Mlsa generated suppressor cells. I. Suppression is mediated by double-negative (CD3+CD5+CD4-CD8-) alpha/beta -bearing cells. This information is current as of October 2, 2021. M Bruley-Rosset, I Miconnet, C Canon and O Halle-Pannenko J Immunol 1990; 145:4046-4052; ; http://www.jimmunol.org/content/145/12/4046 Downloaded from

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision http://www.jimmunol.org/

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription by guest on October 2, 2021 Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1990 by American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. 0022-1767/90/14512-4046$02.00/0 THEJOURNAL OF IMMUNOLCGY Vol. 145.4046-4052. No. 12. December 15. 1990 Copyright 0 1990 by The American Association of lmmunologists Printed In U.S.A.

Mls" GENERATED SUPPRESSORCELLS

I. Suppression is Mediated by Double-Negative (CD3+CDVCD4-CD8-) a/@ T Cell Receptor-Bearing Cells'

Grafting of cells from B10.D2 (H-2d)donors into H- The GVHR3 remains a major problem after bone mar- 2 compatible lethally irradiated (DBA/2 x B1O.DZ)Fl row transplantation, even in MHC-compatible donor-re- hostsresults in a severe graft-vs-host reaction cipient combinations. Although elimination of T cells

(GVHR), developed against DBA/2 non-H-2 Ag, with from donor marrow inocula reduces incidence of GVHR, Downloaded from only 0 to 10% of animals surviving. This GVHR it also increases the incidence of engraftment failure, mortality rate is dramatically reduced (90 to 100% justifying the search for other treatments thatwill pre- of animals survive) by donor preimmunization vent GVHR development. In our model using adult mice, against Mls" determinants. The protection against grafting of cells from B10.D2 (H-2d)donors intoH-2 com- GVHR correlates with a decreased B10.D2 anti-DBA/ patible lethally irradiated (DBA/2 X BlO.D2)Fl recipients 2 proliferative response in vitro. Bothin vivo and in results in a severe GVHR, with only 0 to 10% of the vitro phenomena are associated with activation of http://www.jimmunol.org/ CD5+suppressor T cells in the spleens of immunized animals surviving (1). Survival is considerably increased mice. The present work was designed to study the by donor immunization, shortlybefore grafting, with par- origin of these suppressor cells and to further char- ent-strain DBA/2 spleen cells (incompatible for numer- acterize their phenotype. Theresults show that sig- ous non-H-2 Ag, including Mls") with a peak of efficiency nificant suppression is not inducible in "B" mice. In seen at 3 to 4 days after immunization (2). Other studies contrast, in mice that were only thymectomizedor showed that abrogation of GVHR mortality is in fact due else pretreated in vivo with anti-CD4 or anti-CD8 to the immunization against Mls" determinants and that mAb, the suppressor cells are activated as effi- it correlates with a decrease of the B10.D2 anti-DBA/2 ciently as in normal mice. Thesuppression of GVHR proliferative response in vitro (3, 4). For protection by guest on October 2, 2021 mortality and proliferative responses in vitro is lost against GVHR to be observed, the expression of Mls" in after depletion from preimmunized splenocytes of grafted recipients is necessary (5)or not (6),depending CD5+ T cells and remains unaltered after depletion upon the genetic combination studied. Therefore, con- of CD4+ or CD8+ T cells or both. Depletion of asialo trasting with the induction phase, the protection is not GM1+ cells removes all NK activity, whereas the specific for Mls" at the effector phase. Both the abroga- suppression is decreased only slightly.FACS analy- tion of GVHR and thedecreased proliferative response in sis showed that double-negative (DN) cells from nor- vitro are associated withthe generation in the spleensof mal and immunized mice contain both CD3+ and immunized donors of noncytotoxic CD5' suppressor T CD3- cells: the vast majority of the CD3+ DN T cells cells (4, 7). These suppressor cells are distinct from veto express the a/@ T cell receptor. Suppression of GVHR T cells effective against thespecific cytotoxic response to and of proliferative responses in vitro are abrogated non-H-2 antigens, and theireffect is detectable in mixing after elimination of CD3' cells. These results sug- experiments (7). The experiments described here were gest that Mls" generated suppressor cells: 1) are designed to characterize better the origin and the phe- derived from post-thymic long-lived T cell precur- notype of the M1s"-generated suppressor cells effective in sors: 2) are low