CELLULAR IMMUNOLOGY 113,329-340 (1988)

Different Signals for Stimulation of Proliferation and Lymphokine Secretion by a CD3+ WT31- Cloned Cytotoxic Lymphocyte’

GKAHAM P. PAWELEC, HANS-J~~RG BUHRING, FRIEDRICH W. BUSCH, FRANK KALTHOFF, AND PETER WERNET~

Immunology Laboratory, Medizinische Klinik, D-7400 Tiibingen, Federal Republic of Germany

Received September 8, 1987; accepted December 15, 1987

CD3+ WT3 I T cells were sorted from peripheral blood ofa normal healthy donor by a FACS IV and cloned by limiting dilution in the presence of a phorbol ester (tetradecanoyl phorbol acetate, TPA), calcium ionophore (ionomycin, IO), interleukin-2 (IL-2) allogeneic cells, and phytohemagghttinin (PHA). One of the derived clones, 290-2, was investigated in detail. 290- 2 mediated strong natural killer (NK) but not lymphokine-activated killer (LAK) activity. It proliferated in the presence of IL-2 but not IL-4. It carried the surface phenotype CD3+ WT3 l- CD4”‘e”k+ CD8-, CD16-, and Leu 19+. Expression of CD4 was heterogeneous within the clone, since two of three subclones were also CD4”eak+ but one was CD4-. NK activity was blocked by monoclonal antibody (moAb) to CD1 la (LFAl), but not by monoclonal antibody (moAb) to either CD3 or CD4. Northern blotting revealed T-cell receptor (TCR-7) but not (Y- or full-length P-chain mRNA. 290-2 proliferated autonomously when stimulated with a combination of TPA + IO, with PHA or CD3 moAb and autologous B-cell lines (B-LCL) (and this was inhibited by an anti-IL-2 receptor moAb), but not to allogeneic B-LCL or any of the other stimulating agents alone. Unexpectedly, the TPA + IO stimulus which resulted in maximal proliferative responses did not trigger interferon-y or granulocyte/macrophage colony-stimulating factor production, although both lymphokines were secreted in the presence of B-LCL + TPA + IO. Proliferative responses were not enhanced by the presence of B-LCL. Thus, activation signals sufficient for autocrine proliferative responses were insufficient for secretion of other lymphokines. Such clones will provide valuable reagents for investigating the biology of the TCR--y+ T cell. o 1988

Academic Press, Inc.

INTRODUCTION

The CD3 complex is coordinately expressed at the T-cell surface with the hetero- dimeric antigen receptor (TCR)3 (1). The majority of mature T cells express TCR composed of an o( chain and a p chain, well characterized at the protein and molecular genetic levels (1). The monoclonal antibody (moAb) WT3 1 reacts with a public deter-

’ Supported by the Deutsche Forschungsgemeinschaft, SFB 120. * Present address: Blood Bank, University of Diisseldorf, Federal Republic of Germany. 3 Abbreviations used: B-LCL, B-lymphoblastoid cell line; Con A, concanavalin A; FCS, fetal calf serum;

FITC, fluroscein isothiocyanate; IFN-7, interferon-y; IL-2, -4 (R), interleukin-2, -4 (receptor); IO, iono- mycin; LAK, lymphokine-activated killer; LU, lytic unit; MoAb, monoclonal antibody; NK, natural killer; PBMC, peripheral blood mononuclear cells; PD, population doubling; PE, phycoerythrin; PHA, phytohem- agghttinin; TCR, T-cell receptor; TPA, tetradecanoyl phorbol- 13-acetate; [‘H]TdR, tritiated thymidine.

329

330 PAWELEC ET AL.

minant of the CY//~ TCR (2) and fails to bind T cells which do not express this receptor. It has been shown that a small percentage of peripheral CD3’ WT3 l- T cells can express a non-a//3 TCR containing the y-chain gene product (3, 4) which may be present as a y/6 heterodimer (3) or possibly a y/y homodimer (5). Human T cells of this phenotype have been isolated and propagated with interleukin-2 (IL-2) from immunodeficient patients (3) from thymocytes (6) and fetal lymphocytes (7) and, with more difficulty (3), from normal adult peripheral blood mononuclear cells (PBMC). Such cell lines were found to be CD3+ CD4- CD8- or rarely CD8+, a pheno- type shared with some extant natural killer (NK)-active T-cell clones which were later shown also to be WT3 1 - (5,8). We show here by direct sorting and cloning of CD3+ WT3 1~ cells that, after stimulation with phorbol ester (tetradecanoyl phorbol acetate, TPA) and calcium ionophore (ionomycin, IO), T cells expressing TCR y-chain, but not a-chain, mRNA can be obtained from normal adult donors. One such clone was found to require different signals for proliferation compared to lymphokine secretion. Maximal autocrine proliferation was stimulated by interleukin-2 or TPA + 10, but secretion of interferon-y and colony-stimulating factors required the additional pres- ence of a B-cell line.

