Colony formation by subpopulations of human T lymphocytes: VI. Further studies on colony phenotype, function, and cloning efficiency

CELLULAR IMMUNOLOGY 109,407-4 18 ( 1987)

Colony Formation by Subpopulations of Human T Lymphocytes

VI. Further Studies on Colony Phenotype, Function, and Cloning Efficiency

ANDERS PEDERSEN, CHRISTIAN NOLSBE, CARSTEN R~~PKP, AND MOGENS H. CLA&SON

Laboratory OfExperimental Hematology and Immunology, Institute ofMedical Anatomy, The Panum Institute, University of Copenhagen, DK-2200 Copenhagen N, Denmark

Received April 14,1987; accepted May 20,1987

Phytohemagglutinin (PHA)-induced colony formation in semisolid agar medium by human peripheral blood T lymphocytes showed an increasing cloning efficiency with decreasing numbers of cultured cells. Ninety percent of CD4+ cells (inducer/helper phenotype) and 20% of CD8+ cells (cytotoxic/suppressor phenotype) formed colonies when cultured at lo-200 cells/ml culture in the presence of sheep red blood cells (SRBC) and a source of interleukin-2 (IL-2). Probably all T-colony-forming cells, but none of the subsequent colony cells, expressed the Leu- 8 antigen. The cloning efficiencies of FACS-sorted cells expressing the natural killer antigenic phenotypes Leu-7* and CD16+ were found to be less than 1%. The costimulatory effect of red blood cells for colony formation was specific for SRBC and not observed in the presence of red cells obtained from seven other species including man. All T-lymphocyte colonies obtained from unseparated peripheral blood mononuclear cells expressed the CD25 antigen (IL2 recep tor) and colonies were always composed of either CD4+ or CD8+ cells. None of the colony cells expressed the Leu-8 or the CD16 antigens. By their specific morphology in agar culture the majority of colonies composed of CD4+ cells were easily recognized, but approximately one- third of the CD4+ colonies could not be distinguished from colonies composed of CD8+ cells. On expansion of individual colonies in liquid subculture in the presence of interleukin-2, ap proximately 15% of the colonies developed natural killer (NK)-like cytotoxic activity, being capable of direct killing of K562 tumor cells. It is concluded that the present method for growing human T colonies exhibits the same cloning efficiency as the most efficient liquid culture systems. Individual T colonies are composed exclusively of T inducer/helper or T cytotoxic/sup pressor cells, they are never of mixed phenotype, and they do not contain cells of natural killer phenotype. Regulatory mechanisms influencing colony formation are operating between and within the various subsets of T lymphocytes. o 1987 Academic PTSS, IIIC.

INTRODUCTION

Human peripheral T lymphocytes, cultured in semisolid agar medium in the presence of phytohemagglutinin (PHA), form colonies composed of mature and functionally differentiated T lymphocytes (l-3). This culture system has been extensively used to study the influence of hormones, lymphokines, and pharmacological agents on clonal T-lymphocyte growth in vitro (4-l 1). The T-colony system has also been used to examine the number of T-lymphocyte colony-forming cells (TL-CFC) in peripheral blood in pathological conditions such as autoimmune diseases, cancer, and

407

0008-8749187 $3.00 Copyright Q 1987 by Academic Press, Inc. All rights of reproduction in any form resewed.

408 PEDERSEN ET AL.

acute and chronic virus infections, including AIDS ( 12- 19). To understand the immunological significance of data obtained in the T-colony assay system, it is of impor- tance to know the contribution to colony formation by the various T-cell subsets under normal conditions. Although it has been shown that both helper/inducer and cytotoxic/suppressor T-cell subsets form T colonies in agar medium (1) the maximal cloning efficiency of these subsets is not established at the present time. Moreover, colony formation by T cells, characterized by the recent important phenotypical and functional marker the Leu-8 antigen (20) and colony growth by cells exhibiting the natural killer (NK) phenotypes, the Leu-7 and the CD 16 antigens (2 1 ), has not been studied previously in the present T-colony assay system.

