107
Journal of Immunological Methods, 156 (1992) 107-114
© 1992 Elsevier Science Publishers B.V. All rights reserved 0022-1759/92/$05.00
JIM06494
Quantitation and phenotyping of T cell clones by flow cytometry S t e n S t e m m e and Birgitta K a l l b e r g Department of Clinical Chemistry, Gothenburg University, Gothenburg, Sweden
(Received 29 April 1992, accepted 8 June 1992) In order to efficiently analyze large numbers of T cell clones, a method for analysis of small cell cultures by flow cytometry was developed. The aim was to assess cloning efficiency, growth rate and phenotype of T cell clones. The reliability of the flow cytometer for quantitation of cell populations was documented by repeated analysis of manually counted cell samples. In the concentration range from 3000 to 195,000 cells/ml, the correlation coeffient between counts obtained by the flow cytometer and manual cell counts was 0.999 and within-assay coefficients of variation were below 4%. Cell cultures containing less than 200 ceils were reliably quantified. The technique was applied for analysis of T cell clonings with different mitogens and in the presence of varying amounts of serum. To reduce time and labour, the cultures were analyzed only 10 days after cloning, when the clones contained less than 100,000 cells. The sensitivity of the flow cytometer in the detection of immunolabeled cells made further expansion of cell cultures unnecessary, thus greatly reducing manual labour and experiment turnover time. The commonly used mitogens phytohemagglutinin (PHA) and the antibody OKT3 resulted in comparable cloning efficiencies and clone sizes. Human serum was essential for high cloning efficiency as well as for continued growth, and could not be substituted with an increased amount of fetal calf serum. When cloning with interleukin-2 at 20,000 U / m l , two growing cell types were identified. The majority of the clones contained CD3 +, CD4 +, or CD8 + T cells. Ten out of 60 cultures however, contained cells with the C D 3 - 1 6 / 5 6 + NK cell phenotype, indicating that the culture conditions stimulated proliferation of two different cell types. The described method can be applied for rational analysis of large numbers of minute cell cultures, in for example evaluation of different cloning conditions and estimation of precursor cell frequencies in limiting dilution analysis. The simultaneous phenotyping allows precursor cell analysis under conditions that stimulate growth of more than one cell type. Key words: T lymphocyte;Cell cloning; Cell culture; Flow cytometry
Introduction
The development of techniques for culture of T lymphocytes has provided tools for the investiCorrespondence to: S. Stemme, Department of Clinical Chemistry, Sahlgren's Hospital, Gothenburg University, S-413 45 Gothenburg, Sweden. Tel.: (46) 31-603825; Fax: (46) 31-827610.
gation of antigen specific responses in vitro. Antigen-specific T cell clones can be generated through stimulation with the appropriate antigen together with autologous antigen-presenting cells under limiting dilution conditions. In other instances, for example when the antigen is unknown, a 'library' of T cell clones is desired. Such libraries can then be screened for antigen specificity or clonal composition. There are several
108 ways of achieving polyclonal T cell activation and sustained growth of human T lymphocytes. The most widely used techniques employ stimulation with mitogens such as phytohemagglutinin (PHA) or the mitogenic monoclonal antibody OKT3 (Meuer et al., 1984; Londei et al., 1988). It is generally believed that these different mitogens are truly polyclonal, i.e., stimulate all T cells to the same degree. When a T cell clone library is generated from a mixed lymphocyte population it is essential that the cloning efficiency is high and that potential bias in the clone sample is considered. In the present study, a method was developed for T cell phenotyping and simultaneous counting of cells by flow cytometry. The protocol was applied for evaluation of cloning efficiency, T cell growth rate and possible bias in the clone phenotype, when cloning with different mitogens and in the presence of varying amounts of serum. The sensitivity of the flow cytometer in the detection of immunolabeled cells made it possible to analyze the cultures at a very early stage, only 10 days after cloning, when the clones contained less than 100,000 cells.
Materials and methods
Validation of the flow cytometer for absolute cell enumeration To estimate between-assay variability in flow cytometric measurements of cell concentration, peripheral blood mononuclear cells (PBMC) were isolated from healthy blood donors by FicollPaque density gradient centrifugation, serially diluted in 1 ml 1% paraformaldehyde in phosphatebuffered saline (PBS) (150 mM NaC1, 15 mM phosphate buffer, pH 7.2) and analyzed in duplicates with a 24 h interval. A Becton-Dickinson FACScan flow cytometer with LYSYS II software was used for the analysis. The flow rate was set at position 'high' in all experiments, corresponding to 60 /~1 sample volume/min. Data corresponding to 10,000 counted particles or a maximal analysis time of 180 s was collected from each sample. This procedure was repeated after 4 days. In the computer analysis a gate was set isolating lymphocytes according to their light scatter properties and the number of registered lympho-
cytes/s (rate) was calculated. Lymphocyte concentrations in the diluted samples were obtained by multiplying the cell count from manual counting in a Biirker hemocytometer with the percentage of lymphocytes obtained from the FACS analysis. The correlation between lymphocyte rates and known cell concentrations was then calculated by the least squares method. From all the eight determinations at each concentration in the interval from 380 to 195,000 lymphocytes/ml, a regression line was calculated. The slope was then used to convert cell rates to cell numbers in samples from T cell clonings.
Direct staining of cell cultures for flow cytometry To estimate reliability in the analysis of cell cultures immunostained in Terasaki plates, PBMC were serially diluted and seeded in triplicates in Terasaki plates at a volume of 15 pA. 5 /~1 of antibody solution (FITC-labeled anti-CD3 and phycoerythrin-labeled anti-CD19, directed to human T and B cells respectively, Simultest, Becton-Dickinson, Mountain View, CA) diluted 1 / 8 in PBS, with 0.5% bovine serum albumin and 0.1% NaN 3, were added to each well. The plates were incubated for 15 min at room temperature and the content of each well was then transferred to 200 /.d of 1% paraformaldehyde in PBS and analyzed in the flow cytometer. Data corresponding to 5000 events or 120 s time of analysis was collected.