asialo GM-1+ but do not exhibit NK vivo and in vitro. We found that the suppressor cells DN activity: 3) belongto a subset of peripheral CD5+ originate from post-thymic long-lived T cell precursors, T cells bearing a CD3-associated a/&heterodimer. and thatthey areperipheral DN (CD4-CDK)CD5' T cells, bearing a CD3-associated alp-heterodimer of the TcR. Received for publicatton February 26, 1990. Accepted for publication August 24, 1990. The costs of publication of this article were defrayed in part by the MATERIALS AND METHODS payment of page charges. This article must therefore be hereby marked Mice. Female DBA/2 (H-2" Mls'). B10.D2 (H-2" Mlsb),and (DBA/2 advertisement in accordance with 18 U.S.C. Section 1734 solely to indi- x B1O.DS)Fl (H-2"ldMls"@) mice were purchased from Centre Na- cate this fact. tional de la Recherche Scientifique Laboratories. OrliansLa Source. This work was supported by Institut de la Santl et de la Recherche Medicale. Centre National de la Recherche Scientifique, Federation Na- France. tionale des Groupements des Entreprises Francaises dans la Lutte con- Donor immunization and induction of GVHR mortallty. B10.D2 trele Cancer, Fondationpour la Recherche MCdicale, by Grant 6324from donors were immunized by i.v. injection of 2 X lo7 DBA/2 spleen Associatlon pour la Recherche contre le Cancer. and by private funds cells 3 days before grafting (2).(DBA/2 X B1O.DS)Fl adult recipient from Association contre le Cancer en Val de Seine and la Mairie de mice were givena lethal dose (10 Gy) of whole-body irradiation using Vernouillet. a cesium source (RX 30/55 M irradiator; Gravatom Industries Ltd., Address correspondence and reprint requests to Dr. Martine Bruley- Rosset. INSERM U.267 "Immunoglnetique des Allogreffes." Hbpital Paul 3Abbreviations used in this paper: GVHR. graft-vs-host reaction: DN. Brousse. 14 avenuePaul Vaillant Couturier,94800 Villejuif. France. double negative; NS, natural suppressor: PE, phycoerythrin. 4046 Mls“-GENERATED DOUBLE-NEGATIVE SUPPRESSOR CELLS 4047 Gosport, Hampshire, England)at a dose rate of approximately 1 Gy/ were prepared as described in Materials and Methods. min; 24 h after irradiation. they were grafted [i.v.) with lo7 bone Normal and isografted non-“Tx” mice servedas controls. marrow and 8 X 10‘ spleen cells from normal donors. Mortality was recorded five times a wk from day 1 (the day after grafting) to the Two to 3 mo later, the different groups of B10.D2 mice end of the experiment (1 50 days).At that time. surviving mice with were immunized or not by injecting DBA/2 cells. Three strong signsof scleroderma, alopecia, and loss of weight were con- days after immunization, spleencells from these donors sidered to be mice developing GVHR (8).To test the capacity of preimmunized cells to suppress GVHR mortality, different spleen were mixed with normalBlO.D2 bone marrowand spleen cell populations were mixed with normal bone marrow and spleen cells and injected intolethally irradiated (DBA/2 X cells before grafting. Treatment procedures of preimmunized cells B 10.D2)F1 recipients. The spleencells from allgroups of and numbersof cells added are indicated In Results. Preparation of thyrnectomlzed “TX” mice, “B” mfce, and iso- preimmunized mice protect 65 to 95%of recipients (Table grafted mice. Thymectomy of young B10.D2 adult mice was carried I), with one exception: cells from B mice protected only out 2 to 3 mo before grafting. “Bmice were prepared by performing 30% of recipients,but this protectiondoes not differ thymectomy on young B10.D2 adult mice followed by lethal irradia- significantly from that observedwith cells from non- tion and reconstitution with T-depleted syngeneic bone marrow cells. Isografted mice consist of lethally irradiated and syngeneic bone preimmunized B mice, and is significantly less efficient marrow reconstituted B10.D2 mice. (p < 0.01) than thatobserved after transferof cells from In vitro depletion.Spleen cells. either unseparatedor nylon wool, preimmunized isografted mice. The generation of sup- nonadherent (T cell-enriched fraction), were incubated for 15 min at 4°C with the appropriatedilution of one of the following antibod- pressor cells is equally efficient in “Tx” and in normal ies: Rat mAb against Thy-1, CD4. CD8 (Biosys), CD5 (Becton Dick- mice. These results indicate that the suppressor cells inson Mountain View, CA), rabbit anti-asialoGM1 antiserum (Wako originate from a long-lived peripheral T cell. Industries) and rat anti-murine CD3 mAb (C363.29B. kindly pro- In vivo administration of either anti-CD4 or anti-CD8