In the present work colony formation and the antigenic phenotype of the colony progeny were examined after individual cells and colonies were labeled with fluorochrome-conjugated monoclonal antibodies. Furthermore, the cloning efficiency in our standard colony assay system was compared with the efficiency in a microversion of this system and in a limiting dilution system, and the role of sheep red blood cells (SRBC) as feeder cells compared with red blood cells from other species was studied. Finally, the direct cytotoxic activity against the NK-sensitive target cell K562 of indi- vidually expanded colonies was examined.

MATERIALS AND METHODS

Cell Preparations

Heparinized peripheral blood from adult humans was separated on Lymphoprep (Nygaard, Oslo, Norway). Blood mononuclear cells (BMC) were washed three times and resuspended in RPM1 1640 (Gibco, Paisley, Scotland) supplemented with 10% fetal calf serum (Flow, Erwin, Scotland), r&utamine (1.2 r&Q, penicillin (500 IU/ ml), streptomycin (500 pg/ml), and 2-mercaptoethanol(5 X 1 Oe5 n/i). In some experiments BMC (107/ml) were incubated in petri dishes for 2 hr and nonadherent cells were recovered and used for culture.

Fresh peripheral blood for red cell preparations was obtained from the lymphocyte donors (see above) and from horse, sheep, cow, dog, rabbit, guinea pig, and rat. The cells were washed three times and irradiated (3000 rad) prior to use.

Antibodies

The following fluorochrome-labeled monoclonal antibodies against human lymphocyte antigens (in parentheses) were used: anti-Leu-3a (CD4, T helper cells), anti- Leu-2a (CD8, T cytotoxicfsuppressor cells), anti-Leu-3a/Leu-2a mixture labeled with phycoerythrin and fluorescein, respectively, anti-Leu-8 (T cell subsets), anti-Leu- 11 a (CD 16, NK cells), and anti-IL-2 receptor (CD25). Nonconjugated anti-Leu-7 antibody (NK cells) was used in indirect immunofluorescence. All antibodies were pur- chased from Becton Dickinson.

Flow Microjluorometry and Cell Sorting

BMC were labeled with antibodies following the instructions of Becton Dickinson. The cells were resuspended in phosphate-buffered saline (PBS) containing 2% fetal calf serum (FCS). A Becton Dickinson FASC III instrument with a tunable argon-

COLONY FORMATION BY T-LYMPHOCYTE SUBPOPULATIONS 409

ion laser (Spectra Physics, Model 164/05) was used. To sort fluorescent cells an area was selected which contained cells presenting light scatter signals of small lymphocytes combined with fluorescence intensity above the channels containing background fluorescent cells. The exact level for positive cells was decided so less than 1% of background fluorescent cells were included in the sort area. Thus monocytes, unlabeled lymphocytes, and dead cells were excluded from the sorted population of the fluorescence-positive cells. Cells were sorted into sterile tubes containing 20% fetal calf serum in RMPI 1640 medium. Both cell samples and sorted cells were kept at 4°C during sorting.

Agar Culture

Standard version. Cells were cultured in plastic dishes in 1 ml 0.45% agar medium containing 1% sheep red blood cells (SRBC), 5 X lop5 A4 2-mercaptoethanol, 5% FCS, 5% of a IL-2 source (Lymphocult-T, Bionetics, Munich, West Germany), pretested for optimal colony-stimulating activity, and 0.1% phytohemagglutinin (PHA, Wellcome, Kent, UK). The medium used was Dulbecco’s modified Eagle’s medium. Cultures were incubated for 6 days in a fully humidified atmosphere containing 7% COz after which time the SRBC were lysed in 1 ml NH&l. Aggregates of more than 40 cells were scored as colonies at Day 6 of culture using a dissection microscope. Due to the diffise growth of colonies, the upper limit for accurate count- ing is approximately 600 colonies per plate.

Microversion. The same culture medium, incubation time, and type of microscopy were used as in the standard version. Aliquots of 20 ~1 agar medium containing the BMC were plated in replicates of 30 wells in Terasaki culture plates (Nunc, Denmark).

Liquid Culture

Limiting dilution. The same culture medium as above but without agar was used for these cultures. Cells were cultured in 20-~1 volumes using Terasaki plates as above. Cultures were grown in the absence or presence of 1% SRBC. At Day 7 of culture the number of wells containing proliferating cells were scored by an inverted microscope.