Evaluation of T cell clonings T cell cloning was performed by limiting dilution. PBMC from healthy donors were isolated as described above, serially diluted 1/2, and seeded in three Terasaki plates (Nunc, Roskilde, Denmark) at each dilution, together with 1 × 10 4 irradiated autologous PBMC feeder cells in a volume of 25/xl/well. RPMI 1640 (Gibco BRL, Uxbridge, UK) with 1 0 0 / z g / m l streptomycin and 100 U / m l penicillin G was used as culture medium. Clonings were performed under different conditions; with either 3 /xg/ml PHA (Welcome, London, UK) or 25 n g / m l OKT3 monoclonal antibody (Ortho Diagnostic Systems, Raritan, N J) as mitogens, and in the presence of varying concentrations of fetal calf serum (FCS) and pooled heat inactivated human serum (HS). Recombi-
109
nant interleukin-2 (IL-2) (Amersham, Buckinghamshire, England) was added at 133 U / m l . One cloning was performed with 20,000 U / m l IL-2, 10% FCS and 10% HS in the absence of OKT3 or PHA. After 10 days of culture in humidified chambers, growing clones were detected visually and 5 /zl of antibody diluted with PBS with 0.5% bovine serum albumin and 0.1% NaN 3, was added to the culture wells. The antibodies used were FITClabeled anti-CD4 and phycoerythrin-labeled antiCD8, or FITC-labeled anti-CD3 and phycoerythrin-labeled anti-CD16 and anti-CD56 (Simultest, Becton-Dickinson, Mountain View, CA). Anti-CD3 recognizes all T cells while anti-CD4 and anti-CD8 recognizes the helper and cytotoxic T cell subsets respectively. CD16 and CD56 is present on subsets of CD3-negative lymphocytes displaying natural killer activity (Lanier et al., 1986). The plates were incubated for 15 min at room temperature and the content of each well was then transferred to 200 /zl of 1% paraformaldehyde in PBS and analyzed in the flow cytometer. Data was collected during 30 s from each sample. Cloning efficiencies (the proportion of seeded cells that give rise to a growing clone) were calculated from the formula (Lefkovits and Waldmann, 1984):
F0=e-U where F 0 is the fraction of negative cultures per total number of cultures and u is the average number of seeded precursor cells per culture.
Results
Reliability of the flow cytometer in the evaluation of cell concentrations in small cell samples To evaluate the stability of the flow cytometer, PBMC from a normal blood donor were serially diluted and analyzed twice with a 24 h interval. This procedure was repeated 4 days later with PBMC from a different donor. Fig. 1 shows the linear relationship between the measured lymphocyte rate in the flow cytometer and the cell concentration obtained by microscopic counting.
y
=
0.000842X
+
1.00
R
:
0.999
200
150 -
g
1O0 -
~
50-
,-I 0
I
50000
I
I
100000
150000
200000
Lymphocytes/mL Fig. 1. P B M C from two different healthy donors were serially diluted in duplicates and analyzed twice with a 24 h interval. The two donors were analyzed with a 4 days interval to estimate stability of the flow cytometer. Lymphocytes were detected according to their light scatter properties. Cell concentration obtained by manual counts on the x axis is plotted against m e a n values of cell rates obtained from flow cytometric analysis on the y axis. The cell concentration range was 3000-195,000 lymphocytes/ml.
In the range from 3000 to 195,000 lymphocytes/ ml, the correlation coefficient was 0.999. Between-assay variability was estimated by comparing the means of the determinations on the 4 days. The coefficients of variation increased from 5% at 195,000 lymphocytes/ml, to 32% at 3000 lymphocytes/ml. The high variablity at the lower cell concentrations may to a large part be explained by differences in the dilution of the two PBMC preparations. The determinations on day two were consistently lower than on day one, indicating a deterioration of the fixed PBMC during overnight storage. This difference was more pronounced at lower cell concentrations and may reflect a protective effect of the formation of a larger cell pellet during storage. This contributes to between-assay variability, which also includes the error in the manual countings. Within-assay variability was calculated from the duplicate determinations of each sample and was less than 4% in the range between 3000 lymphocytes/ml and 195,000 lymphocytes/ml. However, at 390,000 lymphocytes/ml the curve deviated from a straight line, leading to underestimation of the cell concentration (Fig. 2). This indicates that the flow cytometer is stable enough for reliable estimates of cell concentra-
110 300
• ..
200-
•
t~
B.,
I~1
t,9 D.
100 -
..I
¢~
. .~.'..:
.,~
~--..:... :~'¢i. °, . ,
. .'.::
,,~
|
/
!
0
100000
200000
300000
Lymphocytes/mL
tions directly from cell rate values. For accurate comparisons at low cell concentrations between measurements on different days, a calibration may be performed on each day.
Direct labeling of T cell cultures for quantitative flow cytometry To reduce culture time as well as the loss of clones in cell passages, the earliest possible analysis of the clones was desired. The cells were therefore immunostained by addition of antibody directly into the culture medium. The antibody concentration was titrated in preparatory experiments. Addition of 5 lzl antibody prediluted 1/5 resulted in a strong signal with low non-specific binding (Fig. 3). To test overall reliability, serially diluted PBMC were seeded in triplicates in Terasaki plates, immunostained in the wells, transferred to 200 /zl 1% paraformaldehyde in PBS and analyzed in the flow cytometer. In the range tested, from 128 to 32,800 lymphocytes/ well, the correlation between lymphocyte rates (lymphocytes/sec) obtained by flow cytometry and the lymphocyte concentrations obtained by manual counting, was high (Fig. 4) and the CVs of the triplicates were below 12% (range 1.111.0%).
Evaluation of T cell cloning by flow cytometry In order to determine possible differences in cloning efficiency, clone size and phenotype of
.
.K-
•
.. :.. "" i.,*,
- " i ' ~ . . . . "i'~2 . . . . "i'i~3. . . . "i': ,"
CD4-F I TC
400000
Fig. 2. Experimental data as in Fig. 1, in the cell concentration range from 381 to 390,000 lympbocytes/ml, plotted as a linear graph. PBMC were analyzed twice with a 24 h interval (1 and 2) in the flow cytometer. Two different preparations of PBMC, A and B, were analyzed with 4 days interval.
:.
•
"~A.-,....," ~ ~"" "":it ~0"
0
-.
. ~..~.
Fig. 3. PBMC seeded in a Terasaki plate at 67,000 cells/well were stained with diluted antibodies (FITC-labeled anti-CD4 and phycoerythrin-labeled anti-CD8) directly in the well, transferred to 200 ~1 1% paraformaldehyde in PBS and analyzed in the flow cytometer. A gate was set to include lymphocytes according to their light scatter properties.
growing clones, T cells were cloned in Terasaki plates with different mitogens and with varying additions of FCS and HS. Growing cultures were identified by visual screening and 60 clones from each cloning condition were selected for analysis in the flow cytometer. A total of 579 cultures
y=O.OO235x + 0.729 O
R=0.998
80
60 o
0I- 40 E
20
_J
0
10000
20000
30000
Lymphocytes/well Fig. 4. Serially diluted PBMC were seeded in triplicates in Terasaki wells, stained with diluted antibodies directly in the wells, transferred to 200/zl 1% paraformaldehyde in PBS and analyzed in the flow cytometer. A gate was set to detect lymphocytes according to their light scatter properties, x axis shows number of lymphocytes/well as calculated from manual PBMC cell counts and the proportion of lymphocytes obtained from FACS analysis, y axis shows cell rates obtained by flow cytometry as lymphocytes per sec. Nine points in the range from 128 to 32,800 seeded lymphocytes per well are shown.