MATERIALS AND METHODS

CelZsorting. Double staining of normal PBMC was carried out by labeling first with moAb WT31 (a kind gift of Dr. W. Tax, Nijmegen) and the F(ab’h fragment of an FITC-conjugated goat anti-mouse antiserum. The latter had been absorbed with hu- man, bovine, and horse serum proteins (Dianova, Hamburg, FRG). After washing, cells were incubated for 15 min with mouse IgG to block free binding sites of the second-step antibody. In the final step cells were stained with PE-conjugated CD3 moAb of the same IgG subclass as WT3 1 (Leu 4, Becton-Dickinson). Prior to sorting CD3+ WT3 I- cells with a FACS IV, a “lymphocyte” window was set by dual-scatter parameters. The green fluorescence of FITC-labeled cells was measured through a 530-nm bandpass filter (Becton-Dickinson) and the yellow fluorescence of the PE- labeled antibodies through a 570-nm bandpass filter (Oriel). The yellow and green components of the emitted light were separated by a beamsplitter with a dichroic (560-nm) filter. A differential log-amplifier was used to correct for overlap of the green fluorescence emission spectrum of FITC into the yellow fluorescence emission spectrum of PE. The amplifier was adjusted so that the signal from cells stained only with FITC was orthogonal to the yellow fluorescence axis of the two-parameter yellow and green fluorescence display. For cell sorting the sort window was set on the WT3 1~ Leu 4+ population. One-droplet sorting without “full deflection envelope (coinci- dence)” was performed at a 20-kHz droplet formation rate (diameter of nozzle orifice 80 pm). Cells were sorted into sterile tubes containing 20% FCS in Hanks’ buffer.

CZoning and culture. Sorted CD3+ WT3 l- T cells were distributed at limiting dilu- tion into 1 -mm-diameter culture wells (Greiner, Niirtingen, FRG) in the presence of irradiated PBMC pooled from 20 random donors. Sixty wells were seeded with an average of 45 cells/well, 60 with 4.5 cells/well, and 300 with 0.45 cells/well. Culture medium was RPM1 1640 buffered with 25 mM Hepes and supplemented with 16% heat-inactivated pooled human serum, 20 U/ml purified natural IL-2 (Lymphocult T-HP, Biotest, Frankfurt, FRG), 10 rig/ml tetradecanoyl phorbol-13-acetate (Sigma, Munich, FRG), and 400 rig/ml ionomycin (Calbiochem, San Diego, CA), 1% phyto-

CD3+ WT3 l- CLONE 331

hemagglutinin (PHA, Gibco-M, Scotland) plus antibiotics. Positive wells were identi- fied and propagated as described previously for alloactivated T cells (9), in medium supplemented as above but without TPA and 10.

Phenotyping of T-cell clones. Cells were analyzed on a FACS IV. Fluorescence was excited at 488 nm with a 5-W argon laser. Power was held constant at 200 mW. Fluorescence of FITC was measured through a 530-nm bandpass filter (Becton-Dick- inson). Dead cells were gated out by forward-scatter characteristics. Fluorescence in- tensity of 10,000 cells/probe was measured on a logarithmic scale. To indicate back- ground staining, histograms of cells labeled with test antibodies were superimposed on those labeled with the nonbinding variant of the anti-MHC class I moAb W6/ 32.HK. Antibodies were obtained from Ortho Pharmaceutical (OKT4), Becton- Dickinson (Leu series), or were gifts: WT31 and WT32 (from W. Tax, Nijmegen, Holland), SPV-L 11 (CD 11 a, from H. Spits, Dardilly, France), and B7/2 1 (anti-HLA- DP, from I. Trowbridge, San Diego, CA). Tubingen moAb were TU 14 (CD7), TU22 (HLA-DC!), TU30 (pan T, unclustered), TU33 (CD6), TU34 (HLA-DR), TU39 (broad class II), TU69 (CD25), TU7 1 (CD5), and TU 102 (CD8).

