Phenotype and Function of Peripheral Blood and Bone Marrow T-Cell Colonies: Identification of C D 3 - , 4 - , 8 Autoreactive T Cells Brigitte Autran, Guy Gorochov, Ioannis Theodorou, Marie Claude Couty, and Patrice Debre
A B S T R A C T : The existence of immature autoreactive T cells has been examined in peripheral blood tPB) and bone marrow (BM) derived T-lymphocyte colonies (TLC) that have previously been shown to be potentially generated from CD3-negative B M - T cells. An extensive phenotypical analysis of total and T-depleted TLC showed that both PB- and BM-derived TLC contained a mean of 5 % immature T lymphocytes (ITL), which were negative for the CD3, CD4, and CD8 antigens but displayed the CD2 and CD7 antigens. The only detectable immune functions of these isolated ITL were an allo- and an autoreactivity without cytolytic activity. The self-reactivity of l T L was not detected in bulk non T-depleted TLC cells and seemed to be actively suppressed by autologous mature T cells. In addition, the auto-MLR of l T L was totally inhibited by a n t i - H L A class II but not by anti-class I monoclonal antibody (MoAb) and only partially by anti-CD4 moAb, whereas the anti-CD3 and anti-CD2 MoAbs gave no inhibition. Once activated, ITL could acquire in culture a mature T cell CD3 + CD4 + phenotype. The C D 3 - , 4 - , 8 - auto-reactive T cells present in T colonies could represent pre- orpost-thymic cells that have not yet undergo or that have escaped the thymic selection. ABBREVIATIONS BM bone marrow CD cluster of differentiation FITC fluorescein isothiocyanate ITL immature T lymphocytes leukocyte-conditioned medium LCM MHC major histocompatibility complex
MoAb PBL TCR TLC
monoclonal antibody peripheral blood lymphocytes T-cell receptor T-lymphocyte colonies
INTRODUCTION The significance and the role of self-reactive T lymphocytes in the T-cell repertoire is a critical question in immunology. First defined as cells responsible for the autologous mixed lymphocyte reaction (A-MLR) [1], little is known about the origin of autoreactive T ceils [ 2 - 6 ] . It has been suggested that their precursors
From the Laboratoire d'Immunologie, UA CNRS 625." CHU Piti~-Salpgtri~re, Paris, France. Address reprint requests to." Dr. B. Autran, Laboratoire d'Immunologie, D~partment d'H~matologie, CHU Pitig-Salpgtri~re, 91 Bd de l'Hapital, 75013 Paris, France. Received February 26, 1988; acceptedAugust I, i988.
Human Immunology24, 11I- 124 (1989) © AmericanSocietyfor Histocompatibilityand Immunogenetics,1989
111 0198-8859/89/S3.50
112
B. Autran et ai. may belong to a particular subset that could require activating signals to differentiate into functional helper T cells [6]. In the thymus, epithelial cells could play a role in activating subsets of human autologous double-negative thymocytes to proliferate [7,8]. Although controversy exists whether double immature thymocytes can migrate to peripheral organs, data in mice have suggested that aucoreactive T cells could derive from immature thymus-dependent precursors m peripheral organs [9]. Indeed, recent observations in the murine model have demonstrated that Thyl +, L y t l - , Lyt2-, and L3T4- cells can be activated and cal, differentiate into a subset of autoreactive mature L3T4+ helper T cells [5]. In order to demonstrate a similar phenomenon in humans, T-cell colonies were used as a source of immature T cells. T-lymphocyte colonies (TLC) provide an interesting model for the study of the in vitro T-cell differentiation since they can be generated from C D 3 - bone marrow (BM) mononuclear cells [10-13~. Indeed, we recently established the conditions by which T-cell clones can be isolated from mature T-cell-depleted BM cells and we identified the growth properties of C D 3 - , 4 . 8 - precursor cells [13]. In the data presented herein. we have analyzed the phenotype and function of peripheral blood and BM T colonies derived from healthy human volunteers and identified a subpopulation of phenotypically immature CD2 +, CD7 +, C D 3 - , 4 - , 8 - T cells. I n functional tests, these cells were not cytotoxic nor did they secrete interleukin 2 cIL2). In contrast, they were able to proliferate in autologous as well as in allogeneic MLR. Their self-reactivity, suppressed by mature T cells, was inhibited by anti-HLADR and, only partially, by anti-CD4 monoclonal antibodies (MoAbs), whereas it was not suppressed by anti-CD2 nor anti-CD3 MoAbs. These data suggest that self-reactive immature T cells can be isolated in vitro from human TLC.
METHODS
Cell suspensions. Peripheral blood lymphocytes (PBL) from normal volunteers and BM mononuclear cells (MNC) from normal allogeneic BM-transplant donors were isolated by centrifugation on a Ficoll-Hypaque density gradient (Pharmacia, France).