111 TABLE I RESULTS OF T CELL CLONINGS PERFORMED WITH DIFFERENT MITOGENS AND VARYING SERUM CONTENT AFTER ANALYSIS BY FLOW CYTOMETRY In exp. 1 and 2 IL-2 was added at 133 U / m l and in exp. 3 at 20,000 U/ml. Culture condition Exp. 1 PHA, 10% FCS, 10% HS OKT3, 10% FCS, 10% HS PHA, 10% FCS OKT3, 10% FCS EXp. 2 PHA, 10% FCS, 10% HS OKT3, 10% FCS, 10% HS PHA, 10% FCS, 20% HS OKT3, 10% FCS, 20% HS PHA, 20% FCS OKT3, 20% FCS Exp. 3 IL-2, 10% FCS, 10% HS
Cloning efficiency
no. of clones
Mean clone size (lymphocytes/well + SD)
%CD4 clones
43% 31% 6.4% 3.8% 61% 35% 28% 36% 10% 8% 17%
47 41 9 6 59 60 60 60 60 57 120
19910_+ 13034 16717+_ 18079 6 987 _+8925 3103 _+1 116 13625_+ 13964 12532_+ 13284 15257_+ 14896 17269_+ 13550 7378 + 5 242 4 075 _+2042 15021 -+ 16875
72 51 ND a ND 51 79 84 91 71 55 57
a ND, not done•
were analyzed and the result is summarized in Table I. The cloning efficiency, with either PHA or OKT3 as mitogen, was very low in the absence of human serum. This is in accordance with general experience and the addition of HS is recommended when cloning human T cells (Nutman, 1991). Increasing the amount of FCS to 20% increased cloning efficiency slightly but only to maximally 10%. Addition of HS enhanced T cell growth and a commonly used protocol with 10%
T~
~ L~ 0,,.~
1.1.1 0.. I i
,~. :.'-"
..
".,¢" ~ :...'. :.~ .:I ~.',
.
• ,- .:.".
:.. ~ ..".tJ"
i: ~o
FCS and 10% HS resulted in cloning efficiencies between 28% and 61%. Increasing the amount of HS to 20%, however, did not further enhance cloning efficiency. PHA gave somewhat higher cloning efficiencies than OKT3, 43% and 61% compared with 31% and 35%. With 10% FCS the mean clone sizes after 10 days of culture were also lower than in the presence of HS, and clone sizes were not significantly affected by increasing the amount of FCS
.... Ti~t'"T~2
."
:
. . . . i ' ~ 3 . . . . "i"~
CDI-FITC
"/i o . . . . i ' 6 ' . . . . i ' ~ 2 . . . . i i ~ " . . . . "i"
CD4-FITC
o
b
..-"
. ".•
~, o . . . . i i ~ c ' ~ - ~
.... i~, ....
i;
C D 4 - F I TC
c
Fig. 5. PBMC seeded at limiting dilution in the presence of PHA with 10% FCS, 10% HS were cultured for 10 days. They were then immunostained with phycoerythrin-labeled anti-CD8 and FITC-labeled anti-CD4, in the Terasaki wells and analyzed by flow cytometry, a: CD4-positive clone; b: CD8-positive clone; c: identical analysis of a well judged to be negative by visual inspection• A gate was set to include lymphocytes according to their light scatter properties.
112
to 20%. Increasing the amount of HS from 10% to 20% in the presence of 10% FCS, did not significantly increase clone sizes. No significant differences were seen in the clone sizes generated by P H A or OKT3. In general, clone sizes showed substantial variation, as indicated by the high SDs in Table I. CD4 ÷ and CD8 ÷ cultures were detected in all conditions tested (Fig. 5). The proportion of CD4 ÷ helper T cell clones and CD8 + cytotoxic T cell clones varied. The percentage of CD4 ÷ clones was between 51% and 91%. The percentage of CD4 ÷ cells in the seeded T cell populations were 42%, 60%, and 78% in the three experiments, respectively. Variability in the C D 4 / 8 ratio was substantial and no relation between the C D 4 / 8 ratios and cloning conditions could be seen. Cloning in the presence of 10% FCS and 10% HS with IL-2 at high concentration gave rise to growing cultures with a cloning efficiency of 17% calculated from the average number of seeded T ceils per well. Stimulation with IL-2 may, however, also stimulate growth of a C D 3 - C D 4 C D 8 - lymphocyte population (Vie et al., 1986). The clones were therefore analyzed with a combination of antibodies detecting CD3, CD16, and CD56, which identifies subsets of natural killer (NK) ceils (Lanier et al., 1986) or with antibodies to CD4 and CD8. More than 90% of the clones expressed either CD4, CD8 or CD3 (Fig. 6a), 2"
indicating that the majority of the clones were CD3 ÷ T cells. Ten out of 60 cultures however, contained C D 3 - 1 6 / 5 6 ÷ ceils (Fig. 6b), showing that the culture conditions stimulated the proliferation of two different cell types.
Discussion
For comparison of different cloning conditions, large numbers of T cell clones must be generated and analyzed. Continuous culture of large numbers of T cell clones, however, involves a substantial amount of effort. To reduce labour, the T cell clones were therefore analyzed at a very early stage, directly from the 25/zl wells in which the T cells were cloned. In order to detect possible differences in growth rate as well as in cloning efficiency, the actual number of T cells in each culture was determined. This was facilitated by the development and application of a method for T cell phenotyping and simultaneous counting of cell numbers by flow cytometry. The range and stability of the flow cytometer, when used as a cell counter, was evaluated through repeated analysis of peripheral blood mononuclear cells. In contrast to flow cytometers of other brands, which draw and inject fixed volumes into the flow cell, the FACScan flow cytometer used in the present study injects the
~r
0
¢..i ~
...:. ".. "
......
8":
~:!-
ft.-.
.... ;-A~ . . . . .
. ,/-:,
~
CD3-FITC
a
....
"tl ~ o ' " T 6 ~ .... "~'~2...." / ~ .... T
CD3-FITC
b
~'"T6~
.... " ~
....i ' ~ .... T
CD3-FITC
c
Fig. 6. Clones from P B M C which had been seeded at limiting dilution and cultured in the presence of 20,000 U / m l IL-2, were immunostained with FITC-labeled anti-CD3 and phycoerythrin-labeled anti-CD16 and anti-CD56 antibodies in the Terasaki wells 10 days after start of culture, a: CD3-positive culture; b: CD16/56-positive culture; c: identical analysis of a well judged to be negative by visual inspection. A gate was set to include lymphocytes according to their light scatter properties.