Assaysjbr proliferation. Cloned cells were distributed at 1 X 1 04/well into U-form microtiter plates in RPM1 + 10% human serum. They were cocultured with 2.5 X lo4 irradiated cells from B-lymphoblastoid cell lines (B-LCL), and titrated amounts of IL-2 (20, 10,5 U/ml), TPA ( 100, 10, 1 rig/ml), 10 (800,400,200 rig/ml), PHA (l%, 0.2%), concanavalin A (Con A, 10, 1 pg/ml) and CD3 moAb WT32 (10, l,O. 1 pg/ ml), or combinations of these, for different periods (24,48,72 hr). IL-2 was obtained from Biotest, TPA from Sigma, and 10 from Calbiochem. PHA and Con A were from Gibco. Recombinant IL-4 was the kind gift of Dr. Banchereau (Unicet, Dardilly, France). Proliferation was assessed by measuring the incorporation of tritiated thymi- dine ([3H:ITdR, sp act 185 GBq/pmol, Amersham-Buchler, Braunschweig, FRG), 37 kBq of which was added to each culture well 18 hr before termination of cultures and beta-scintillation spectroscopy.

Lymphokine secretion. Cloned cells were incubated for different periods in me- dium containing 10 U/ml IL-2 and various stimulating agents. Supernatants were assayed for II++ content using a radioimmunoassay kit (Cell-Tech, Slough, Bucks, England) and for granulocyte/macrophage colony-stimulating factor (GM-CSF) us- ing a bioassay as follows: Briefly, human mononuclear bone marrow cells were plated in 0.9% methylcellulose supplemented with 20% prescreened FCS and human pla- centa-conditioned medium (control source of GM-CSF) or the supernatants to be tested. Colonies were counted on Day 14 in triplicate. Morphological comparison of the consti.tution of colonies supported by recombinant GM-CSF in comparison to T- cell clone supernatants suggested that the latter contained only GM-CSF (G. P. Pa- welec et wl., manuscript submitted for publication).

Assays fir cytotoxicity. Target cells were taken from cell lines maintained my- coplasma.-free and frozen in large batches for repeated use in sequential experiments. The targets were K562, Daudi, Jurkat, IL-2-dependent normal T cells, or B-LCL derived from the donor of the clones and from allogeneic donors. The origins of the cell lines, labeling with sodium “Cr-labeled chromate, and the performance of the assay have been described previously (10). Briefly, 4 X lo4 radiolabeled target cells were cocultured with titrated amounts of effector cells for 4 hr, radioactivity in the supematants was measured by gamma-spectroscopy, and the percentage specific iso- tope release was calculated. Data were expressed as the percentage specific isotope

332 PAWELEC ET AL.

release (=% lysis) or as lytic units (LU)/ 1 O7 effector cells, where one LU was defined as the number of effector cells required to achieve 25% specific 5’Cr release from 4 X lo4 target cells. Lymphokine-activated killer cells were prepared by culturing PBMC at 1 X 106/ml with RPM1 1640 + 10% human serum supplemented with 100 U/ml IL-2 for 3 days before use as effector cells in cytotoxicity assays.

Northern blotting for TCR message. Total cellular RNA was isolated from frozen cells lysed in 4 Mguanidium isothiocyanate and 0.5% sodium sarcosine by centrifu- gation through a 5.7 MCsCl cushion at 120,OOOg for 28 hr. Approximately 5 pg RNA from each cell sample was run through 1.2% agarose containing formaldehyde and transferred to Hybond-N filters (Amersham) in 10 X SSC. Hybridization was carried out at 62°C. Final washings of the filters were done in 0.1 X SSC and 0.1% SDS, and autoradiographed for 24-36 hr at -80°C using intensifying screens (DuPont). The cDNA probe specific for the TCR p chain (pCPREX) was a kind gift of Dr. H. D. Royer (Heidelberg, FRG) and was described previously ( 11). cDNA probes for the TCR a! (pY 14) (12) and y (pTy) chains were kindly provided by Dr. T. W. Mak (Toronto, Canada). Inserts of all probes were labeled to a specific activity of ca. 1 X lo9 cpm/pg using random oligonucleotide primers (Pharmacia) and the Klenow fragment of DNA polymerase I (Boehringer).