T colony assay. The T-colony assay was performed in a semisolid culture using a double layer method as previously described [10]. Briefly, undertayers were constituted of RPMI-1640 medium supplemented with 10% human pooled pretested sera, 1% glutamine, nonessential aminoacids and sodium pyruvate (Gibco), 1% PHA-M (Difco Laboratories), 2% Leucocyte-Conditioned-Medium (LCM), and 0.5% agar (Bacto-Agar, Difco Lab). MNC (1.5 × 106/ml) were plated in the upper layer, which consisted of 0.3% agar-supplemented culture medium without PHA and LCM. Cells were cultivated in Petri dishes (55 mm and 100 mm, Greiner, France) at 37°C in fully humidified 5% CO2 armosphere~ Within 7-8 days after plating, clusters of at least 50 cells were scored as colonies under an inverted microscope. Colonies growing at the surface of the upper layer were harvested after resuspension in RPMI-1640, pooled and desegregated, and then extensively washed. Leukocyte-conditioned medium (LCM) preparation. Large batches of LCM were prepared as previously described [14]. Briefly, PBL were incubated in RPMI-1640 medium supplemented with 2% human pooled sera, PHA-P 1/~g/ml (Wellcome Laboratories, UK), PMA × 1 × 10-7M (Sigma, USA) and an aUogeneic EpsteinBarr virus (EBV)-transformed lymphoblast0id cell line for 2 hr. After extensive washings, cells were cultured for 2 days. Culture supernatants were then col-
CD3-, 4 - , 8 - Autoreactive T Cells
113
lected and stored at -80°C until use. Samples were tested for their interleukin 2 (IL2) activity on an IL2-dependent cell line.
Complement-mediated cytolysis. Suspensions of pooled colony cells were incubated at 4°C for 30 min with optimal concentrations of two cocktails of anti-T cell MoAbs: (1) IOT4 (CD4, Immunotech France), T3 (CD3, D. Bourrel, Rennes, France), and A 50 (CD5, A. Bernard, Villejuif, France), and (2) C D 2 - - G 1 4 4 ; C D 5 - - A 5 0 ; CD7--I21 (kindly provided by A. Bernard, Villejuif, France). Two baths of baby rabbit complement, previously tested for its absence of spontaneous cytotoxicity (CRTS, Besan~on, France), were then successively added and mixtures were incubated for 30 min at 37°C. Controls consisted of cells incubated with medium alone and complement. Cells were then extensively washed and used for phenotypic and functional analysis. The T depletion was assessed by immunofluorescence (IF) analysis using Ethidium Bromide (EB) (Sigma, USA) as a marker of mortality. Percentages of residual viable T cells (FITC +, E B - ) were then quantified. Immunofluorescence analysis. Cell surface markers of total and T-depleted colony cells were analyzed with an indirect IF technique using a large panel of MoAbs as first step reagents: C D 1 - - O K T 6 (Ortho Pharmaceutical, France); CD2--T11 (S. Carrel, Lausanne, Switzerland); OKT11 (Ortho Pharmaceutical France); C D 3 - T3 (Dr. D. Bourrel, CDTS; Rennes, France); C D 4 - - I O T 4 (Immunotech, France); CD5--T101 (S. Carrel) and A50 (A. Bernard); C D 6 - - M B G 6 (G. Janossy, London, UK); CD7--Lau A1 (S. Carrel) and I21 (A. Bernard); C D 8 - IOT8 (Immunotech); CD11a--LFA1 (D. Olive, Marseille-Lumigny, France); C D 1 1 b - - O K M 1 (Ortho); CD16--Leu 11b (Becton-Dickinson, France); C D 1 9 - - B 4 (Coultronics, France); and CD25--Tac (H. Waldmann, Bethesda, USA). Nonclusterized MoAbs were also used: anti-HLA class II antigens; antiD R - - D 1-12 (D. Charron, Paris, France); anti-DP--B7.21 (Dr. Fritz-Bach, Minneapolis, USA); anti-DQ--Leu-10 (Becton-Dickinson, France); Leu 7 (Becton Dickinson); D-44 (A. Bernard) [15]. Several anti-TCRod/3 heterodimer MoAbs were used: WT31 was a gift of Dr. De Vries (UNICET, Lyon, France); 31/162, 42/161, and 65 were gifts of Dr. A. W. Boylston [16]. After washings, cells were incubated with FITC-coupled goat-anti-mouse F(ab')2 fragments (Cappel). All incubations were performed at 4°C for 20 min. Cells were analyzed for their IF with a Leitz microscope (Orthoplan) and data were confirmed by a cytofluorometric analysis (FACSTAR, Becton Dickinson).
Proliferation assay ofT-colony cells. Single cells suspensions of total or T-depleted colony cells harvested at day 8 from agar cultures were cultured (1 x 106/ml) in liquid RPMI-1640 culture medium supplemented with 10% human pooled sera, 1% glutamine, 1% nonessential aminoacids, and 1% sodium pyruvate. T-cell proliferations were induced in triplicate cultures in round-bottom, 96-well plates (Greiner, France), either by 25-Gy irradiated allogeneic or autologous PBL (1 x 106/ml) or by optimal concentrations of soluble antigens: Streptodornase-Streptokinase (Varidase, VAR, 5-20 ~g/ml, Lederle, France) or PPD (2-10/zg/ml, Institut Pasteur, Paris, France) in the presence of 50-Gy irradiated autologous PBL (1 × 105/ml) as antigen-presenting cells. The capacity of colony cells to proliferate was determined by tritiated (3H)-thymidine, (37Bq/well, Amersham, France) incorporation during the last 18 hr of a 5-day or a 6-day culture for the MLR assays or the antigen-proliferation assays, respectively. Cells were harvested (Skatron) and radioactivity incorporated to D N A was measured in a beta scintillation counter (Packard).