113 sample into the flow cell by administering an overpressure in the tube containing the sample. This means that an estimation of the cell number in the sample must be based on the cell rate, i.e., registered ceils per unit time. The flow rate, and thus the measured cell rate, past the flow cell is dependent on the resistance in the air- and liquid-filled tube system of the instrument. Since these conditions may be assumed to vary with time and condition of the instrument, the stability must be documented if the cell rate is to be used to obtain absolute cell concentrations in the samples. Both the within-assay variability, estimated by repeated analysis of identical samples, and between-assay variability from day to day must be determined. Furthermore, to allow a simple conversion of cell rates into cell concentration in the sample, the linear range must be defined. In the present study, the correlation and variability was tested in two ways: by serial dilution of PBMC in test tubes and by serial dilution of PBMC in 25/xl Terasaki wells. The results from cell dilutions in test tubes demonstrate good linearity in the relation between manually counted cell concentrations and cell counts obtained with the flow cytometer at cell concentrations within the range of approximately 3000-200,000 cells/ml (r = 0.999). At cell concentrations below 3000/ml, the within-assay variability increased, but was below 10% even at 760 lymphocytes/ml. At the lowest cell concentrations, the between-assay variability was up to 30%. This error may, however, to a large part be explained by increasing dilution errors and possibly more rapid cell decomposition at lower cell concentrations during storage. At cell concentrations up to 400,000 lymphocytes/ml, on the other hand, the curve deviated from a straight line leading to an underestimation of cell numbers. This may be explained by a cell crowding effect in the flow chamber. Analysis of PBMC immunostained in 25 ~1 culture wells showed a high correlation with manual cell count (r = 0.998) and within-assay CVs below 12% were obtained with wells containing between 128 and 32,800 cells. This corresponds to between 570 and 145,800 cells/ml, supporting the working range obtained from direct cell dilutions. Reliable cell counts could be obtained from wells containing less than 1000 cells. In fact, CV from
triplicate wells was below 10% also when less than 100 lymphocytes were registred by the flow cytometer. In conclusion, the high degree of linearity, the high correlation with manual counts, and the low within-assay variability in the range from approximately 1000 lymphocytes/ml to 200,000 lymphocytes/ml, indicates this to be a reliable working range. Terstappen et al. (1989) established the validity of the FACScan flow cytometer in the analysis of cell suspensions at concentrations between approximately 200,000 and one million cells/ml. The present results document the reliability of the flow cytometer in enumerations at very low concentrations, below 200,000 cells/ml. In our study, however, the relation between flow cytometric counts and manual counts at cell concentrations above 200,000 cells/ml were no longer linear. This is probably explained by different instrument settings. In the present study, the highest flow rate was used to allow rapid analysis of small cell cultures. The technique was applied for analysis of 579 T cell clones grown with different mitogens and in the presence of varying amounts of serum. PHA and the mitogenic antibody OKT3 resulted in comparable cloning efficiencies and clone sizes. The proportions of CD4 ÷ and CD8 ÷ clones varied substantially, and could not be related to the culture conditions. Human serum was essential for high cloning efficiency as well as for continued growth, and could not be substituted with an increased amount of fetal calf serum. Increasing the amount of human serum from 10% to 20% in the presence of 10% fetal calf seum, however, did not further enhance cloning efficiency or clone growth. When cloning with interleukin-2 at 20,000 U / m l , the majority of growing cultures were CD3 ÷, CD4 +, or CD8 +. This probably reflects stimulation of resting peripheral blood T ceils, expressing the p70, low-affinity IL-2 receptor subunit (Bich-Thuy et al., 1987), as well as a low number of circulating activated T cells expressing the high-affinity IL-2 receptor (Robb et al., 1981). Ten out of 60 cultures, however, contained C D 3 - 1 6 / 5 6 + cells, indicating that the culture conditions stimulated proliferation of different cell types. This illustrates the use of simultaneous
114
phenotyping in the analysis of growing clones. Using different combinations of antibodies, the conditions for clonal growth and precursor cell frequencies of different cell types may be determined. In conclusion, the present study documents the reliability of the flow cytometer for simultaneous quantitation and phenotyping of very small cell samples. The protocol may be applied for rational analysis of large numbers of clones from limiting dilution cloning and for quantitative analysis of heterogeneous cell populations.
Acknowledgements This work was supported by the Swedish Medical Research Council (proj. no. 6816), the Swedish Heart-Lung Foundation, King Gustav V 80th anniversary fund, the National Society against Rheumatism, the Gothenburg Medical Society, and research funds of Gothenburg University.
References Bich-Thuy, L.T., Dukovich, M., Peffer, N.J., Fauci, A.S., Kehrl, J.H. and Greene, W.C. (1987) Direct activation of human resting T cells by IL-2: The role of an IL-2 receptor distinct from the Tac protein. J. Immunol. 139, 1550.
Lanier, L.L., Le, A.M., Civin, C.I., Loken, M.R. and Phillips, J.H. (1986) The relationship of CD16 (Leu-ll) and Leu-19 (NKH-1) antigen expression on human peripheral blood NK cells and cytotoxic T lymphocytes. J. Immunol. 136, 4480. Lefkovits, I. and Waldmann, H. (1984) Limiting dilution analysis of cells of the immune system. I. The clonal basis of the immune response. Immunol. Today 5, 265. Londei, M., Grubeck-Loebenstein, B., DeBerardinis, P., Greenall, C. and Feldmann, M. (1988) Efficient propagation and cloning of human T cells in the absence of antigen by means of OKT3, interleukin-2, and antigen-presenting cells. Scand. J. Immunol. 27, 35. Meuer, S.C., Hussey, R.E., Cantrell, D.A., Hodgdon, J.C., Schlossman, S.F., Smith, K.A. and Reinherz, E.L. (1984) Triggering of the T3-Ti antigen-receptor complex results in clonal T cell proliferation through an interleukin-2-dependent autocrine pathway. Proc. Natl. Acad. Sci. USA 81, 1509. Nutman, T.B. (1991) In: J.E. Coligan, A.M. Kruisbek, D.H. Margulies, E.M. Shevach and W. Strober (Eds.), Current Protocols in Immunology. Greene Publishing and WileyInterscience, New York, pp. 7.19.5. Robb, R.J., Munck, A. and Smith, K.A. (1981) T cell growth factor receptors. Quantitation, specificity and biological relevance. J. Exp. Med. 154, 1455. Terstappen, L.W.M.M., Meiners, H. and Loken, M.R. (1989) A rapid sample preparation technique for flow cytometric analysis of immunofluorescence allowing enumeration of cell populations. J. Immunol. Methods 123, 103. Vie, H., Bonneville, M., Sondermeyer, P., Moreau, J.-F. and Soulillou, J.-P. (1986) Limiting dilution analysis (LDA) of cells responding to interleukin-2 without previous stimulation: Evidence that all responding cells are lymphokineactivated potent effectors. Immunology 57, 351.