RESULTS

Derivation of CD3’ WT31- Clones

PBMC from a normal donor were labeled with Leu 4-PE and WT3 1 -FlTC, fol- lowed by sorting on a FACS IV. The purity of the sorted WT3 I- population was >98% on reanalysis. These cells were distributed into culture wells at limiting dilution in the presence of irradiated pooled PBMC, 20 U/ml IL-2, and 1% PHA. Under these conditions, no clones were obtained, and the bulk cultures also failed to proliferate. However, when this experiment was repeated in medium supplemented with 10 ng/ ml TPA and 400 rig/ml 10, some clones were obtained, although the cloning efficiency was still only 19%. Of 17 clones which could be propagated to > 1 O7 cells, three were found to be CD3+ WT31-. The remainder were WT31f, suggesting that the latter can be much more easily propagated under these conditions.

Of the WT3 I- clones, only one, designated 290-2, could be extensively propagated (>109 cells). However, this clone also had only a relatively limited life span of ca. 35-40 population doublings (PD). Even a number of modifications to the culture conditions, including “restimulation” with TPA and 10, did not result in an extension of this life span. Subcloning, again by limiting dilution (with an efficiency of 58%), did result in the isolation of lines which grew for ca. 25 PD after the second limiting dilution (i.e., for a total of 48-50 PD), but “immortalized” lines were not obtained. The experiments reported in the following were performed with the original clone 290-2, except the Northern blots, which were done with a pool of subclones.

Surface Phenotype of Clone 290-2

Clone 290-2 strongly expressed CD3 but was completely negative with moAb WT31 (Fig. 1). It expressed CD2, CD5, CD6, and CD7 mature T-cell markers and lacked CDla thymocyte antigen. It was CDs- but, interestingly, weakly, but clearly CD4’ in repeated analyses. This was confirmed in analyses of subclones, where two

CD3+ WT31- CLONE 333

of three showed a CD4 fluorescence histogram identical to that of clone 290-2, al- though a third subclone was CD4- (data not shown). Clone 290-2 lacked Leu 11 (Fc receptor for IgG) but a subpopulation of cells expressed Leu 19 (NKHl). This clone strongly expressed HLA class I as well as all three major class II antigens, HLADR, - DQ, and -DP at approximately equal fluorescence intensities (Fig. 1). Finally, TU30, a pan-T-cell marker of uncertain specificity, was also strongly positive.

Clone 290-2 Expresses the TCR-7 but Not a-Chain Gene

Messenger RNA was isolated from pooled subclones of 290-2 and from other, allo- reactive clones, and Northern blotting was performed. Figure 2 shows that clone 290- 2, but not a representative alloreactive clone, 277G 15, expressed y-chain mRNA. In contrast, clone 277C-15, but not 290-2, expressed a-chain message. Accordingly, clone 277C-15 also expressed a full-length 1.3-kb P-chain message, whereas the y- chain-expressing clone 290-2 displayed only the shorter 1 .O-kb mRNA known to lack a V-region gene segment and thought to be derived from cryptic promoters upstream of rearranged DJ recombinations (13).

Cytotoxic Activity of Clone 290-2

Clone 290-2 was tested in standard 4-hr “Cr-release assays using NK-susceptible K562 targets and LAK-, but not NK-susceptible Daudi cell targets. Cloned cells were strongly cytotoxic toward K562 in the presence or absence of 1% PHA, but killed Daudi targets only in the presence of the lectin (data not shown). The “NK-like,” but not LAK activity of clone 290-2 was further demonstrated in experiments using a wider range of target cells, as illustrated in Table 1. Unlike LAK cells, clone 290-2 did not lyse autologous B-LCL to any degree.

Addition of moAb to the cytotoxicity assay showed that only CD1 la (LFAl) sig- nificantly reduced lysis by clone 290-2, whereas CD3 or CD4 did not do so, despite their expression on the cloned cell surface (Table 2). In contrast, however, note that clone 28Z!-16, a representative CD3+ WT3 I+ NK-active cell, was blocked by CD3 moAb.

Proliferative Capacity of Clone 290-2

Clone 290-2 was derived by stimulating sorted cells with alloantigen, lectin, IL-2, phorbol ester, and calcium ionophore, and propagating them with alloantigen, lectin, and IL-2. It was therefore investigated whether any of these agents could stimulate autocrine clonal proliferation of the cells in the absence of exogenous IL-2. In these experiments, all stimulating agents were titrated and kinetic studies were carried out. For clarity of presentation, the results illustrated in Table 3 have been selected as maximal iattainable values with that particular stimulator or combination of stimula- tors. Thus, [3H]TdR incorporation on stimulation with PHA, TPA, or 10 alone was not higher than that measured after culture in medium alone. Soluble CD3 moAb WT32 also failed to stimulate. [3H]TdR incorporation in the presence of autologous or allogeneic B-LCL or PBMC was slightly increased, possibly due to a “feeder” effect. Marginally positive responses were obtained by combining TPA, but not 10, with autologous B-LCL, but more strongly positive responses were observed when CD3 moAb, TPA, and autologous B-LCL were combined, suggesting that the y-TCRICD3