114
B. Autran et ai.
Limiting dilution analysis of autoreactive cells in TLC cells. Responder TLC cells were plated in 96-we! l microtiter plates in liquid RPMI medium supplemented as described above for proliferation assays at dilutions varying from 1 to 250 cells per well. For each dilution; 144 wells were plated. Autologous, 25-Gy irradiated stimulating PBL were added in each well at a concentration of 1 × 104 cells per well in a final volume o f 200 Izl. In control wells, responder cells were plated alone. Recombinant IL2 (r-IL2) was added to cultures (50 U/ml) at day 1. At day 7, supernatants were replaced by new culture medium plus IL2. At day i4, proliferating wells were counted and effector cell frequencies were estimated by Poisson distribution analysis [ 17].
Inhibition of MLR assays, Monoclonal antibodies in ascitic fluids were added at the beginning o f the allogeneic and autologous MLR assays. The following MoAbs were used: D 1112, directed against the D R molecule, W6.32, a n t i - H L A class I antigen, XT1 (A. Bernard, France) anti-CD2, T3 (D. Bourrel, France), anti-CD3, and BL4 (J. Brochier, France); anti-CD4 antigen. The inhibiting dilutions had been previously determined from dose response curves. Percentages of inhibition were calculated as followed: 1 -
inhibited cpm - control cpm × 100 noninhibited cpm - control cpm
IL2 production assay. Single~cell suspensions o f total and T-depleted Colony cells were incubated at 1 × 106/ml in liquid supplemented R P M I - t 6 4 0 medium with P H A - P (1/zg/ml) and Cultured for 48 hr. After centrifugation, culture supernatants were Collected and tested for their IL2 activity on a human IL2±dependent T-cell line obtained after 4 weeks of continuous culture o f healthy donoVs PBL initiated with P H A , P ~l/zg/ml) and maintained with LCM alone. T h e n 6 × 10! washed IL2-sensitive cells were cultured in 200/21 o f supplemented RPMI-1640 medium with serial dilutions o f the samples to be tested. After a 30-hr incubation, the proliferation was measured by 3 H - T d r incorporation. The !L2 activity of the tested samples was determined from the activity of a reference IL2 preparation simultaneously tested.
Direct cytotoxicity assay. T h e cytotoxicity of the T-colony cells was assayed against K562 or autologous PHA-blasts target cells. Then 2 × 106 target cells were incubated overnight at 37°C with 100/zCi sodium-chromate (51Cr)(Amersham, France) and extensively washed before plating. Effector cells were added in six serial dilutions from 100:1 to 3 : 1 effector-target cell ratios and cukured in duplicate experiments for 4 hr in RPMI-1640-supplemented culture medium at 37°C in a humidified 5% CO2 incubator. After centrifugation, supernatants were collected and measured for their radioactivity in a Gamma Counter (Packard, USA). Maximal 51Cr release was obtained in adding the N P 4 0 detergent onto target cells. Percentage o f 5lCr release was determined as followed: tested cpm - spontaneous cpm 1 - maximal cpm - spontaneous cpm × 100
Immunofluorescence analysis of cultured immature T lymphocytes. After complementmediated T depletion of PB-derived TLC cells, the residual viable cells devoid of detectable C D 3 + , 4 + , 8 + tymphocytes were seeded at 5 cells per well in microtiter culture plates with 1 × t05 25-Gy irradiated, allogeneic or autologous PBL in culture medium defined above supplemented with r-IL2. Culture medium was replaced by fresh medium plus r-IL2 every 3 days and irradiated stimulating
C D 3 - , 4 - , 8 - Autoreactive T Cells
115
cells were added every 12 days. A growth of TLC-derived cells could be detected in more than 50% of wells. The-cell surface phenotype of cultured cells was determined, as described above, 8 days after the third stimulation with PBL. RESULTS Phenotypical Analysis of TLC from PBL and BM Mononuclear Cells An IF analysis of PB- or BM-derived TLC pooled cells was performed at day 8 of in vitro PHA-driven, semisolid cultures. As shown in Table 1, 95% TLC cells had the phenotype of mature T cells, mostly CD4+, and displayed the od/3 T-cell receptor as shown by their positivity for the WT-31 MoAb and a lower but consistent number of cells reactive with other anti-TCRc~//3 MoAbs: 31/162, 42/
TABLE 1
Immunofluorescence analysis of PB- and BM-derived T colony cells: existence of a C D 3 - , 4 - , 8 - cell subset O r i g i n o f T - c o l o n y cells
Monoclonal antibodies Cluster of
Peripheral blood
Bone marrow
Untreated
T-depleted b
Untreated
T-depleted
Antibody
cells
ceils
cells
cells
CD1
OKT6
.
CD2
OKT11
98
95
97
93
CD3 CD4
OKT3 T3 IOT4
94 93 78
----
96 94 72
----
CD5 CD6 CD7
T101 MBG6 I21
90 89 99
--96
91 90 97
--94
CD8
LauA1 IOT8
98 16
95 --
97 22
95 --
WT31 31/162
90 10
---
91 NT d
-N T 't
42/161 65 Tac Dl-12 B7.21 L e u 10 D44 LFA-1 OKM1 B4 Leul lb Leu 7
4 5 43 49 47 45 43 75 . . . .