Journal of Immunological Methods, 156 (1992) 107-114
© 1992 Elsevier Science Publishers B.V. All rights reserved 0022-1759/92/$05.00
JIM06494
Quantitation and phenotyping of T cell clones by flow cytometry S t e n S t e m m e and Birgitta K a l l b e r g Department of Clinical Chemistry, Gothenburg University, Gothenburg, Sweden
(Received 29 April 1992, accepted 8 June 1992) In order to efficiently analyze large numbers of T cell clones, a method for analysis of small cell cultures by flow cytometry was developed. The aim was to assess cloning efficiency, growth rate and phenotype of T cell clones. The reliability of the flow cytometer for quantitation of cell populations was documented by repeated analysis of manually counted cell samples. In the concentration range from 3000 to 195,000 cells/ml, the correlation coeffient between counts obtained by the flow cytometer and manual cell counts was 0.999 and within-assay coefficients of variation were below 4%. Cell cultures containing less than 200 ceils were reliably quantified. The technique was applied for analysis of T cell clonings with different mitogens and in the presence of varying amounts of serum. To reduce time and labour, the cultures were analyzed only 10 days after cloning, when the clones contained less than 100,000 cells. The sensitivity of the flow cytometer in the detection of immunolabeled cells made further expansion of cell cultures unnecessary, thus greatly reducing manual labour and experiment turnover time. The commonly used mitogens phytohemagglutinin (PHA) and the antibody OKT3 resulted in comparable cloning efficiencies and clone sizes. Human serum was essential for high cloning efficiency as well as for continued growth, and could not be substituted with an increased amount of fetal calf serum. When cloning with interleukin-2 at 20,000 U / m l , two growing cell types were identified. The majority of the clones contained CD3 +, CD4 +, or CD8 + T cells. Ten out of 60 cultures however, contained cells with the C D 3 - 1 6 / 5 6 + NK cell phenotype, indicating that the culture conditions stimulated proliferation of two different cell types. The described method can be applied for rational analysis of large numbers of minute cell cultures, in for example evaluation of different cloning conditions and estimation of precursor cell frequencies in limiting dilution analysis. The simultaneous phenotyping allows precursor cell analysis under conditions that stimulate growth of more than one cell type. Key words: T lymphocyte;Cell cloning; Cell culture; Flow cytometry
Introduction
The development of techniques for culture of T lymphocytes has provided tools for the investiCorrespondence to: S. Stemme, Department of Clinical Chemistry, Sahlgren's Hospital, Gothenburg University, S-413 45 Gothenburg, Sweden. Tel.: (46) 31-603825; Fax: (46) 31-827610.
gation of antigen specific responses in vitro. Antigen-specific T cell clones can be generated through stimulation with the appropriate antigen together with autologous antigen-presenting cells under limiting dilution conditions. In other instances, for example when the antigen is unknown, a 'library' of T cell clones is desired. Such libraries can then be screened for antigen specificity or clonal composition. There are several
108 ways of achieving polyclonal T cell activation and sustained growth of human T lymphocytes. The most widely used techniques employ stimulation with mitogens such as phytohemagglutinin (PHA) or the mitogenic monoclonal antibody OKT3 (Meuer et al., 1984; Londei et al., 1988). It is generally believed that these different mitogens are truly polyclonal, i.e., stimulate all T cells to the same degree. When a T cell clone library is generated from a mixed lymphocyte population it is essential that the cloning efficiency is high and that potential bias in the clone sample is considered. In the present study, a method was developed for T cell phenotyping and simultaneous counting of cells by flow cytometry. The protocol was applied for evaluation of cloning efficiency, T cell growth rate and possible bias in the clone phenotype, when cloning with different mitogens and in the presence of varying amounts of serum. The sensitivity of the flow cytometer in the detection of immunolabeled cells made it possible to analyze the cultures at a very early stage, only 10 days after cloning, when the clones contained less than 100,000 cells.
Materials and methods
Validation of the flow cytometer for absolute cell enumeration To estimate between-assay variability in flow cytometric measurements of cell concentration, peripheral blood mononuclear cells (PBMC) were isolated from healthy blood donors by FicollPaque density gradient centrifugation, serially diluted in 1 ml 1% paraformaldehyde in phosphatebuffered saline (PBS) (150 mM NaC1, 15 mM phosphate buffer, pH 7.2) and analyzed in duplicates with a 24 h interval. A Becton-Dickinson FACScan flow cytometer with LYSYS II software was used for the analysis. The flow rate was set at position 'high' in all experiments, corresponding to 60 /~1 sample volume/min. Data corresponding to 10,000 counted particles or a maximal analysis time of 180 s was collected from each sample. This procedure was repeated after 4 days. In the computer analysis a gate was set isolating lymphocytes according to their light scatter properties and the number of registered lympho-
cytes/s (rate) was calculated. Lymphocyte concentrations in the diluted samples were obtained by multiplying the cell count from manual counting in a Biirker hemocytometer with the percentage of lymphocytes obtained from the FACS analysis. The correlation between lymphocyte rates and known cell concentrations was then calculated by the least squares method. From all the eight determinations at each concentration in the interval from 380 to 195,000 lymphocytes/ml, a regression line was calculated. The slope was then used to convert cell rates to cell numbers in samples from T cell clonings.
Direct staining of cell cultures for flow cytometry To estimate reliability in the analysis of cell cultures immunostained in Terasaki plates, PBMC were serially diluted and seeded in triplicates in Terasaki plates at a volume of 15 pA. 5 /~1 of antibody solution (FITC-labeled anti-CD3 and phycoerythrin-labeled anti-CD19, directed to human T and B cells respectively, Simultest, Becton-Dickinson, Mountain View, CA) diluted 1 / 8 in PBS, with 0.5% bovine serum albumin and 0.1% NaN 3, were added to each well. The plates were incubated for 15 min at room temperature and the content of each well was then transferred to 200 /.d of 1% paraformaldehyde in PBS and analyzed in the flow cytometer. Data corresponding to 5000 events or 120 s time of analysis was collected.