Colony Expansion and Cytotoxicity Assay

Individual colonies were picked up at Day 7 and expanded in liquid culture for 10 days in round-bottom 200~~1 microplates (Nunc, Roskilde, Denmark). The culture medium was RPM1 1640 supplemented with 10% FCS, 5 X IO-’ 2-mercaptoethanol, 2% fresh glutamine, and 5% of a source of IG2 (Lymphocult-T, Biotest, Dreieich, West Germany).

Prior to cytotoxicity assay, 100~1 supematant was removed from each microcul- ture and replaced with 100 ~1 RPM1 1640 medium containing 2000 5’Cr-labeled K562 cells. Total releasable counts was obtained from incubation of target cells in 5% acetic acid, and spontaneous releasable counts by assaying wells containing only target cells in normal culture medium.

410 PEDERSEN ET AL.

]-1---;I 6 loo 200 300 400.800 2000

z Number of cultured cells

I%. 1. (a) Limiting dilution analysis of BMC in the presence (0) or absence @) of 1% SRBC. Each point indicates the percentage of wells in a 30-microwell tray without growing cells. The dotted lines represent the 95% confidence limits. (b) Colony formation in 20-PI(O) or l-ml (0) agar culture in the presence (00) or absence (WI) of 1% SRBC. Each point represents the mean colony counts of thirty 20-4 or twelve l-ml cultures, respectively. (c) Relative frequency of colony formtion as a function of BMC seeded in l-ml agar cultures. Each point represents the mean colony count of four to eight 1 -ml cultures.

Labeling and Phenotyping of Individual Colonies

Individual colonies were picked up from the agar at Day 6 to 7 of culture using a fine Pasteur pipet. Colonies were placed in nondiluted antibodies for 90 min at 4°C. The colonies were then transferred to a test tube and washed twice at 600g in 10 ml serum-free RPM1 1640 culture medium. Individual colonies were examined in a fluorescence microscope at 500X.

RESULTS

T-Lymphocyte Colony Proliferation Obtainedfrom Blood Mononuclear Cells: Com- parison among Three DQ@erent Culture Systems and Dependence of SRBC

Limiting dilution. Figure la shows the results of a typical experiment assessing the ability of BMC to form T-cell colonies in a liquid culture system using 30-60 repli-


cates in 20-~1 microwells for culture. Cells were grown in the presence or absence of 1% sheep red blood cells. Each point in the figure represents the results from scoring of 30-60 microwells. By Day 7 positive wells contained one or more tight cell aggregates consisting of at least 100 cells. As seen in the figure only cells cultured in the presence of SRBC formed aggregates. According to one-hit limiting dilution theory, the loga- rithm to the fraction of nonresponding cultures when plotted versus the number of cells per culture should fit a straight line if each of the responding cultures is developed from one cell. As shown in Fig. la this was certainly the case. The frequency of clon- ally proliferating cells in BMC was calculated to be 150.

Agur culture. Figure lb shows T-cell colony formation by BMC cultured in our standard colony medium including SRBC. These cultures were performed either in the same type of microwells (i.e., 20 el) as those used for the limiting dilution experiments or in one culture dish. There is a linear cell dose relationship between the number of cultured cells and the number of developing colonies, the frequency of colony formation being 1: 10.

When cells are plated in standard 1 -ml cultures there exists a slight but significant decrease in the number of developing colonies with an increasing number of cultured cells (Fig. lc). No colonies are formed in this culture system in the absence of SRBC when less than 1000 BMC are cultured. There exists a specific requirement for SRBC as feeder cells since the number of colonies in the presence of autologous erythrocytes, as well as erythrocytes from other species including rabbit, guinea pig, rat, dog, and horse, was highly inferior to that with SRBC for stimulation of T-cell colony formation (Fig. 2).

Antigenic Phenotype ofIsolated T-Cell Colonies BMC were cultured according to our standard method at 200 cells per milliliter of

culture. At Day 7 of culture individual colonies were removed from the agar and labeled with fluorochrome-conjugated antibodies.