PAWELEC ET AL. 334

HLA Class I

CD7

Leull

CD3

Leul9

CD5

HLA-DR HLA-DP

WT31

HLA-DQ

Tii39

Tii30

CD3+ WT31- CLONE 335

alpha bet a gamma

- 28S-

-lSS-

28s -

18 S -

FIG. 2. Expression of TCR gene messages in clone 290-2. Northern blotting demonstrated the presence of y-chain but not a-chain message in clone 290-2. A representative WT3 l+ clone expressing an 01 but not a y message is shown for comparison. This clone also expressed a full-length 1.3-kb mRNA whereas clone 290-2 expressed only an incomplete I .O-kb message lacking V-region segments.

complex is, to some extent, competent to transduce cell-activating signals. However, stronger stimulation was seen with PHA + autologous B-LCL, and the most strongly positive responses were observed when TCR/CD3 signaling was bypassed by stimula- tion with a combination of TPA and 10, suggesting the possibility that signal transduc- tion by unidentified structures other than CD3 may also play a part in activating these cells.

The only single agent to which clone 290-2 readily responded by proliferation was IL-2, including purified recombinant IL-2 (not shown), again showing that WT3 l- T cells use IL-2 as a growth factor (Table 3). Proliferation supported by IL-2 was also significantly inhibited by moAb Tfi69, as was autocrine proliferation stimulated by the combinations of agents described above. Furthermore, recombinant IL-4 failed to support proliferation of clone 290-2 cells even in the presence of autologous B-LCL (Table 3) although the concentration used was known to be optimal for supporting proliferatilon of certain other IL-2-dependent clones (data not shown).

Lymphokine Secretion by Clone 290-2

Cells from clone 290-2 were stimulated by various agents, and supernatants were harvested at different time points and assayed for IFI+-/ content by radioimmunoas-

FIG. 1. FACS IV-generated histograms of indirect immunofluorescence surface marker phenotyping of clone 290-2. The dotted lines superimposed on each histogram represent the negative control (nonbinding moAb W6/32.HK). The positive control was provided by HLA-class I-specific moAb W6/32.HL.

336 PAWELEC ET AL.

TABLE 1

Lytic Specificity of WT3 I- T-Cell Clone 290-2

Targetb

Effector’ K562 Daudi Jurkat T cells Auto1 B-LCL Al10 B-LCL

290-2 823’ -co. 1 <o. 1 <O.l co. 1 <o. 1 LAK cells 1878 1500 1622 <o. 1 80 30 NK cells 429 <o. 1 <o. 1 <o. 1 10.1 <o. 1

’ Effector cells were all derived from the same donor: LAK cells = PBMC cultured for 3 days with 100 U/ml IL-2, NK cells = adherent cell-depleted fresh PBMC.

b T cells, IL-2 cultured alloactivated lines; auto1 & allo B-LCL derived from the clone donor and alloge- neic donor, respectively.

’ Data given as LU (25%)/ 10’ effecters on 4000 target cells. r > 0.9 except where LU < 0.1. Results are from one of two similar experiments.

say and for GM-CSF by bone marrow colony formation. Table 4 shows that when the clone was incubated in IL-2-containing medium alone, no IFN-7 or GM-CSF was found in the supernatant, although cells retained viability. Even under circumstances where stimulating agents had been able to trigger maximal proliferation (TPA + IO), no IFN-y or GM-CSF was secreted. However, when 290-2 cells were incubated with B-LCL in addition to TPA + 10, although proliferation was not increased (Table 3), both IFN-7 and GM-CSF were found in the supernatants. Thus, in order to secrete these lymphokines, clone 290-2 usually requires signals in addition to those sufficient to trigger maximal autocrine proliferation.