--39 48 52 47 41 --
" 47 45 42 43 55 70
" 5 51 50 50 NT e --
differentiation"
TCR
CD25 MHC Class II N.CI. CD1 la CDllb CD19 CD16 N.C1.
.
.
. . . .
.
. . . .
.
. . . .
"As proposed by the Workshops on Human Leucocyte Differentiation Antigens (Paris, France, 1982; Boston, USA, I984; Oxford, UK, I986). The anti-D44 and Leu 7 MoAbs did not enter in a cluster definition and were "nonclusterized" (NC). bT-colony cells were depleted of mature T cells after a complement-mediated cytolysis using a cocktail of three MoAbs, CD3, CD4, and CD5, as described in Methods. 'Results are expressed in percentages of positive cells in the cell population analyzed; or in absence of specific positive staining (--) as compared to negative controls (~<1.5%). aNot tested.
116
B: Autran et al. 161, 65, which detect subsets of normal PBL [16]. Nearly half of both the PBand BM-derived TLC cells were activated: 35 % - 5 0 % o f them displayed the class II M H C molecules (DR, DP, DQ) and carried the IL2 receptor. Finally none of them expressed cell-surface antigens o f the killer or natural killer cells, B cells, or myelomonocytic cells. Since about 5% PB- and BM-derived T-colony cells remained regularly negative for the CD3, CD4, and CD8 markers, cell suspensions were enriched in those cells after a complement-mediated cytolysis using the anti-CD3, CD4, and CD5 MoAbs: a 95% cell T-depletion was regularly obtained after treatment o f the T-colonies. Table 1 shows that the residual viable cell~ from both PB- and BM-derived colonies were consistently negative for the following T-cell markers: CD1, CD3, CD4, CD5, CD6, and CD8 as for the T C R c~! 3. These cells were C D 2 + and C D 7 + (Table 1) as further assessed by their sensitivity to the CD2, CD5, CD7 cytolytic cocktail (99.99% mortality) while they were resistant to the CD3, CD4, CD5 cytolytic cocktail Mata not shown) Table 1 shows that these cells were also L F A 1 - and D 4 4 + . About 50c~ of residual cells displayed the HLA-DR, DP, and D Q molecules; 39% of PB* derived immature T cells but only 5% of BM o n g m were Tac posiuve. Finally, none o f these cells expressed non-T-cell lineage markers (CD 16, H N K 1, CD 19. CD1 lb). Figure 1 shows the fluorescence histograms of PB-colony treated cells. Identical results were obtained with BM-colony treated cells I data not shown~. We will further refer to these C D 3 - , C D 4 - , C D S - cells detected in TLC after MoAb complement-dependent cytolysis as phenotypically immature T lymphocytes (ITL).
FIGURE 1 C D 3 - C D 4 - C D 8 - CD2+ CD7+ cells identified after mature T-ceil depletion. Immature T cells were purified as detailed in Methods, They were labeled with a second step fluoresceinared reagent (a) alone preceded by the following primary MoAbs: (b) OKT11--CD2; (c) I21--CD7; (d) T3--CD3; (e) IOT4--CD4; and (f) IOT8--CD8. Fluorescence histograms of PB-derived immature T cells from a representative experiment were obtained after a FACSTAR flow cytometer analysis. The top row shows back~ ground staining with fluoresceinared antibody alone (a), which is indicated in (b)-(f) by ~ dashed line on the histograms. a
/'",,~ d
t/
iI
F'luorescence Intensity
C D 3 - , 4 - , 8 - Autoreactive T Cells
119
100
50
J
i
1O 0
i
i
i
25
•
6 AUTOLOGOUS
PBL ADDED
x 10 3
FIGURE 3 Suppressive effects of mature T cells addition on the proliferative activities of ITL in MLR assays. ITL (1 × 106/ml), purified as described in Methods, were cultured for 5 days against (1 × 106/ml) autologous irradiated PBL in the presence of increasing dilutions of nonirradiated autologous PBL in triplicate wells.
o b t a i n e d by i n c u b a t i o n o f cells with the a n t i - H L A - D r M o A b ( D l - 1 2 ) , w h e r e a s , in contrast, t h e a n t i - H L A class I M o A b ( W 6 - 3 2 ) had no effect on the A - M L R ; (2) a d d i n g a n t i - C D 2 M o A b ( X T 1 ) i n d u c e d a slight i n h i b i t i o n ( 1 5 % ) o f I T L a u t o r e a c t i v i t y ; (3) no i n h i b i t i o n was seen w h e n the a n t i - C D 3 M o A b was a d d e d to
F I G U R E 4 MLR and A-MLR inhibition patterns obtained with a n t i - H L A class I and class II, anti-CD2, anti-CD3, and anti-CD4 MoAbs. ITL, purified as described in Methods, were assayed in auto-MLR alone ([~) or with distinct MoAbs ([3) directed against either the HLA class I (W6-32) and HLA class II (D1-12), the CD2 (XT1), the CD3 (T3), or the CD4 (BL4) antigens. Control responder cells were cultured alone ([]) and with the corresponding MoAbs (D). Results are expressed in percentages of proliferation of the control auto-MLR of ITL (100c~). PERCENTAGES OF
PROLIFERATION
100
CONTROL
W6/32
OI-1Z
×T1
T3
120
B. Autran et aI. the cultures; on the contrary, a slight increase of proliferation of ITL was d e t e c t e d in the A-MLR assay whereas no proliferation was seen when the anti-CD3 MoAb was added on control ITL alone; and (4) finally, the anti-CD4 (BL4) MoAb induced a partial inhibition (5 7 %) of the A-MLR, although it was lower than the inhibition induced by the a n t i - H L A class II MoAb.