Anti-CD4/CD8 antibodies. This combination of antibodies was used to determine the frequency of cells in individual colonies with one of the antigenic phenotypes, cells with double phenotype, and cells without either of the two antigens. Table 1 shows the results: 60-75% of all colonies were composed of pure CD4+ cells whereas 25-30% of the colonies were composed of pure CD8 cells. None of the colonies contained CD4-/CD8- cells and none of the colonies contained double positive cells. Considering a frequency of B cells in BMC of approximately 30%, the ratio between CD4+ and CD8+ cells in peripheral blood equals the ratio between CD4+ and CDV T-cell colonies.

Anti-Leu-8 and CD16 antibodies. Individual colonies were tested for Leu-8+ cells. Table 1 shows that none of the colonies contained Leu-8+ cells, a finding which sharply contrasts the high frequency (approx 75%) of Leu-8+ cells in BMC. Table 1 also shows that none of the colonies grown from BMC contained CD16+ cells, whereas all colonies contained between 50 and 100% IL-2 receptor-expressing cells (CD25).

Morphology of CD4’ and CD8+ Colonies In the present culture system all colonies develop within the agar matrix and not

at the surface, as found in many other systems (3, 5, 7, 9, 10, 12, 13). Two different

412 PEDERSEN ET AL.

500 1000

Number of cultured cells

FIG. 2. Numbers of T colonies obtained with increasing numbers of seeded BMC using 1% red blood cells from different species incorporated in the agar matrix. Data are means of quadruplicate cultures. 0, SRBC; 0, pig; A, rabbit; 0, horse; 0, cow, *, guinea pig; n , dog; A, human.

growth characteristics of colonies are apparent: Colonies may grow in a very diffuse fashion with individual colony cells scattered over a large area within the agar. An example of such a diffuse colony is shown in Fig. 3a. Colonies may also grow in a more dense fashion, the individual colony cells being confined to a rather limited

TABLE 1

Antigenic Phenotype of Cells in T-Lymphocyte Colonies Grown from Peripheral Blood Mononuclear Cells”

Percentage of labeled T-lymphocyte colonies

Expt CD4+ CD8+ CD4+/CD8+ CD4/CD8- Leu-8+ CD16+ CD25+

I 62 (31)b 38 (22) O(1) 0 (47) 0 (75) 0 100’ 2 86 (49) 14(13) O(1) 0 (38) 0 (75) n.d. n.d.

D Two hundred cells were plated per milliliter of culture and 100 colonies were analyzed in each experiment except for CD1 6 and CD25 where 50 colonies were analyzed.

b Numbers in parentheses show the percentage of labeled cells in peripheral blood mononuclear cells. ’ Colonies contained 50- 100% positive cells.


FIG. 3. Morphology of T colonies grown from BMC. (a) A colony composed of very diffusely growing cells scattered over a large area within the agar. Individual colonies of this type contained either CD4+ or CD8+ cells. (b) A colony growing in a very dense fashion. The individual colony cells are confined to a rather limited area within the agar. This colony type contains exclusively CD4+ cells.

area (Fig. 3b). Approximately 20-30% of the colonies grown from the BMC exhibit a morphology between that of typical diffuse and dense colonies. Attempts were made to evaluate the CD4 and CD8 phenotypes of diffuse and dense colonies, respectively. Out of 50 diffuse colonies stained with antiCD4/CD8 antibodies, 30 colonies were composed of CD4+ cells and 20 colonies of CD8+ cells. In contrast, 50 out of 50 dense colonies analyzed contained only CD4+ cells. Thus, in the present culture system one can morphologically separate a subset of T-cell colonies-approximately 40% of all colonies derived from BMC-composed of CD4+ cells.

Colony Formation by FACS-Sorted Cell Subsets

BMC were labeled with following fluorescein-conjugated monoclonal antibodies: anti-CD4, anti-CD8, anti-Leu-8, antXD16, and anti-Leu-7 antibodies. In separate experiments cells were labeled with anti-Leu-8 and sorted on a FAG, followed by labeling with CD4 or CD8 antibodies and resorting. Figure 4 shows colony formation by the various subsets as a function of numbers of cells cultured at very low cell density. Data are from one single donor. The slopes of the curves in Fig. 4 were used to calculate the maximum frequency of colony-forming cells within the individual subsets. The calculated frequencies are shown in Table 2: CD4+, 90%; CD8+, 20%, Leu-8+, 60%; Leu-8-, 5%; Leu-8+/CD4+, 60%; Leu-8+/CD8+, 20%; CD16+, 0.3%; and Leu-7+, 0.3%. As mentioned above (Fig. lc), a common finding in these experiments was that cells plated at cell numbers higher than 500 cells/ml showed a gradual decrease in cloning frequency (data not included). When FACS-sorted CD8+ and CD4+ cells were mixed prior to agar culture there was a net decrease in colony formation compared with the sum of colony numbers obtained from the two subsets seeded alone. These data are not included here since they are similar to our previous results (1). Leu-8+ cells as opposed to Leu-8- cells showed a high cloning frequency contrast-