Evaluation of T cell clonings T cell cloning was performed by limiting dilution. PBMC from healthy donors were isolated as described above, serially diluted 1/2, and seeded in three Terasaki plates (Nunc, Roskilde, Denmark) at each dilution, together with 1 × 10 4 irradiated autologous PBMC feeder cells in a volume of 25/xl/well. RPMI 1640 (Gibco BRL, Uxbridge, UK) with 1 0 0 / z g / m l streptomycin and 100 U / m l penicillin G was used as culture medium. Clonings were performed under different conditions; with either 3 /xg/ml PHA (Welcome, London, UK) or 25 n g / m l OKT3 monoclonal antibody (Ortho Diagnostic Systems, Raritan, N J) as mitogens, and in the presence of varying concentrations of fetal calf serum (FCS) and pooled heat inactivated human serum (HS). Recombi-
109
nant interleukin-2 (IL-2) (Amersham, Buckinghamshire, England) was added at 133 U / m l . One cloning was performed with 20,000 U / m l IL-2, 10% FCS and 10% HS in the absence of OKT3 or PHA. After 10 days of culture in humidified chambers, growing clones were detected visually and 5 /zl of antibody diluted with PBS with 0.5% bovine serum albumin and 0.1% NaN 3, was added to the culture wells. The antibodies used were FITClabeled anti-CD4 and phycoerythrin-labeled antiCD8, or FITC-labeled anti-CD3 and phycoerythrin-labeled anti-CD16 and anti-CD56 (Simultest, Becton-Dickinson, Mountain View, CA). Anti-CD3 recognizes all T cells while anti-CD4 and anti-CD8 recognizes the helper and cytotoxic T cell subsets respectively. CD16 and CD56 is present on subsets of CD3-negative lymphocytes displaying natural killer activity (Lanier et al., 1986). The plates were incubated for 15 min at room temperature and the content of each well was then transferred to 200 /zl of 1% paraformaldehyde in PBS and analyzed in the flow cytometer. Data was collected during 30 s from each sample. Cloning efficiencies (the proportion of seeded cells that give rise to a growing clone) were calculated from the formula (Lefkovits and Waldmann, 1984):
F0=e-U where F 0 is the fraction of negative cultures per total number of cultures and u is the average number of seeded precursor cells per culture.
Results
Reliability of the flow cytometer in the evaluation of cell concentrations in small cell samples To evaluate the stability of the flow cytometer, PBMC from a normal blood donor were serially diluted and analyzed twice with a 24 h interval. This procedure was repeated 4 days later with PBMC from a different donor. Fig. 1 shows the linear relationship between the measured lymphocyte rate in the flow cytometer and the cell concentration obtained by microscopic counting.
y
=
0.000842X
+
1.00
R
:
0.999
200
150 -
g
1O0 -
~
50-
,-I 0
I
50000
I
I
100000
150000
200000
Lymphocytes/mL Fig. 1. P B M C from two different healthy donors were serially diluted in duplicates and analyzed twice with a 24 h interval. The two donors were analyzed with a 4 days interval to estimate stability of the flow cytometer. Lymphocytes were detected according to their light scatter properties. Cell concentration obtained by manual counts on the x axis is plotted against m e a n values of cell rates obtained from flow cytometric analysis on the y axis. The cell concentration range was 3000-195,000 lymphocytes/ml.
In the range from 3000 to 195,000 lymphocytes/ ml, the correlation coefficient was 0.999. Between-assay variability was estimated by comparing the means of the determinations on the 4 days. The coefficients of variation increased from 5% at 195,000 lymphocytes/ml, to 32% at 3000 lymphocytes/ml. The high variablity at the lower cell concentrations may to a large part be explained by differences in the dilution of the two PBMC preparations. The determinations on day two were consistently lower than on day one, indicating a deterioration of the fixed PBMC during overnight storage. This difference was more pronounced at lower cell concentrations and may reflect a protective effect of the formation of a larger cell pellet during storage. This contributes to between-assay variability, which also includes the error in the manual countings. Within-assay variability was calculated from the duplicate determinations of each sample and was less than 4% in the range between 3000 lymphocytes/ml and 195,000 lymphocytes/ml. However, at 390,000 lymphocytes/ml the curve deviated from a straight line, leading to underestimation of the cell concentration (Fig. 2). This indicates that the flow cytometer is stable enough for reliable estimates of cell concentra-
110 300
• ..
200-
•
t~
B.,
I~1
t,9 D.
100 -
..I
¢~
. .~.'..:
.,~
~--..:... :~'¢i. °, . ,
. .'.::
,,~
|
/
!
0
100000
200000
300000
Lymphocytes/mL
tions directly from cell rate values. For accurate comparisons at low cell concentrations between measurements on different days, a calibration may be performed on each day.
Direct labeling of T cell cultures for quantitative flow cytometry To reduce culture time as well as the loss of clones in cell passages, the earliest possible analysis of the clones was desired. The cells were therefore immunostained by addition of antibody directly into the culture medium. The antibody concentration was titrated in preparatory experiments. Addition of 5 lzl antibody prediluted 1/5 resulted in a strong signal with low non-specific binding (Fig. 3). To test overall reliability, serially diluted PBMC were seeded in triplicates in Terasaki plates, immunostained in the wells, transferred to 200 /zl 1% paraformaldehyde in PBS and analyzed in the flow cytometer. In the range tested, from 128 to 32,800 lymphocytes/ well, the correlation between lymphocyte rates (lymphocytes/sec) obtained by flow cytometry and the lymphocyte concentrations obtained by manual counting, was high (Fig. 4) and the CVs of the triplicates were below 12% (range 1.111.0%).
Evaluation of T cell cloning by flow cytometry In order to determine possible differences in cloning efficiency, clone size and phenotype of
.
.K-
•
.. :.. "" i.,*,
- " i ' ~ . . . . "i'~2 . . . . "i'i~3. . . . "i': ,"
CD4-F I TC
400000
Fig. 2. Experimental data as in Fig. 1, in the cell concentration range from 381 to 390,000 lympbocytes/ml, plotted as a linear graph. PBMC were analyzed twice with a 24 h interval (1 and 2) in the flow cytometer. Two different preparations of PBMC, A and B, were analyzed with 4 days interval.
:.
•
"~A.-,....," ~ ~"" "":it ~0"
0
-.
. ~..~.
Fig. 3. PBMC seeded in a Terasaki plate at 67,000 cells/well were stained with diluted antibodies (FITC-labeled anti-CD4 and phycoerythrin-labeled anti-CD8) directly in the well, transferred to 200 ~1 1% paraformaldehyde in PBS and analyzed in the flow cytometer. A gate was set to include lymphocytes according to their light scatter properties.
growing clones, T cells were cloned in Terasaki plates with different mitogens and with varying additions of FCS and HS. Growing cultures were identified by visual screening and 60 clones from each cloning condition were selected for analysis in the flow cytometer. A total of 579 cultures
y=O.OO235x + 0.729 O
R=0.998
80
60 o
0I- 40 E
20
_J
0
10000
20000
30000
Lymphocytes/well Fig. 4. Serially diluted PBMC were seeded in triplicates in Terasaki wells, stained with diluted antibodies directly in the wells, transferred to 200/zl 1% paraformaldehyde in PBS and analyzed in the flow cytometer. A gate was set to detect lymphocytes according to their light scatter properties, x axis shows number of lymphocytes/well as calculated from manual PBMC cell counts and the proportion of lymphocytes obtained from FACS analysis, y axis shows cell rates obtained by flow cytometry as lymphocytes per sec. Nine points in the range from 128 to 32,800 seeded lymphocytes per well are shown.
