712 + 456

12.9

73

71

~' n = 4 experiments, results expressed as mean _+SEM. b fluorescent cells x 100. total cells

Fig. 4. Detection of HLA-DR antigens on sorted epidermal cells by the immunogold technique. The suspension is composed primarily of Langerhans cells (LC) with some contaminating keratinocytes (KI identified by the presence of cytoplasmic tonofilaments (magnification: × 9800).

85

Sorting of unlabelled Langerhans cells according to their light scattering properties and ultrastructural appearances These findings prompted us to define a strategy for sorting unlabelled cells. A small aliquot (10 6 cells) of a suspension was treated with OKT6 antibody and analysed as described above to establish the R A S / F A S distribution of the positive cells. 10 3 cells from this labelled aliquot were sorted in the selected R A S / F A S region to determine the presence of fluorescent cells. The cells from the unlabelled suspension were then subjected to sorting with the R A S / F A S defined window. Cells from this sorting were further labelled with BE2 antibody and processed for electron microscopy. Indirect immunogold labelling with BL2 antibody indicated the presence of HLA-DR antigens on cell membrane of about 75% of sorted epidermal cells (Fig. 4). The majority of BL2-reactive cells were dendritic Langerhans cells identified by cytoplasmic Birbeck granules (Fig. 5). Besides Langerhans cells only small keratinocytes (identified by the presence of desmosomes and tonofilaments) were seen and these were negative with the anti-HLA-DR antibody (Fig. 4). The majority of typical Langerhans cells with Birbeck granules showed weak labelling but some typical Langerhans cells and Birbeck granule-lacking cells showed a strong labelling

Fig. 5. HLA-DR-positiveLangerhanscell with Birbeck granules (g) (magnification: ×12,000).

86

Fig. 6. Two HLA-DR-positive cells. The left-hand cell shows weak labelling with a low number particles along the membrane. The right-hand cell shows strong labelling with a large number particles (magnification: x 24,000).

(Fig. 6). This procedure allowed the sorting 20% of the whole population, in 2 h.

of lo6 cells. representing

of gold of gold

an average

of

Discussion Their Langerhans cells comprise about 2-5s of the epidermal cell population. procedures from cell characterisation and functional study require enrichment suspensions obtained by trypsinisation. Several approaches have been used at this aim including rosette formation (Morhenn et al., 1982; Bjercke et al., 1984). monolayer panning (Scheynius et al., 1982) and flow cytometry sorting (Morhenn et al., 1982; Scheynius et al., 1982). The latter was shown to give better yields and purity. It has been assumed that sorted cells retain their stimulation potency in mixed skin cell-lymphocyte reaction after mild trypsinisation (Morhenn et al.. 1982). but a drawback of the isolation technique and of these different enrichment procedures might be that they result in functionally defective Langerhans cells, since long-term culture of these cells has not been carried out to assess their ability to regenerate cell membrane components. Additional treatment with antibody and enzymes would further perturb function and lead to erroneous results in any

87 systematic investigation. We felt it necessary to develop an alternative method for obtaining Langerhans cells in reasonable numbers without labelling or further enzymatic treatment following the initial trypsinisation. Using flow cytometry analysis of the light scattering properties and immunofluorescence of epidermal cells we were able to show that OKT6-positive cells could be defined as low RAS-high FAS cells. Using this definition, we found that most of the BL2-positive cells displayed the same characteristics. A majority of human Langerhans cells has been shown to express both H L A - D R and T6 antigens (Fithian et al., 1981) although T6 is proposed as a better marker for these cells in skin sections (Harrist et al., 1983). Surprisingly, in cell suspensions the number of HLA-DR-positive cells is greater than the expected number of T6-positive cells (Fithian et al., 1981; Dezutter-Dambuyant et al., 1984). In the present work, the selection of cells according to their light scattering properties gave a slight difference in favour of OKT6-reactive ceils. The cell population defined as low RAS-high FAS excludes contaminating HLA-DR-positive cells such as lymphoid cells or specialised keratinocytes, acrosyringial keratinocytes (Murphy et al., 1983), that are capable of showing reactivity with antibodies against H L A - D R antigens. Nevertheless the 2 distinct types of HLA-DR-positive cells previously described (Dezutter-Dambuyant et al., 1984) were identified in the sorted cell population. Thus it appears that the selection of cells according to light scattering properties is capable of identifying the Langerhans cell population a n d / o r cells related to the Langerhans cell population which lack Birbeck granules. The selection method described gave a theoretical yield of 70% which is similar to that reported after sorting by others (Morhenn et al., 1982; Scheynius et al., 1982). In conclusion, the sorting of Langerhans cells defined as low RAS-high FAS cells was possible and yielded fixed or viable Langerhans cells in reasonable numbers and purity. This emphasises the relevance of cytometry to the sorting of viable Langerhans cells for use in functional studies and for mammalian somatic cell hybrid production (Berman and France, 1980) and to solving the problem of long-term maintenance of Langerhans cells in tissue culture, by identifying proliferating cell-lines expressing Langerhans cell characteristics.