414 PEDERSEN ET AL.

300

260

220 o

5 180 r: 3 110 f -

I

100 I

'E

0

g 75-

';;

:

2

so-

2

25/-&gg) IO 0 .

8

,

12 20 50 loo 200 LOO 600 1600 3200

Number of cultured cells

FIG. 4. Colony formation by subpopulation of T cells positively selected by FACS sorting and cultured at low cell density. Each point represents the mean colony count of four to eight replicate cultures. 0, CD4; 0, CD8; A, Jm-8; v, CD4/Leu-8; V, CDI/Leu-8; 0, Leu-8-; w, Lew7/CD16.

ing the total absence of Leu-8+ cells in the developing colonies (see above). The few colonies which developed from the LeuX subset might represent “contamination” of the sorted cells by weakly fluorescent Leu-8+ cells. Obviously, our cloning system for T cells does not allow proliferation of NK cell subsets, i.e., Leu-7+ and the CD 16+ cells.

NK-like Cytotoxicity of Expanded Colonies

Table 3 shows that approximately 14% of colonies expanded for 10 days in IL-2- containing medium are capable of direct killing of NK-sensitive K562 tumor cells (P < 0.05) in a 4-hr “Cr-release assay. In the presence of 1% PHA, 15% of the expanded colonies killed K562 cells. Apparently, none of these NK-like cytotoxic colonies exhibits the CD 16 phenotypic marker (Table 1).

DISCUSSION

The present work shows that the T-cell cloning efficiency in the standard version of our T-colony assay system is five times higher than in limiting dilution culture and

TABLE 2

Cloning Frequency of Subsets of T Lymphocytes”

BMC CD4+ CD8+ Leu-8+ Leu-8- Leu-8+/CD4+ Leu8+/CD8+ Leu-7+ CD16+

1 O-20% 90% 20% 60% 5% 60%

LI Calculated from the steepest part of the curves in Fig. 4.

20% <I% <I%


TABLE 3

Expanded Human T-Cell Colonies Can Kill K562 Tumor Cells’

PHA added b Cytotoxic colonies’

- 8/56 (14%) + 9/60 (15%)

u Four-hour 5’Cr-release assay. b One percent per assay culture. ’ Number of cytotoxic colonies/total number of colonies assayed, percentage cytotoxic colonies is in

parentheses. Colonies were considered cytotoxic when the “Cr release (cpm) was three times above the standard deviation of the mean cpm of cultures without colony cells.