vided by C. A. Janeway). Fresh guineapig complement (diluted 1/8) Downloaded from was added and the mixture was incubated 1 h at 37°C. In each mAb, before immunization of B10.D2 donors, consider- experiment. cytotoxic treatment was performed at least twice. The ably depleted the CD8+ and CD4+ cells (respectively de- cells were then washed and counted. creased from 10.8 to l%and from 17.8 to 5%;Table 11). In vivo depletion. Antl-CD8 or anti-CD4 mAb (0.3 ml; diluted 1/ 5) was injected i.v. to B10.D2 donors on days -15 and -10 before Addition of cellsfrom immunized donors inhibits the immunization with DBA/2 spleen cells. GVHR developed by normal cells regardless of whether Immunofluorescence staining andflow cytometry analysis.Ali- donor immunization was performed in normalmice (62% quots of the cell populationobtained after different purification procedures were labeled with the appropriate mAb after removing suppression) or in mice previously depleted of CD4+ (77% http://www.jimmunol.org/ the dead cells by Ficoll hypaque gradient: FITC-conjugated rat anti- suppression) or CD8’ (86%suppression) cells (Table 11). Thy-1. CD5 or CD8. PE-conjugated rat anti-CD4 mAb (Becton Dick- This observation suggests that the suppressorcells orig- inson). FITC-conjugated hamster anti-murine CD3 (145.2Cll) (9). inate from a cell subpopulation that is negative for CD4 FITC-conjugated goat anti-rat Ig (MARK-1, Biosys), and FITC-conju- gated goat anti-mouse Ig (GAM, Nordic). For two-color analysis of and CD8 markers. Immunofluorescence. FITC-conjugated hamsteranti-murine CD3 The phenotypeof the suppressor cells active against (145.2Cll) wasused together with biotinylated-anti-a/P TcR (H57- GVHR CD5+, CD4-,and CD8-. We have demonstrated 157) (10) andavidin-PE as second reagent. Cells were incubated 30 is min at 4°C In the uresenceof the coniueated mAb and washed twice previously that the suppressive activity of cellsfrom with PBS buffer, ind fluorescence OF go00 cells was analyzed on a immunized donors is exerted by the nylon nonadherent FACS (EPICSVor ATC 3000). subpopulation, and that itis removed by anti-Thy-1 and by guest on October 2, 2021 NK cell activity. Various numbers of spleen cells were incubated for 4 h at 37°C with lo4”Cr-labeled YAC-1 cells in U- anti-CD5 + C treatments (7). In the present work, the well microtest plates. E:T ranged from 200:1 to 6:1. The amountof suppressor cells were further characterized by depleting ”Cr released intothe supernatants was measured in a gamma the CD4+ or the CD8’ T cells fromthe splenic population counter (LKB). Cultures were carried out in triplicate. and the per- centage of specific lysis was calculated as follows: TABLE I ActIuatlon ofsuppressor cells In “Tx”but not In “E”mIce (experimental release - spontaneous release) 100 - Experiment Transfer of Spleen No. of GVHR Mlceb/ (maximal release - spontaneous release) Cells“ from: Total No. ofGrafted Mice ’GVHR In thistest, spontaneous release wasmeasured in sixreplicate 1 None 20/20 100 “Tx”miceC cultures containing “Cr-labeled YAC-1 cells as targets and 20/20 100 PI miced cells as negative effectors. 