I T L D o N o t R e s p o n d to Soluble A n t i g e n s n o r D o T h e y P r o d u c e IL2 in R e s p o n s e to M i t o g e n Since total and T-depleted TLC displayed different MLR reactivities, we examined their ability to develop a secondary response to soluble antigens or m produce IL2 in response to PHA. Non-T-depleted TLC showed a weak proliferative response to streptococcal antigens or to PPD, whereas control P B L of the original donor were highly responsive, but no antigen-specific proliferative re-sponses could be detected in ITL although antigen-presenting cells were added to the cultures (data not shown), In addition, after a 48-hr PHA-stimulation, the culture supernatants of the TLC-untreated cells were able to give a consistent proliferation o f an IL2-dependent T-cell line, while ITL did not produce any detectable amount of IL2 activity, as shown in Table 3. Those data indicate a lack o f antigen specific precursor cells in ITL and suggest the inability of ITL to secrete IL2 in response to PHA. C D 3 + a n d C D 4 + T Cells Can Be D e r i v e d f r o m I T L T o determine whether ITL activation gave rise to mature T cells, the phenotype o f the C D 3 - , 4 - , and 8 - was reevaluated after co-culture with irradiated PBL. Although ITL were devoid o f detectable residual mature T cells after T depletion (Figure 1), they were cultured in limiting dilution conditions (3 cells per well) in order to avoid a contamination. The phenotype of individual wells could be determined after expansion in the presence of either autologous or allogeneic PBL and r-IL2. Each culture o f ITL originated mature T cells whether they were stimulated by autologous or allogeneic PBL. A representative phenotype is given TABLE 3
Lack of IL2 production of immature T lymphocytes in response to P H A
E f f e c t o r cells ~ PBL
Untreated T - c o l o n y cells T-depleted b T-colonycells
Dilution of supernatants
Proliferative responses
1/ 3 1/ 9 1/27 1/ ~ 1/ 9 1/27 1/ 3 1/ 9 1/27
40 9 9 4 ( 9 8 1 ) ~ 21 8 0 6 ~ 1 ~ 0 6 ) 13 7 4 9 ( 1 1 7 2 ) 26 290 ( 9 5 3 ) 7 24(3 ( 1 3 3 ) 3 829(176) 9961 (212) 1 1581 141) 1 0 7 8 ( 229"1 1 654 ~ 147)
None
~Effector cells were preincubated with PHA-P for a 48-hr culture. Supernatants were then collected and tested for their IL2 activity as described in Methods. AT-cell depletion was performed with the CD3, 4, 5 cytolytic MoAb cocktail and rabbit complement as described m Methods. ~Results expressed in mean cpm ~± SD) of duplicate cultures.
C D 3 - , 4 - , 8 - Autoreactive T Cells TABLE 2
117
Proliferative responses in autologous and allogeneic MLR of the PB- and BM-derived T-colony cells Stimulating cells h
Effector cells a PB T c o l o n y cells Untreated M o A b treated C o m p l e m e n t treated T-depleted B M T c o l o n y cells Untreated T-depleted
None
Autologous PBL
Allogeneic PBL
958' (137) 1029 (168) 868 (155) 1319 (221)
1511 (189) 1818 (124) 1809 (201) 15746 (2224)
2923 v (3409} 15753 (1182) 21050 ~2857) 23068 (32711
2140 (134) 2190 (287)
3145 (581) 14878 (1570)
19090 (2831} 6944 (1246)
~PB- and BM-derived T colony cells from distinct donors were tested in auto- and allo-MLR, either before or after T-cell depletion with anti-CD3, CD4, CD5 MoAbs and rabbit complement. Control cells were treated with MoAb or complement alone.
bA total 1 x 105 T-colony cells were mixed with 1 x 105 irradiated (25 Gy) autologous or allogeneic PBL and incubated for 5 days in MLR cultures as described in Methods. 'Results are expressed in mean cpm ± SD of triplicate cultures from a representative experiment.
CD3-,
4 - , 8 - TLC Cells Proliferate in A u t o l o g o u s MLR Since PB- and BM-derived TLC were not only composed of mature T cells but also contained ITL, the functional properties of both cell populations were investigated. Firstly, their capacities to respond to an allogeneic or an autologous stimulus were studied. Table 2 shows the results of MLR assays. Total colony cells of both PB and BM origins were able to proliferate against allogeneic irradiated PBL but were unresponsive to equal amounts of autologous stimulating PBL. In contrast, C D 3 - , C D 4 - , C D 8 - cells obtained from both PB and BM TLC cells were markedly activated and proliferate when cultured with autologous or allogeneic irradiated PBL in the same conditions. The same pattern of responses could be obtained with P B - and B M - I T L derived from at least five donor origins (data not shown). N o autoreactivity could be detected when TLC cells were treated with MoAbs or rabbit complement alone. Contrasting with these proliferative responses in auto- and allo-MLR, ITL were devoid of specific or nonspecific cytotoxic activities. Indeed, these cells were totally unable to lyse the NK-target cell line K562; in addition, they could not lyse autologous or allogeneic PHA-blasts neither in a direct cytotoxicity assay nor after co-culture (data not shown). This lack of cytotoxicity was in accordance with the cell-surface phenotype of ITL, since they were C D 8 - , H N K I - , and CD16-.