comparable to the cloning efficiency obtained in a microversion of the agar culture system. Thus although studies have indicated that cell cultures in very minute quanti- ties of liquid culture medium improve cloning efficiency for T cells (22,23), it appears that our l-ml agar culture system is an optimal method for growing human T-cell colonies. The phenotype of individual colonies grown from BMC showed that all colony cells were either CD4+ or CD8+. None of the colonies contained a mixture of the two cell subsets and none of the colony cells were double labeled. This observation is in disagreement with our previous study (1) in which we observed that a minority, approx 3%, of all T colonies contains a mixture of both CD4+ and CD@ cells. How- ever, in contrast to the previous work (I), the present results are based on direct labeling of colonies with a anti-Leu-2+/Leu-3+ antibody combination, a technical improvement which might explain the observed discrepancy. The finding of either pure CDS+ or CD4+ T-cell colonies is also in disagreement with more recent work which demonstrated a mixture of the two T-cell phenotypes in the majority of developing T colonies (3). There exist, however, some important differences between the two culture systems. In the work by Farcet et al. (3) BMC were plated in 0.3% agar culture medium at lo4 to 5 X lo4 cells/ml and all developing colonies grew at the surface of the agar culture medium. In such a cell-crowded system there is a great risk for nonclonality for the majority of developing colonies (24,25) which would explain the presence of mixed colonies. Although mixing experiments with two sets of donor cells containing one or both of two isoenzymes, followed by enzyme analysis of individual colonies, appeared to rule out nonclonality in (3); isoenzyme typing of expanded colonies although used in a number of studies, is not the most optimal control for clonality. Prior to isoenzyme typing, individual colonies have to be expanded to lo5 cells (24). During the expansion period of an initially mixed colony (composed of cells from two donors and/or syngeneic CD4’ and CD8+ cells) one of the cell types might obtain a growth advantage due to both immunological and nonimmunological regulatory mechanisms. For example, CD4+ cells have been shown to become refrac- tory to IL-2 growth promotion earlier than CD8+ cells (26). CDS+ colony cells inhibit growth of CD4+ colony cells (1). Or one cell type from a mixed colony might on IL- 2-induced expansion turn into lymphokine-activated killer cells (27) and eliminate other cell types growing in the culture. Finally, the existence of a slow-growing cell population within the progeny of a mixed colony might not be detected because of the limited sensitivity of the isoenzyme analysis. In addition to the use of very low cell

416 PEDERSEN ET AL.

numbers in our standard method (lOO-lOOO/ml), the claim that this colony system is true clonal was based on mixing experiments (1) where cells from the same donor were separated on a FACS sorter in Leu-2at and Leu-3a’ subsets and then mixed prior to agar culture. The developing colonies from such experiments assayed at Day 6-7 of culture never did show a “mixed” phenotype but were composed of either Leu-2a+ or Leu-3a+ cells, data which strongly indicate the clonality of the present assay system.

The high cloning efficiency of the present culture system is dependent on the presence of SRBC in the agar medium whereas erythrocytes from other species including the human failed to stimulate T-colony formation to the same extent as SRBC. When SRBC were present in an agar layer below that containing the seeded lymphocytes the cloning efficiency decreased by one or two orders of magnitude (data not shown). The alternative pathway of T-cell activation can be triggered by antibodies specific for certain epitopes (T 11 2 and T 11 3) on the SRBC receptor (E-receptor, CD2) present on all peripheral T cells (28,29) and antibodies reacting with the SRBC-binding site of the CD2 block PHA responses (30), indicating that PHA may activate T cells through this surface receptor. The present observation, that SRBC are a prerequisite for PHA-induced T-colony formation in culture systems with very low numbers of cultured cells, may suggest that SRBC under these conditions bind specifically to the CD2 antigen and produce conformational changes which potentiate the stimulatory effect of PHA also bound to this molecule. In contrast to our results, Woods and Lowenthal(3 1) showed that bovine erythrocytes, although not as effective as SRBC, also stimulate human T-cell colony formation. These results may indicate cross-reac- tivity for the E-receptor by SRBC and some types of bovine erythrocytes.

All colonies tested for the Leu-8 antigen were negative. This is a surprising observation because the majority (approx 80%) of peripheral T cells express the Leu-8 marker (20, 2 1) and we found that probably all T-colony-forming cells themselves also express the Leu-8 antigen. Since both expanded T-colony cells and the anti-Leu-8 antibody have been found to identify identical regulatory subsets of human T lymphocytes (1, 2, 3, 20,2 1) our present findings suggest that the Leu-8 antigen is primarily expressed on functionally active, but nonproliferating peripheral T cells.

Colony formation by hematopoietic cells give rise to colonies composed of macrophages, granulocytes, mixed granulocyte-macrophages, and cells which contain natural cytotoxicity. These colony types can be distinguished from each other on the basis of their morphology and growth pattern in agar culture (32,33). Likewise, using our standard culture method, the majority of colonies composed of CD4+ cells could be morphologically identified. Thus approximately 40% of all colonies derived from BMC grow as very dense aggregates of cells and belong to this category of T colonies. Work is in progress to study whether this type of CD4+ colony contains a specific subset of functionally active T-helper lymphocytes.