6/20 30 PI mice 7/20 35 Mixed reaction. Responding spleen cells (4 x lo5/ “Tx” well] from normal B10.D2 mice were stimulated with irradiated (20 2 None 13/13 100 Gy) spleen cells fromnormal B10.D2 (control) or DBA/2 (experimen- Isografted mice‘ 12/13 92 tal) mice. To test their capacity to suppress theB10.D2 anti-DBAl2 “Bmice’ 13/13 100 proliferative response, spleen cells from Immunized mice were irra- PI mice 3/12 25 diated (12 Gy) and added either unseparated (4 X 105/well) or after PI isograftedmice 2/13 15 variouspretreatment procedures:enriched. T or Thy- 911Pi ‘B mice 3 708 l.CD5.CD4.CD8, or CD3-depleted populations (2 X 105/well). Cul- In all groups. lethally irradiated (DBA/2 X B1O.DP)Fl recipients were tures were performed in Click medium containing antibiotics and grafted 1.v. with lo7bone marrow and 8 X lo6 spleen cells from normal 5% FCS. Themixture was incubated for 3 days at 37°C under B10.D2 donors. Before grafting, spleen cells (8 X 10‘) fromindicated standard conditions; [3H]thymidine (1 pCi/well) was added for the experimental groups were mixed with bone marrowand spleen cells from last 18 h of culture, and incorporated radiolabel was determined normal donors. with a @-scintillationcounter. bGVHR mice include dead mice and mice showing strong signs of Statlstics. Survival curves were compared onday 150 after graft- GVHR 150 days after transplantation. ing using Wilcoxon’s -sum nonparametric test. Significance of ‘B10.D2 adult mice were subjected to thymectomy “Tx” 2 mo before the results observed in in vitro mixed lymphocyte reaction experi- grafting. ments was determined using Student’s t-test (t statistic for the PI. B10.D2 mice prelmmunized (i.v.1 with 2 X lo7DBA/2 spleen cells comparison of two means). Significance between two percentages 3 days before transfer. was evaluated by using x2 test. e 810.D2 isografted mice wereprepared by lethal irradiation and recon- stitution with T-depleted syngeneicB10.D2 bone marrow (lO’ceils). ’B10.D2 B mice were prepared by performing adult thymectomy fol- RESULTS lowed 2 mo later by lethal irradiation and reconstitution with T-depleted Origin of suppressor cells. To study the origin of the syngeneic B10.D2 bone marrow(10’ cells). Not significant as compared with “Bmice and p < 0.01 (X’ test] as Mlsa-generated suppressor cells, B10.D2 “Tx”or “Bmice compared with preimmunized isografted mice. 4048 MlS*-GENERATED DOUBLE-NEGATIVE SUPPRESSORCELLS