B. Autran et al.
118
NESPONDEN
loo
~,
7o
~
37
..3 e
12
2,.a
5,0
DOSE
190
,
150
.
2QO
5o
i
~
2o
FIGURE 2 Limiting dilution analysis of the autoreactive cells in TLC celis. Replicate cultures (144 per group) containing 1.5 to 200 cells per well were stimulated with or without autologous irradiated PBL. 1L2 (50/~/ml) was added at day 2. Growth was scored at day 14 poststimulation. F0 = 1/173 between the origin of the plot and the point o~ inflexion.
T h e A u t o r e a c t i v i t y o f I T L Is A c t i v e l y S u p p r e s s e d b y M a t u r e T Cells The above data suggested that the autoreactivity of ITL was not detectable when these cells were mixed to the total TLC cell population. In order to estimate the frequency o f autoreactive cells in TLC, we performed limiting dilution analysis of the proliferating TLC cells against autologous PBL. A typical partition analysis of autoreactive cells in this model is shown on Figure 2. with a plot suggesting an active suppression of the p h e n o m e n o n [17]. At low cell concentrations, the fraction (f0) o f nonresponding wells decreased to increase again at higher cell concentration, with an f0 returning to the origin at 200 cells per well. Those data suggest that (1) a subset of autoreactive cells do exist among the total TLC cells with a frequency estimated to 1/173. and (2) the TLC cells could also contain a subset o f suppressive cells, with a lower frequency than the above one. which negatively regulate the autoreactive cells. W e then tried to test the inhibiting effects of the addition or- autologous mature T cells on the autoreactivity of isolated ITL. As expected, the addition of nonirradiated autologous PBL induced a 9 0 % inhibition of the proliferation m auto-MLR (Figure 3) with a d o s e - r e s p o n s e curve, confirming the active suppression o f the autoreactivity of ITL by mature T cells. Effect of Anti-CD2, CD3, CD4, Anti-MHC the ITL Autoreactivity
class I a n d class I I M o A b s on
Though ITL differ from c o m m o n autoreactive mature T cells by the absence of surface expression o f the CD3, CD4, C D 8 antigens, we investigated the inhibiting effects of monoclonal antibodies directed against the H L A class I and class II molecules and against the CD2, CD3, and CD4 antigens. Four patterns of reactivities were obtained w h e n these MoAbs were added to the A-MLR cultures (Figure 4)" (1) a 9 0 % inhibition of the proliferation of ITL in A-MLR could be
C D 3 - , 4 - , 8 - Autoreactive T Cells
121
a
b
c
- .7"~'~'~'~L'Jr
d
~/,~,~
....... I
~' ""~'ir' I
; ,3~,,
e
~,~'r/~,, i
FLUORESCENCE INTENSITY
FIGURE 5 Immunofluorescence analysis of ITL after long-term culture. After purification, ITL, whose phenotype is given in Figure 1, were cultured in limiting dilutions (5 cells per well) with either autologous or allogeneic 1 × 104 irradiated PBL and r-IL2. After a 45-day expansion, individual wells could be tested for their phenotype. A representative analysis of ITL cultured with allogeneic PBL is shown, where cells were labeled either with the FITC-second step reagent alone (a) or with the anti-CD2 (b), CD3 (c), TCRa//3 (WT31) (d), CD4 (e), CD8 (f) MoAbs.
in Figure 5, where ITL, whose initial phenotype was shown in Figure 1, differentiate in long-term cultures into C D 3 + , C D 4 + mature T cells. DISCUSSION Previous reports demonstrated that T colonies may be driven from putative Tcell progenitors considered as prethymic ceils [ 10-13]. After T-cell depletion by rosetting, the ability of E - TLC cells to generate secondary mature T colonies and the clonogenic nature of this subset has been reported [ 11,12]. The T-colony assay is then an attractive model for the in vitro study of the T-cell differentiation; however, the phenotype and immune functions of the T-colony-forming cells have not been fully investigated. The data presented here gave an extended phenotypical analysis of PB- and BM-derived TLC cells. Repetitive cell-surface markers analysis of the TLC cells showed that about 5% of C D 3 - , C D 4 - , and C D S - ceils could be isolated after a cytotoxic treatment of TLC cells with anti-Tcell MoAbs. The CD3, 4, 5 resistant cells derived from both the PB and BM T colonies displayed the same absence of surface expression of the T-cell receptor and of the CD3, CD4, CD5, CD6, and CD8 molecules. In contrast, both the CD2 and CD7 antigens were consistently expressed on those cells designed as immature T cells (ITL). Indeed, the expression of the CD7/gp40 molecule on immature T cells, before the other T-differentiation antigens, has been demonstrated on leukemic T-cells and in normal thymus [ 1 8 - 2 0 ] . T h e ITL expressed in the same extend the 28-kd molecule characterized by the D44 MoAb, which is in accordance with previous data showing its expression on the BM-myeloid precursor cells [15]. In contrast, the C D 3 - , 4 - , 8 - cells from both PB- and BM-derived TLC were negative for the LFA 1 antigen that has been shown to be absent on myelo-monocytic precursor cells [21].