The cloning efficiency of peripheral blood T lymphocytes cultured in semisolid culture medium is generally found to be at least one order of magnitude below that obtained by limiting dilution in liquid culture (22, 34). Thus the most efficient agar culture systems including our previously reported ones (1) showed that approximately 5-10% of BMC form colonies, each being composed of at least 40 cells. In contrast, in the liquid feeder cell- and factor-dependent system developed by Moretta and co-workers (34) virtually all cultured T cells produced a colony. However, as


shown in the present work, when very low numbers of purified T cells were seeded in agar culture in the presence of SRBC and a source of IL-2, cloning efficiencies comparable to those reported for the limiting dilution systems were obtained. More- over, the present use of stimulator-y agents (SRBC, IL-2) in a limiting dilution micro- culture system, known to be extremely efficient for growing T cells from BMC and bone marrow (22,23), resulted in cloning efficiencies which were inferior to the agar culture system. At present we cannot explain present and previous findings that the cloning efficiency of BMC and even purified T-cell subpopulations (1) appears to decrease with increasing numbers of cultured cells. However, in light of work in the mouse (35) and recent data obtained with human T cells (36) we suggest the existence of suppressor activity for cell proliferation inherent in various subpopulations of oth- erwise phenotypically identical T cells. Such suppressor activity may simply be di- luted out of the culture system when very low cell numbers are seeded.

Studies on FACS-sorted T-cell and NK cell subpopulations showed that approximately 90 and 20% of helper and cytotoxic/suppressor phenotype cells, respectively, form colonies, whereas less than 1% of cells of the NK phenotypes, CD 16+ and Leu- 7+, proliferate in the present system. The lower cloning efficiency of T cells present in cultures of BMC as compared to FACS-purified T-cell subpopulations is in line with previous results showing that CDS+ cells in physiological cell numbers exhibit natural suppressor activity against colony formation by CD4+ cells ( 1).

Although Leu-7- and CD 16-positive cells do not proliferate in the present system approximately 14% of all expanded T colonies were directly capable of killing NK- sensitive K562 tumor cells. These cytotoxic T-cell colonies did not express the CD 16 antigen typical for human NK cells (2 1) and may thus be considered to belong to the so-called non-MHC-restricted subset of cytotoxic T cells (37). Clonal proliferation of such NK-like T cells in the present colony system strongly suggests that IL-2, which strongly augments T-colony formation (1) and is necessary for the expansion culture, may be a trigger of this form of cytotoxic T-cell activity.

In conclusion the present results show that our standard method for growing human T-cell colonies exhibits the same cloning efficiency as liquid culture systems. The vast majority of T colonies are grown from Leu-8+ BMC and are composed of either CD4+ or CD8+ cells. Colony formation is specifically dependent upon the presence of SRBC in the agar culture. The majority of CD4+ colonies can be recognized in the agar culture by their specific morphological appearance. Although cytotoxic activity against NK-sensitive K562 tumor cells can be induced in a number of expanded T colonies, the cytotoxic colony cells do not display the CD16+ NK phenotype.

REFERENCES

1. Claesson, M. H., Petersen, J., Senderstrup-Hansen, G., Rapke, C., and Serensen, T., Cell. Immunol. 81,276, 1983.

2. SBnderstrupHansen, G., Petersen, J., and Claesson, M. H., Scund. J. Immunol. 19,48 1, 1984. 3. Farcet, J. P., Gourdin, M.-F., Calvo, C., Oudrhiri, N., Divine, M., Bouguet, J., Fradelizzi, D., Senik,

A., and Reves, F., Eur. J. Immunol. 15, 1067, 1973. 4. Cla&son, M. H., and Rijpke, C., Clin. Exp. Immunol. 54,554, 1983. 5. Winkelstein, P., Simon, L., Wood, D., Machen, L. L., Shadduck, R. K., and Waheed, A., Immunology

58, 173, 1986. 6. Woods, G. M., and Lowenthal, R. M., Exp. Hematol. 12,301, 1984.

418 PEDERSEN ET AL.

7. Oudrhiri, N., Farcet, J. P., Gourdin, M. F., Divine, M., Bouguet, J., Fradelizi, D., and Reyes, F., J. Immunol. 135,1813, 1985.