TABLE I1 CD4-CD8- phenotupe of the precursorsof sumressor cells % PI Cells Positive forb Transfer of Spleen Cells No. of GVHR MlceCj from PI Mice" CDB CD4 Total No. of Grafted Mlce x GVHR

~~ ~ None 12/12 100 Un treated 10.8 Untreated 5/12 17.8 38 Anti-CD8 mAb-treatedd 1 25 2/14 14 Anti-CD4 mAb-treatedd 5.5 10.2 3/13 23 "See Table I. footnotes a and d. Splenic populationsof immunized mice untreated or pretreated in vivo with CD8 or CD4 mAb were evaluated by FACS analysis for their percentage of CD8- and CD4-positive cells before transfer with the mixture of bone marrow and spleen cells from normal mice. See Table1. footnote b. 'B10.D2 mice received i.v. 0.3 ml of CD8 mAb or of CD4 mAb (diluted 1/5) on day -14 and -10 before transfer.

of preimmunized mice, before mixing with bone marrow TABLE 111 and spleen cells from normal donors and grafting of the Removal of NK cell acttutty of spleen cellsfrom normal or prelmmunlzed mlce by ln uftro treatment wtth antl-aslalo GMI but mixture into lethally irradiated (DBAI2 x BlO.D2)F1 re- not wtth antt-CD5 + C cipients. Specificity for Mls" inthe induction of the . 7% Cytotoxlclty of Spleen Cells" from: suppression was observed previously (3)in experiments Responding Splenlc Populatlon in which the protective effect induced by immunization Normal mlce Preimmunlzed mlce Unseparatedspleen cells 25.7 f 1.9 29.3 f 1.8 Downloaded from against Mls" was compared with that obtained after im- NAdb cells 33.2 f 2.5 38.4 f 2.1 munization against othernon-H-2 or H-2 Ag. In the pres- Anti-CD5 + C' treated NAd 23.2 f 1.2 35.8 * 2.2 ent experiments,the specificity for Mls" in theinduction cells Anti-asialo GM1 + C'Of0 Of0 of the suppression was confirmed in a control group by treated NAd cells transferring splenic cells from B10.D2mice activated " Percent cytotoxicity was evaluated by measuring the amount of "Cr polyclonally in vivo by ConA together with bone marrow released by 10' YAC cells in the presence of spleen cells at an E:T of and spleen cells from normal mice. Results indicate that 1OO:l. http://www.jimmunol.org/ NAd, nonadherent cells. spleen cells derived from Con A-stimulated mice are de- T cellswere incubated for15 min at 4°C with the appropriate dilutlons void of suppressive activity (Fig. la), whereas, as ex- of indicated mAb; then fresh guinea pig complement (diluted 1/8) was pected, the suppression of mortality by cells from Mls" added and the mixture was incubated 1 h at 37°C. preimmunized mice is mediated by T cells and is abro- TABLE IV gated by anti-CD5 C treatment (Fig. lb). In addition, + Effect ofantt-CD5 or aslalo GMl treatment on the suppressor actluttu the suppression is not affected by elimination of CD4' or Treatment of T cells' No. of GVHR Mlceb/Total of CD8+T cells (Fig. 1b). These dataconfirm our previous [no.of cells addedfmousel No.of Grafted Mlce x GVHR Mice observations and extend them by indicating that sup-

None 8/8 100 by guest on October 2, 2021 pressor cells are bothCD4 negative and CD8 negative. Untreated (4 X IO') N 8/8 100 Suppressor cells active againstGVHR are low asialo Untreated (4 X 10') PI 1/8 12 CD5-depleted (3 X N 8/8 100 GMI+ but do not exhibit NK actfvtty. Before grafting, CD5-depleted (2 X 10') PI 8/8 100 spleen cells from normalor immunized donors were Asialo GM 1-depleted (3X IO'] N 8/8 100 treated by anti-asialo GM 1 or anti-CD5mAb + C. The NK Asialo GM 1-depleted (2 X 10')PI 4/8 50' activity is unaltered after depletion of CD5+ cells (Table T cellswere obtained by nylon wool purification of spleen cellsderived from normal B10.D2 mice (N) or from B10.D2 mice immunized (I.v.) 3 111), whereas suppression of GVHR by cells from preim- days before transfer with 2 X lo7DBA/2 spleen cells (PI). In all groups, munized mice is completely abolished (Table IV). In con- lethally irradiated (DBA/2 X 810.D2)Fl recipients were grafted i.v. with lo' bone marrow and 8 x 10' spleen cells from normal B10.D2 donors. Before grafting, spleen cells from indicated experimental groups were mixed with bone marrowand spleen cells from normal donors. See Table 1. footnote b. See Table 111. footnote c. 'NS. not statistically significant as compared with untreated preim- munized cells.