122
B. Autran et al. Most o f the C D 3 - , 4 - , 8 - ITL isolated after an 8-day culture with PHALCM displayed surface antigens of activated mature T cells, including the M H C class II molecules and the CD25 antigen, although the latter could be induced by the primary PHA-IL2 stimulus incorporated in soft agar cultures. Such a dissociation between the IL2 receptor and the TCR expression has already been mentioned in vivo on immature Thy 1+, L 3 T 4 - , L y t 2 - murine thymocytes [22,23]. Together, these data suggest that the C D 3 - , 4 , 8 - T cells isolated from the PB and BM T-colonies display phenotypical features of immature T cells. Mature autoreactive cells, studied in both human and murine models, have been shown to recognize self-antigens through a conventional CD3-1inked T-cell receptor [ 2 4 - 2 6 ] . Several reports brought data on the origin of autoreactive T cells in the mouse model; however, little is known about their human counterpart [ 2 - 6 ] . In early experiments, it has been suggested that the recognition o f selfantigens by neonatal murine thymocytes was induced by differentiation antigens of mature spleen cells [2]. Anti-self-MHC cytotoxic T cells have been shown t~ differentiate from T h y l - normal murine BM precursors [3] and were enriched in the same BM-BSA fraction as the prethymic cells [4]. Recently, concanavalin A-reactive, dull T h y 1 + , L y t l - , L 3 T 4 - , L y t 2 - precursors o f autoreactive T cells have been isolated by complement mediated cytolysis m murine spleen [5} Recent data suggest that the peripheral C D 4 - , C D 8 - cells could correspond to post- or extrathymic T-cell precursors that have escaped the thymic selection and are able to differentiate in the periphery [6]. Our data show similar results in the human model: the T depletion ofT-colony cells results m an enrichment of a C D 3 - . 4 - , 8 - subset able to proliferate in auto-MLR. This p h e n o m e n o n seemed to be actively suppressed since the C D 3 4 - , 8 - cells self-reactivity disappeared with the readdition of mature autologous PBL, a result that is m accordance with the T lymphocyte control o f autoreactivity previously described [27,28]. The ITL were also alloreactive, in accordance with the murine model of autoreactive L3T4-, Lyt2-precursor cells, where both reactivities persisted even at the clonal level [5]. The proliferation of ITL raised questions about the membrane antigens involved in the activating signals. The M H C class II but not class I molecules co uld play a role. The involvement of the C D 3 - T C R complex in such a process could not be demonstrated since no inhibition o f the auto-MLR was induced by the anti-CD3 MoAbs, suggesting that at least the initial stimulus was TCR independent. Nevertheless, the anti-CD3 MoAb induced a slight proliferation of ITL only when stimulated by autologous PBL, in accordance with the acquisition of the CD3 antigen during cultures as shown by phenotypic analysis of responding cells after culture. Interestingly, the partial inhibition of the anti-self proliferation by an anti-CD4 MoAb suggested that this antigen, initially absent from the surface of ITL, could play a role at a further step in the proliferation process. The anti-CD2 M o A b did not block the auto-reactivity of ITL. Similarly, the proliferation o f double-negative thymocytes against autologous epithelial cells is not inhibited by anti-CD2 MoAb [7, 8]. Together, these data suggest that a direct cell contact through LFA-3 is not solely responsible for activation of C D 3 - , C D 4 - . C D 8 - cells. Finally, whether the C D 3 - , 4 - , 8 - immature self-reactive T lymphocytes isolated from both the PB and BM compartments correspond to early developmental stages in the T-cell lineage or have escaped the thymic selection remains unclear. Once activated, ITL could acquire mature T-cell phenotypes. Experiments in progress should relate more precisely the relationship between activation events and maturation process in human precursor cells, and they might be useful to analyze the implication of these precursor cells in autoimmune or immunodeficiency pathological models.