DISCUSSION

The protein product of the TCR y-chain gene was recently identified on the surface of thymocyte and immunodeficient patient-derived CD3+ T cells which failed to re- act with moAb WT31 (3, 6). It was noted that this type of T cell was capable of responding by proliferation to alloantigen + IL-2. The same type of CD3’ WT3 l- T

TABLE 2

MoAb Blockade of Clone 290-2-Mediated Lysis of K562 Targets

MoAbb

Effector“

290-2 282-16 LAK cells

None

45.1 f 5.0’ 43.7 f 3.1 81.4 f 9.3

CD3

44.9 iz 3.9 18.1 t-2.0 71.7 f 5.9

CD4

45.9 f 4.2 34.1 + 6.7 70.7 + 6.5

CD1 la

2.4 + 0.3 10.1 & 2.3 58.6 + 6.6

“Clone 282-16 (same donor as 290-2) is a CD3+ CD4+ TCR cy/p’ NK-like clone. LAK cells were also from the same donor. E:T was 25: 1.

b Moab WT32 (1: loo), OKT4 (10 &ml), SPV-LI 1 (1: loo), respectively. c Results are presented as % specific “Cr release, mean cpm of triplicate ? SDM. Results are from one

of two similar experiments.

CD3+ WT31- CLONE 337

TABLE 3

Proliferative Capacity of Clone 290-2

Stimulator” Proliferatiot+ Stimulator Proliferation

Medium Auto1 B-LCL PHA WT31 moAb1 CD3 moAb TPA 10 Pool-lys Allo B-LCL IL-4 IL-4 + Auto1 LCL

342 f 78’ 1031 f 128 567? 80 469 f 140 204 + 46 618 f 101 545 f 54

1176 + 123 1336 & 209 197 + 262 927 f 161

Auto1 LCL + PHA Auto1 LCL + CD3 moAb Auto1 LCL + TPA Auto1 LCL + TPA + CD3 moAb Auto1 LCL + IO TPA + IO Auto1 LCL + TPA + IO TPA + IO + moAb TU69 IL-2 IL-2 + moAb TU69

7,459 k 899 2,582? 460 2,083 f 304 6,131 F 728

997+ 140 15,638 f 1285 14,267 ? 998 3,856 rt 421

26,334 f 1886 8.731 + 945

’ Stimulatmg agents were used at concentrations previously determined by titration to be optimal for stimulation (data not shown), as follows: B-LCL, 2.5 X lO“/well; pooled PBMC (pool-lys), 1 X 105/well; PHA, 1%; CD3 moAb (WT32), l,,Ng/ml; TPA, 10 rig/ml; IO, 400 rig/ml; IL-2,20 U/ml; IL-4,0.5%; MoAb WT3 1, at I: 100 ascites; moAb TU69,25% culture supernatant.

* Proliferation is given as mean cpm of triplicates ? SDM of [jH]TdR incorporation at the time point previously dIetermined to be optimal by kinetic experiments (data not shown).

’ Results are representative of three separate experiments.

cell seems, also to exist in peripheral blood of normal adult donors ( 14). Although not spontaneously cytotoxic, these cells rapidly acquired NK-like activity and NKHl (Leu 19) positivity after culture in IL-2 alone (14). Consistent with this finding, a number of later reports showed that certain cloned T cells which had been selected for their NK-like activity, but which were nonetheless CD3+, did not react with moAb

TABLE 4

Secretion of IFN-7 and GM-CSF by Clone 290-2”

Stimulus IFN-7 (U/ml) GM-CSF

Medium alone B-LCL B-LCL + PHA TPA + IO TPA + 10 + B-LCL

Controls TPA + IO TPA + IO + B-LCL

<I 0 115 2kO.5 <15 NT 115 5 + 0.5

> 1000 35 *4

<7 2&O <15 3 * 0.5

“Clone Z!90-2 cells (2 X IOs/ml) were incubated in RPM1 1640 + 10% human serum + 10 U/ml IL-2 (medium alone) and stimulated with 2.5 X lO’/ml autologus B-LCL, B-LCL + 1% PHA, 10 rig/ml TPA + 400 rig/ml 10, or TPA + IO + B-LCL. Controls were performed without clone 290-2. Supematants were harvested at the optimal time point and assayed in duplicate for IFN-7 content (48 hr) by radioimmunoas- say (Celltech, Slough, Bucks, England) or in triplicate for GM colony formation (24 hr) from normal human bone marrow (results expressed as number of colonies per 3 X IO4 bone marrow mononuclear cells). In this experiment 10% human placenta-conditioned medium supported 54 colonies per 3 X lo4 bone marrow cells. This is one of three experiments with similar results.

338 PAWELEC ET AL.

WT31 (5, 8). It was further shown by biochemical and molecular genetic analyses that these cells expressed a TCR which contained, or was entirely composed of, y- chain gene products (5, 15).