8. Claesson, M. H., Andersen, V., and SBnderstrupHansen, G., Clin. Exp. Immunol. 34,364, 1978. 9. Kashiwa, H., Hyodo, S., Kishi, T., Karakawa, T., Kittaka, E., Suzawa, T., Sakano, T., and Usui, T.,

Clin. Exp. Immunol. 60,509, 1985. 10. Jourdan, M., Commes, T., and Klein, B., Immunology54,249, 1985. 11. Larsson, E.-L., Iscove, N. N., and Coutinko, A., Nature (London) 283,664, 1980. 12. Dao, C., Marie, J. P., Bemadou, A., and Bilski-Pasquier, G., Immunology34,741, 1978. 13. Foa, R., and Catovsky, D., Clin. Exp. Immunol. 36,488, 1979. 14. Wilson, J. D., and Dalton, G., AJEAK54,27, 1976. 15. Davis,S.,Blood58, 1053, 1981. 16. Mukhopadhyaya, R., Advani, S. H., and Gangal, S. G., Acta Haematol. 70,357, 1983. 17. Cldsson, M. H., and Andersen, V., Clin. Exp. Immunol. 38,483, 1979. 18. Bernstein, M. L., Winkelstein, A., and Dobson, S. A., Arthritis. Rheum. 23,385, 1980. 19. Winkelstein, A., Klein, R. S., Evans, T. L., Dixon, B. W., Holder, W. L., and Weaver, L. D., J. Immu-

nol. 134, 151, 1985. 20. Kansas, G. S., Wod, G. S., Fishwild, D. M., and Engleman, E. G., J. Immunol. 134,2995, 1985. 21. Lanier, L. L., Engleman, E. G., Gatenby, P., Babcock, G. F., Warner, N. L., and Herzenberg, L. A.,

Immunol. Rev. 74, 143, 1983. 22. Moreau, N.-F., and Miller, R. G., J. Immunol. 130, 1139, 1983. 23. Reimann, J., and Miller, R. G., J. Exp. Med. 158, 1672, 1983.

24. Farcet, J.-P., and Testa, U., Exp. Hematol. 10, 172m, 1982. 25. Lindahl-Kiessling, K., and Karlberg, I., Stand. J. Immunol. 15,525, 1982.

26. Gullberg, M., and Smith, K. A., J. Exp. Med. 163,270, 1986. 27. Grimm, E. A., Mazunder, A., Zhang, H. Z., and Rosenberg, S. A., J. Exp. Med. 155,1823, 1982.

28. Meuer, S. C., Hussey, R. E., Fabbi, M., Fox, D. A., Acute, O., Fitzgerald, K. A., Hodgdon, J. C., Protentis, J. P., Schlossman, S. F., and Reinherz, E. L., CeN36,897, 1984.

29. Fox, D. A., Schlossman, S. F., and Reinherz, E. L., J. Immunol. 136,1945, 1986. 30. OFlynn, K., Krensky, A. M., Beverley, P. C. L., Burakoff, S. J., and Linch, D. C., Nature (London)

313,686,1985.

31. Woods, G. M., and Lowenthal, R. M., J. Clin. Lab. Immunol. 6, 175, 1981. 32. Metcalf, D., In “Hemopoietic Colonies” (P. Rentchnick, Ed.), pp. l-227. Springer-Verlag, Berlin/

Heidelberg/New York, 1977. 33. Claesson, M. H., Olsson, L., Martinsen, and Brix-Paulsen, P., Exp. Hematol. 10,708, 1982.

34. Moretta, A., Pantaleo, G., Moretta, L., Cerottini, J. C., and Mingori, M. C., J. Exp. Med. 157, 743,

1983. 35. Cooper, J., Eichmann, K., Fey, K., Melchers, I., Simon, M. M., and Weltzien, H. U., Immunol. Rev.

79,63, 1984. 36. Rosenkrantz, K., DuPont, B., and Flomenberg, N., Proc. Natl. Acad. Sci. USA 82,4508, 1985.

37. Lanier, L. L., Phillips, J. H., Hackett, J., Jr., Tutt, M., and Kumar, V., J. Zmmunol. 137,2735, 1986.

Documents

Colony formation by subpopulations of human T lymphocytes: VI. Further studies on colony phenotype, function, and cloning efficiency