C D 3 - , 4 - , 8 - Autoreactive T Cells
123
REFERENCES 1. Von Boehmer T, Adams PB: Syngeneic mixed lymphocyte reaction between thymocytes and peripheral lymphoid cells in mice: strain specificity and nature of target cells. J Immunol 110:376, 1973. 2. Howe ML, Goldstein AL, BattistoJR: Isogeneic lymphocyte interaction: Recognition of self antigens by cells of the neonatal thymus. Proc Natl Acad Sci USA 67:613, 1970. 3. Benveniste P, Miller RG: Differentiation from precursors in normal bone marrow of spontaneously cytotoxic cells showing anti-self. J Immunol 136:4399, 1986. 4. Hurme M, Shivola M: In vivo activation of autoreactive cytotoxic T lymphocytes after bone marrow transplantation: Evidence for the prethymic specificity of T cells. J Immunol 135:1108, 1985. 5. De Talance A, Regnier D, Spinella S, Morisset J, Seman M: Characterization of Thy-l÷, Lyt-, L3T4- precursors in the spleen of normal mice. J Immunol 137:1101, 1986. 6. Morisset J, Trannoy E, De Talance A, Spinella S, Debre P, Godet P, Seman M: Genetic and strain distribution of concanavalin A reactive Lyt-2-, L3T4- peripheral precursors of autoreactive T cells. Eur J Immunol, in press. 7. Denning SM, Tuck DT, Singer KH, Haynes BF: Human thymic epithelial cells function as accessory cells for autologous mature thymocyte activation. J Immunol 138:680, 1987. 8. Denning SM, KurtzbergJ, Le PT, Tuck DT, Singer KH, Haynes BF: Human thymic epithelial cells directly induce activation of autologous immature thymocytes. Proc Natl Acad Sci USA 85:3125, 1988. 9. Scollay R, Bartlett P, Shortman K: T cell development in the adult murine thymus: Changes in the expression of the surface antigens Ly2, L3T4, and B2A2 during development from early precursor cells emigrants. Immunol Rev 82:79, 1984. 10. Triebel F, Robinson WA, Hayward AR, Goube de la Forest P: Existence of a pool of T-lymphocyte colony forming cells (T-CFC) in human bone marrow and their place in the differentiation of the T-cell lineage. Blood 58:911, 1981. 11. Mossalayi MD, Goube de la Forest P, Guilhot F, Lemaire M, Larroque V, Tanzer J: Self-renewal and maturation of immature T-cell progenitors within PHA-induced agar T-cell colonies. IntJ Cell Clin 2:34, 1984. 12. Mossalayi MD, Ahmad M, Tanzer J, Janossy G, Goube de la Forest P: Studies on human prothymocytes by means of in agar T-cell colony assay. Exp Hematol 14:351, 1986. 13. Mossalayi MD, Lecron JC, Goube de Laforest P, Janossy G, Debr~ P, Tanzer J: Characterization of prothymocytes with cloning capacity in human bone marrow. Blood 71:1281, 1988. 14. Triebel F, Missenard-Leblond V, Couty MC, Charron D, Debrd P: Genetic control of antigen specific T cell clones: Effect of monoclonal anti-DR antibodies. J Immunol 132:1773, 1984. 15. Bernard A, Gay-Bellile V, Amiot M, Caillou B, Charbord P, Boumsell L: A novel human leukocyte differentiation antigen. Monoclonal antibody D44 defines a 28-Kd molecule present on immature hematological cells and a subpopulation of mature T cells. J Immunol 132:2338, 1984. 16. Boylston AW, Borst J, Yssel H, Blanchard D, Spits H, De Vries JE: Properties of a panel of monoclonal antibodies which react with the human T cell antigen receptor on the leukemic line HPB-ALL and a subset of normal peripheral blood T lymphocytes. J Immunol 137:741, 1986.
124
B. Autran et al, 17. Lefkovits I, Waldmann H: Limiting Dilution Analysis of Cells of the Immune System. Cambridge, Cambridge University Press, 1979. 18. Vodinelich L, Tax W, Bai Y, Pegram S, Capel P, Greaves MF: A monoclonal antibody (WT1) for detecting leukemias o f T cell precursors. Blood 62:1108, 1983. 19. Sutherland DR, Rudd CE, Greaves MF: Isolation and characterization of a human T lymphocyte associated glycoprotein (gp40). J Immunol 133:327, 1984. 20. Van Dongen JIM, Quertermons T, Bartram CR, Gold DP, Wolvers-Tettero IL Comans-Bitter WM, Hooijkaas H, Adriaansen HJ, De Klein A, Raghavachar A, Ganser A, Dubly D, Seidman JG, Van Den Elgen P, Terhorst C: T cell receptor-CD3 complex during early T cell differentiation. J Immunol 138:1260, 1987. 21. Campana D, Sheridan B, Tidman N, Moffbrand AV, Janossy G: Human leukocyte function associated antigens on lympho-hemopoietic precursor cells. Eur J Irnmunol 16:537, 1986. 22. Raulet DH: Expression and function of interteukin-2 receptors on immature rhymocytes. Nature 314:101, 1985. 23. Shimonkevitz RP, Husmann LA, Bevan MJ, Crispe IN: Transient expression of IL-2 receptor precedes the differentiation of immature thymocytes. Nature 329:157, 1987. 24. Altman A, Katz DH: Existence o f t cells manifesting self-reactivity indistinguishable from alloreactivity. J Immunol 125:1536, 1980. 25. Bensussan A, Meuer SC, Schlossmann SF, Reinherz El: Delineation of an immunoregulatory amplifier population recognizing autologous Ia molecules: Analysis with human T cell clones. J Exp Med 159:559, 1984. 26. Tilkin AF, Michon J, Kayibanda M, Henin Y, Sterkers G, Betuel H, Levy JP! Auto-~ reactive T cell clones of MHC class II specificities are produced during responses against foreign antigens in man. J lmmunol 137:674, 1987. 27. Rosenkrantz K, Dupont B, Flomenberg N: Generation and regulation of autocytotoxicity in mixed lymphocyte cultures: Evidence for active suppression of autocytotoxic cells. Proc Natl Acad Sci USA 82:4508, 1985. 28. Deusch K, Moebius U, Meyer Z, Buschenfelde KH, Meuer SC: T lymphocyte con-. trol of autoreactivity: Analysis with human T cell clones at limiting dilution culture. EurJ Immunol 16:1433, 1986.