This type of cell has been propagated in vitro mostly from fetal donors, thymocytes, and immunodeficient patients (3,6, 7). We have shown in the present communica- tion that CD3+ WT3 I- cells from normal adult donors can also be clonally propa- gated with appropriate stimulation and culture schedules. Thus, stimulation with alloantigen, lectin, and IL-2 was not sufficient to generate clones, but the additional inclusion of phorbol ester and calcium ionophore supported an acceptable cloning efficiency. The CD3+ WT3l cells obtained in this way were shown by Northern blotting to express full-length mRNA for the TCR-7 but not LY or p chains, whereas alloreactive WT3 I+ clones expressed full-length (Y- and p- but not y-chain messages. Thus it is thought most likely that clone 290-2 expresses the “alternative” y-chain+ TCR. The finding that these cells are weakly CD4+ was surprising in light of reports that such cells are CD4-. However, it was found that of three subclones of 290-2 analyzed, two presented a CD4 histogram identical to that of the parental clone, but the third was completely CD4-. Thus there appears to be heterogeneity of expression of CD4 within a monoclonal population, and progressive loss or acquisition of CD4 antigen could not be excluded. This possibility is being investigated on larger num- bers of subclones of different ages, using a panel of CD4 moAb.

Consistent with expectations from previous publications on CD3+ WT3 l- T-cell clones, clone 290-2 exerted potent MHC-unrestricted cytotoxicity. However, it showed a clear-cut NK pattern of cytotoxicity, in contrast to the broader kind of LAK activity which has been noted for certain other clones of this kind (5). The expression of Leu 19 (NKHl) antigen as a marker for NK status is also compatible with this finding. Cytotoxicity of 290-2 was blocked by moAb to CD 11 a (LFA I ), as is also the case with NK cells and other NK-like clones (data not shown), further emphasizing similarities among these cells. The lack of blocking by CD3 moAb may have reflected a potential inhibition offset by the ability of CD3 moAb to redirect cytotoxicity to Fc-receptor-positive target cells. However, the finding that a CD3+ a//3-TCR+ NK- like clone from the same donor was blocked by CD3 moAb argues against this as a general mechanism. Although clone 290-2 unexpectedly expressed low levels of CD4, moAb specific for this structure failed to block cytotoxicity, suggesting that CD4 was not involved in effector-target interactions or in signal transduction.

Regarding the proliferative capacity of clone 290-2, it was shown in kinetic experi- ments that reactivity levels with different stimulators did not vary as a function of culture duration (data not shown). Thus, clone 290-2 was found not to be alloreac- tive. It did respond by autonomous proliferation to a combination of TPA + 10, but to neither agent alone. In this latter respect, it behaved identically to clones which carry the a//3 TCR. That the CD3-TCR complex of the WT3 I- cells was capable of transducing activation signals was suggested by stimulation with CD3 moAb in the presence of TPA and autologous B-LCL, although the maximum attainable level of [3H]TdR incorporation was lower than TPA + IO-stimulated proliferation. This may suggest that structures other than the TCR-CD3 complex could be involved in acti- vating this type of T cell.

Proliferation of the WT3 I- cells could be stimulated very well by IL-2, and this was blocked by the IL-2-receptor-specific moAb TU69, suggesting that this type of T cell used the same growth factor as at least the majority of a/P-TCR-carrying T cells.

CD3+ WT31- CLONE 339

However, certain T cells responding to IL-2 may drive their own autocrine prolifera- tion with IL-4 ( 16). Nonetheless, it seems most likely that IL-2 is the autocrine growth factor for clone 290-2, since its autocrine proliferation stimulated by TPA + 10 was also blocked by anti-IL-2 receptor moAb Tc69. Moreover, proliferation of clone 290-2 cells was not supported by IL-4, even in the presence of B-LCL as a potential source of IL,- 1 (Table 3).

Interestingly, the combination of TPA + IO which maximally stimulated autocrine proliferation of 290-2 failed to trigger secretion of lymphokines other than IL-2 (IFN- y, GM-CSF:), whereas the combination B-LCL + TPA + IO stimulated lymphokine secretion without enhancing proliferation. This is contrary to the general finding with TCR (Y/P’ clones in which stimulation of lymphokine secretion in the absence of proliferation, but not vice versa, is observed. The distant possibility that the B-LCL themselves are capable of secreting lymphokines after 80 Gy irradiation remains to be excluded formally. This could be the case only if the stimulated clones released factors causing the B-LCL to secrete IFN-7 and GM-CSF. The finding that ongoing GM-CSF secretion could still be detected 120 hr after the beginning of coculture of clone and B-LCL, at a time point when the irradiated B cells are disintegrating (data not shown), provides a first argument against this possibility. The fact that 1FN-r and GM-CSF secretion by WT3 I- T cells can be triggered only under extreme, nonphysi- ological conditions may, however, imply the existence of one or more as yet unde- fined specific signals for these cells to produce this lymphokine. The nature of the presumed signal from the B-LCL requires clarification.