Acknowledgements The authors are grateful to C. Boujeon for secretarial assistance. This work was supported by Grant INSERM 1982, CRL 10-39.

References Berman, B. and D.S. France, 1980, J. Invest. Dermatol. 74, 323. Bjercke, S., T. Lea, L.R. Braahlen and E. Thorsby, 1984, Scan& J. Immunol. 19, 255. Brunstig, A. and P.F. Mullaney, 1974, Biophys.J. 14, 439. Cordier, G., J.F. Mornex and J.P. Revillard, 1982, in: Cytofluorom6trieet Anticorps Monoclonauxpour le Suivi des Th6rapeutiques: Quo Vadis, eds. P. Gros, F.K. Jansen, P. Poncelet and R. Roncucci (Groupe Sanofi, Montpellier, France) p. 19.

88 Dezutter-Dambuyant, C., G. Cordier, D. Schmitt, M. Faure, C. Laquoi and J. Thivolet, 1984. Br. J. Dermatol. 111, 1. Fithian, E., P. Kung, G. Goldstein, M. Rubenfeld, C. Fenoglio and R.L. Edelson, 1981, Proc. Natl. Acad. Sci. U.S.A. 78, 2541. Harrist, T.J., J.E. Muhlbauer, G.F. Murphy, M.C. Mihm and A.K. Bhan, 1983, J. Invest. Dermatol. 80, 100. Hoffman, R.A., P.C. Kung, W.P. Hansen and G. Golstein, 1980, Proc. Natl. Acad. Sci. U.S.A. 77, 4914. Klareskog, L., M.U. Tjerlund, U. Forsum and P.A. Peterson, 1977, Nature (London) 268, 248. Mahrle, G. and C.E. Orfanos, 1974, Arch. Dermatol. Forsch. 251, 19. Morhenn, V.B., C.J. Benike, D.J. Charron, A. Cox, G. Mahrle, G.S. Wood and E.G. Engleman, 1982. J. Invest. Dermatol. 79, 277. Murphy, G.F., R.S. Shepard, T.J. Harrist, B.R. Bronstein and A.K. Bhan, 1983. J. Invest. Dermatol. 81, 181. Rowden, G., M.G. Lewis and A.K. Sullivan, 1977, Nature (London) 268, 247. Scheynius, A., L. Klareskog, U. Forsum, P. Matssom L. Karlsson, P.A. Peterson and C. Sundstrom, 1982, J. Invest. Dermatol. 82, 452. Schuler, G., J. Aubock and J. Linert, 1983, J. Immunol. 130, 2008. Shapiro, H.M., 1983, Cytometry 3, 227. Shelley, W.B. and L. Juhlin, 1978, Acta Dermatol. Venereol. 58 (Suppl. 79), 7. Stingl, G., E.C. Wolff-Scheiner, W.J. Pichler, F. Gschnait, W. Knapp and K. Wolff, 1977, Nature (London) 268, 245. Stingl, G., S.I. Katz, k. Clement, 1. Green and E.M. Shevach, 1978, J. Immunol. 121, 2005. Wolff, K., 1972, Curr. Probl. Dermatol. 4, 79. Wolff, K. and R.K. Winkelmann, 1967, J. Invest. Dermatol. 48, 50. Yonish-Rouach, E., G. Cordier, J. Cohen, C. Vincent and J. Brochier, 1983, Prot. Biol. Fluids 30, 487.