111 TABLE I RESULTS OF T CELL CLONINGS PERFORMED WITH DIFFERENT MITOGENS AND VARYING SERUM CONTENT AFTER ANALYSIS BY FLOW CYTOMETRY In exp. 1 and 2 IL-2 was added at 133 U / m l and in exp. 3 at 20,000 U/ml. Culture condition Exp. 1 PHA, 10% FCS, 10% HS OKT3, 10% FCS, 10% HS PHA, 10% FCS OKT3, 10% FCS EXp. 2 PHA, 10% FCS, 10% HS OKT3, 10% FCS, 10% HS PHA, 10% FCS, 20% HS OKT3, 10% FCS, 20% HS PHA, 20% FCS OKT3, 20% FCS Exp. 3 IL-2, 10% FCS, 10% HS
Cloning efficiency
no. of clones
Mean clone size (lymphocytes/well + SD)
%CD4 clones
43% 31% 6.4% 3.8% 61% 35% 28% 36% 10% 8% 17%
47 41 9 6 59 60 60 60 60 57 120
19910_+ 13034 16717+_ 18079 6 987 _+8925 3103 _+1 116 13625_+ 13964 12532_+ 13284 15257_+ 14896 17269_+ 13550 7378 + 5 242 4 075 _+2042 15021 -+ 16875
72 51 ND a ND 51 79 84 91 71 55 57
a ND, not done•
were analyzed and the result is summarized in Table I. The cloning efficiency, with either PHA or OKT3 as mitogen, was very low in the absence of human serum. This is in accordance with general experience and the addition of HS is recommended when cloning human T cells (Nutman, 1991). Increasing the amount of FCS to 20% increased cloning efficiency slightly but only to maximally 10%. Addition of HS enhanced T cell growth and a commonly used protocol with 10%
T~
~ L~ 0,,.~
1.1.1 0.. I i
,~. :.'-"
..
".,¢" ~ :...'. :.~ .:I ~.',
.
• ,- .:.".
:.. ~ ..".tJ"
i: ~o
FCS and 10% HS resulted in cloning efficiencies between 28% and 61%. Increasing the amount of HS to 20%, however, did not further enhance cloning efficiency. PHA gave somewhat higher cloning efficiencies than OKT3, 43% and 61% compared with 31% and 35%. With 10% FCS the mean clone sizes after 10 days of culture were also lower than in the presence of HS, and clone sizes were not significantly affected by increasing the amount of FCS
.... Ti~t'"T~2
."
:
. . . . i ' ~ 3 . . . . "i"~
CDI-FITC
"/i o . . . . i ' 6 ' . . . . i ' ~ 2 . . . . i i ~ " . . . . "i"
CD4-FITC
o
b
..-"
. ".•
~, o . . . . i i ~ c ' ~ - ~
.... i~, ....
i;
C D 4 - F I TC
c
Fig. 5. PBMC seeded at limiting dilution in the presence of PHA with 10% FCS, 10% HS were cultured for 10 days. They were then immunostained with phycoerythrin-labeled anti-CD8 and FITC-labeled anti-CD4, in the Terasaki wells and analyzed by flow cytometry, a: CD4-positive clone; b: CD8-positive clone; c: identical analysis of a well judged to be negative by visual inspection• A gate was set to include lymphocytes according to their light scatter properties.
112
to 20%. Increasing the amount of HS from 10% to 20% in the presence of 10% FCS, did not significantly increase clone sizes. No significant differences were seen in the clone sizes generated by P H A or OKT3. In general, clone sizes showed substantial variation, as indicated by the high SDs in Table I. CD4 ÷ and CD8 ÷ cultures were detected in all conditions tested (Fig. 5). The proportion of CD4 ÷ helper T cell clones and CD8 + cytotoxic T cell clones varied. The percentage of CD4 ÷ clones was between 51% and 91%. The percentage of CD4 ÷ cells in the seeded T cell populations were 42%, 60%, and 78% in the three experiments, respectively. Variability in the C D 4 / 8 ratio was substantial and no relation between the C D 4 / 8 ratios and cloning conditions could be seen. Cloning in the presence of 10% FCS and 10% HS with IL-2 at high concentration gave rise to growing cultures with a cloning efficiency of 17% calculated from the average number of seeded T ceils per well. Stimulation with IL-2 may, however, also stimulate growth of a C D 3 - C D 4 C D 8 - lymphocyte population (Vie et al., 1986). The clones were therefore analyzed with a combination of antibodies detecting CD3, CD16, and CD56, which identifies subsets of natural killer (NK) ceils (Lanier et al., 1986) or with antibodies to CD4 and CD8. More than 90% of the clones expressed either CD4, CD8 or CD3 (Fig. 6a), 2"
indicating that the majority of the clones were CD3 ÷ T cells. Ten out of 60 cultures however, contained C D 3 - 1 6 / 5 6 ÷ ceils (Fig. 6b), showing that the culture conditions stimulated the proliferation of two different cell types.
Discussion
For comparison of different cloning conditions, large numbers of T cell clones must be generated and analyzed. Continuous culture of large numbers of T cell clones, however, involves a substantial amount of effort. To reduce labour, the T cell clones were therefore analyzed at a very early stage, directly from the 25/zl wells in which the T cells were cloned. In order to detect possible differences in growth rate as well as in cloning efficiency, the actual number of T cells in each culture was determined. This was facilitated by the development and application of a method for T cell phenotyping and simultaneous counting of cell numbers by flow cytometry. The range and stability of the flow cytometer, when used as a cell counter, was evaluated through repeated analysis of peripheral blood mononuclear cells. In contrast to flow cytometers of other brands, which draw and inject fixed volumes into the flow cell, the FACScan flow cytometer used in the present study injects the
~r
0
¢..i ~
...:. ".. "
......
8":
~:!-
ft.-.
.... ;-A~ . . . . .
. ,/-:,
~
CD3-FITC
a
....
"tl ~ o ' " T 6 ~ .... "~'~2...." / ~ .... T
CD3-FITC
b
~'"T6~
.... " ~
....i ' ~ .... T
CD3-FITC
c
Fig. 6. Clones from P B M C which had been seeded at limiting dilution and cultured in the presence of 20,000 U / m l IL-2, were immunostained with FITC-labeled anti-CD3 and phycoerythrin-labeled anti-CD16 and anti-CD56 antibodies in the Terasaki wells 10 days after start of culture, a: CD3-positive culture; b: CD16/56-positive culture; c: identical analysis of a well judged to be negative by visual inspection. A gate was set to include lymphocytes according to their light scatter properties.