The functional significance of NK-active T-TCR cells is by no means clear. Al- though normal donors’ peripheral T cells contain a small percentage of the CD3+ WT3 l- subtype (3.07% f 2.09 SDM in our donors, n = 22, data not shown), we have noted that in certain disease states, the proportion of CD3+ WT31l cells may be altered. In one case, a donor with a previously normal level of CD3+ WT31- cells exhibited a. very marked increase to 28%. That the proportions of such cells may be altered in viral infections was suggested from a recent survey of the literature, where a selective expansion of CD3+ CD4- CD8- cells in HIV-infected patients was impli- cated (17). The availability of clonal populations of y-TCR’ T cells from normal donors should help in the analysis of the functional relevance of this subpopulation and in the preparation of moAb specific for this subset.

ACKNOWLEDGMENTS

The assista.nce of Mss. A. Rehbein, I. Balko, C. Schmidt, K. Katrilaka, and G. Hummel is gratefully acknowledgetd.

REFERENCES

I. Oettgen, H. C., and Terhorst, C., Hum. Immunol. 18, 187, 1986. 2. Spits, H., Borst, J., Tax, W., Capel, P. J. A., Terhorst, C., and DeVries, J. E., J. Immunol. 135, 1922,

1985. 3. Brenner, M. B., McLean, J., Dialynas, D., Strominger, J. L., Smith, J. A., Owen, F. L., Seidman,

J. G., Ip, S., Rosen, F., and Krangel, M. S., Nature (London) 322, 145, 1986. 4. Lanier, I,. L., and Weiss, A., Nature (London) 324,268, 1986. 5. Borst, J., Van de Griend, R. J., van Oostveen, J. W., Ang, S.-L., Melief, C. J., Seidman, J. G., and

Bolhuis, R. L. H., Nature (London) 325,683, 1987. 6. Bank, I., DePinho, R. A., Brenner, M. B., Cassimeris, J., Ah, F. W., and Chess, L., Nature (London)

322,179, 1986.

340 PAWELEC ET AL.

7. Moingeon, P., Ythier, A., Goubin, G., Faure, F., Nowill, A., Delmon, L., Rainaud, M., Forrestier, F., Daffost, F., Bohuon, C., and Hercend, T., Nature (London) 323,638, 1986.

8. Nowill, A., Moingeon, P., Ythier, A., Graziani, M., Faure, F., Delmon, L., Rainaut, M., Formstier, F., Bohuon, C., and Hercend, T., J. Exp. Med. 163, 1601, 1986.

9. Pawelec, G., Hum. Immunol. 8,239, 1983. 10. Pawelec, G., Hadam, M. R., Schneider, E. M., and Wemet, P., J. Zmmunol. 129,2271, 1982. 11. Yanagi, Y., Chan, A., Chin, B., Minden, M., and Mak, T. W., Proc. Natl. Acad. Sci. USA 82, 3430,

1985. 12. Acute, O., Campen, T. J., Royer, H. D., Hussey, R. E., Poole, C. B., and Reinherz, E. L., J. Exp. Med.

161,1326, 1985. 13. Yoshikai, Y., Anatoniou, D., Clark, S. P., Yanagi, Y., Sangster, R., Van den Elsen, P., Terhorst, C.,

and Mak, T. W., Nature(London) 312,521, 1984. 14. Lanier, L. L., Ruitenberg, J. J., and Phillips, J. H., J. Exp. Med. 164,339, 1986. 15. Moingeon, P., Jitsukawa, S., Faure, F., Troalen, F., Triebel, F., Graziani, M., Forrestier, F., Bellet, D.,

Bohuon, C., and Hercend, T., Nature (London) 325,723, 1987. 16. Lichtman, A. H., Kurt-Jones, E. A., and Abbas, A. K., Proc. Natl. Acad. Sci. USA 84,824, 1987. 17. Marcos, M. R., Gaspar, M. L., De La Hera, A., Toribio, M. L., Marquez, C., and Martinez, A., Stand.

J. Immunol. 25,321, 1987.