113 sample into the flow cell by administering an overpressure in the tube containing the sample. This means that an estimation of the cell number in the sample must be based on the cell rate, i.e., registered ceils per unit time. The flow rate, and thus the measured cell rate, past the flow cell is dependent on the resistance in the air- and liquid-filled tube system of the instrument. Since these conditions may be assumed to vary with time and condition of the instrument, the stability must be documented if the cell rate is to be used to obtain absolute cell concentrations in the samples. Both the within-assay variability, estimated by repeated analysis of identical samples, and between-assay variability from day to day must be determined. Furthermore, to allow a simple conversion of cell rates into cell concentration in the sample, the linear range must be defined. In the present study, the correlation and variability was tested in two ways: by serial dilution of PBMC in test tubes and by serial dilution of PBMC in 25/xl Terasaki wells. The results from cell dilutions in test tubes demonstrate good linearity in the relation between manually counted cell concentrations and cell counts obtained with the flow cytometer at cell concentrations within the range of approximately 3000-200,000 cells/ml (r = 0.999). At cell concentrations below 3000/ml, the within-assay variability increased, but was below 10% even at 760 lymphocytes/ml. At the lowest cell concentrations, the between-assay variability was up to 30%. This error may, however, to a large part be explained by increasing dilution errors and possibly more rapid cell decomposition at lower cell concentrations during storage. At cell concentrations up to 400,000 lymphocytes/ml, on the other hand, the curve deviated from a straight line leading to an underestimation of cell numbers. This may be explained by a cell crowding effect in the flow chamber. Analysis of PBMC immunostained in 25 ~1 culture wells showed a high correlation with manual cell count (r = 0.998) and within-assay CVs below 12% were obtained with wells containing between 128 and 32,800 cells. This corresponds to between 570 and 145,800 cells/ml, supporting the working range obtained from direct cell dilutions. Reliable cell counts could be obtained from wells containing less than 1000 cells. In fact, CV from
triplicate wells was below 10% also when less than 100 lymphocytes were registred by the flow cytometer. In conclusion, the high degree of linearity, the high correlation with manual counts, and the low within-assay variability in the range from approximately 1000 lymphocytes/ml to 200,000 lymphocytes/ml, indicates this to be a reliable working range. Terstappen et al. (1989) established the validity of the FACScan flow cytometer in the analysis of cell suspensions at concentrations between approximately 200,000 and one million cells/ml. The present results document the reliability of the flow cytometer in enumerations at very low concentrations, below 200,000 cells/ml. In our study, however, the relation between flow cytometric counts and manual counts at cell concentrations above 200,000 cells/ml were no longer linear. This is probably explained by different instrument settings. In the present study, the highest flow rate was used to allow rapid analysis of small cell cultures. The technique was applied for analysis of 579 T cell clones grown with different mitogens and in the presence of varying amounts of serum. PHA and the mitogenic antibody OKT3 resulted in comparable cloning efficiencies and clone sizes. The proportions of CD4 ÷ and CD8 ÷ clones varied substantially, and could not be related to the culture conditions. Human serum was essential for high cloning efficiency as well as for continued growth, and could not be substituted with an increased amount of fetal calf serum. Increasing the amount of human serum from 10% to 20% in the presence of 10% fetal calf seum, however, did not further enhance cloning efficiency or clone growth. When cloning with interleukin-2 at 20,000 U / m l , the majority of growing cultures were CD3 ÷, CD4 +, or CD8 +. This probably reflects stimulation of resting peripheral blood T ceils, expressing the p70, low-affinity IL-2 receptor subunit (Bich-Thuy et al., 1987), as well as a low number of circulating activated T cells expressing the high-affinity IL-2 receptor (Robb et al., 1981). Ten out of 60 cultures, however, contained C D 3 - 1 6 / 5 6 + cells, indicating that the culture conditions stimulated proliferation of different cell types. This illustrates the use of simultaneous
114
phenotyping in the analysis of growing clones. Using different combinations of antibodies, the conditions for clonal growth and precursor cell frequencies of different cell types may be determined. In conclusion, the present study documents the reliability of the flow cytometer for simultaneous quantitation and phenotyping of very small cell samples. The protocol may be applied for rational analysis of large numbers of clones from limiting dilution cloning and for quantitative analysis of heterogeneous cell populations.
Acknowledgements This work was supported by the Swedish Medical Research Council (proj. no. 6816), the Swedish Heart-Lung Foundation, King Gustav V 80th anniversary fund, the National Society against Rheumatism, the Gothenburg Medical Society, and research funds of Gothenburg University.
References Bich-Thuy, L.T., Dukovich, M., Peffer, N.J., Fauci, A.S., Kehrl, J.H. and Greene, W.C. (1987) Direct activation of human resting T cells by IL-2: The role of an IL-2 receptor distinct from the Tac protein. J. Immunol. 139, 1550.
Lanier, L.L., Le, A.M., Civin, C.I., Loken, M.R. and Phillips, J.H. (1986) The relationship of CD16 (Leu-ll) and Leu-19 (NKH-1) antigen expression on human peripheral blood NK cells and cytotoxic T lymphocytes. J. Immunol. 136, 4480. Lefkovits, I. and Waldmann, H. (1984) Limiting dilution analysis of cells of the immune system. I. The clonal basis of the immune response. Immunol. Today 5, 265. Londei, M., Grubeck-Loebenstein, B., DeBerardinis, P., Greenall, C. and Feldmann, M. (1988) Efficient propagation and cloning of human T cells in the absence of antigen by means of OKT3, interleukin-2, and antigen-presenting cells. Scand. J. Immunol. 27, 35. Meuer, S.C., Hussey, R.E., Cantrell, D.A., Hodgdon, J.C., Schlossman, S.F., Smith, K.A. and Reinherz, E.L. (1984) Triggering of the T3-Ti antigen-receptor complex results in clonal T cell proliferation through an interleukin-2-dependent autocrine pathway. Proc. Natl. Acad. Sci. USA 81, 1509. Nutman, T.B. (1991) In: J.E. Coligan, A.M. Kruisbek, D.H. Margulies, E.M. Shevach and W. Strober (Eds.), Current Protocols in Immunology. Greene Publishing and WileyInterscience, New York, pp. 7.19.5. Robb, R.J., Munck, A. and Smith, K.A. (1981) T cell growth factor receptors. Quantitation, specificity and biological relevance. J. Exp. Med. 154, 1455. Terstappen, L.W.M.M., Meiners, H. and Loken, M.R. (1989) A rapid sample preparation technique for flow cytometric analysis of immunofluorescence allowing enumeration of cell populations. J. Immunol. Methods 123, 103. Vie, H., Bonneville, M., Sondermeyer, P., Moreau, J.-F. and Soulillou, J.-P. (1986) Limiting dilution analysis (LDA) of cells responding to interleukin-2 without previous stimulation: Evidence that all responding cells are lymphokineactivated potent effectors. Immunology 57, 351.