trast, asialo GMl -depleted cells lose their NK activity (Table 111). but 50%of suppression remains (Table IV). These results suggestthat theprotection against GVHR mortality is not due to NK cells but to CD5' Ts cells slightly sensitive to anti-asialoGM1 treatment. DN suppressor cells active against GVHR express CD3. To establish definitively that the suppression is 20 - exerted by DN T cells, we prepared nylon nonadherent splenic Tcells, depleted of contaminating B cells on anti- DRYS AFTERGRAFTING Ig-coated plastic dishes or on anti-Ig-coated dynabeads Ftgure 1. CD4-CDEWCD5+phenotype of Ts cells effective against (Dynal); we then eliminated both the CD4+ and CD8+ GVHR. In all groups lethally irradiated (DBA/2 x B1O.DZ)Fl recipients populations. Figure 2 showsa representative example of were grafted witha mixture of bone marrowand spleen cellsfrom normal B10.D2 mice. Mixture of normal cells alone,(--); a. Mixture of normal the immunofluorescence staining and flow microfluo- cells and cells from in vivo Con A-activated mice (- - -1; b. Mixture of rometry analysis of these DN cells from normal (Fig. 2a) normal cells and cells from preimmunized mice, which were either T- and preimmunized (Fig. 2b) mice isolated as described enriched (.-.-.-). T-CD5-depleted (*-*-*-), T-CD8-depleted (. . .). or T-CD4- depleted (- - -). above. It is important to note that thedepletion of CD4+ MlS"-GENERATED DOUBLE-NEGATIVE SUPPRESSORCELLS 4049 - MARK-1CD8 CD4 THY-1 CD5 GAM

Ffgure2. FACS analysis of DN cells iS0- lated from normal (a)or preimmunized (b) mice. Spleencells were stained with the following markers: FITC-conjugated anti- Leu 11 (control).FITC-conjugated MARK-1. FITC-conjugated anti-Thy-1,anti-CD8. anti-CD5. PE-conjugated anti-CD4 mAb. and FITC-conjugated goat anti-mouse: 5000 cells were analyzed onATC 3000 cell Sorter.

FLUORESCENCE INTENSITY and CD8' cells is nearly complete, inasmuch as only 1 to staining. It can be seen that 30%of cells are CD3+ among 3.7% of the cells stain with directly conjugated anti-CD4 the DN population from normal mice and 54% areCD3' or anti-CD8 mAb, and only 4 to 5% of the cells remain among DN cells from preimmunized mice, and that the coated after cytotoxic procedures with these antibodies vast majority of DN cells that are positive for CD3 also as detected by a goat antiserum directed against rat Ig express a/P-TcR (Fig. 3). To examinewhether the (MARK-1). Among DN cells, 34.5% are Thy-l+ innormal suppression is mediated by CD3' cells, the spleen cells Downloaded from mice and 49% areThy-1' in preimmunized mice. The DN from preimmunized mice were depleted of CD4'. CD8'. cells from normal and preimmunized mice also stain with or CD3+ cells and added to the mixture of normal cells CD5 mAb (29.6 and 43%.respectively). Among both nor- grafted to irradiated F1 recipients(Table V). Suppression mal and preimmunized DN cells, the percentage of B cells of GVHR remains after elimination ofCD4' and CD8' foundunder our experimental conditions is relatively cells (Table V, experiments 1 and 2) butis completely lost high (24 and 23%,respectively) for two reasons: 1) if the after elimination of CD3' cells (Table V, experiment 2). conditions we use for purification of T cells are too "dras- Considered together, these results indicate that suppres- http://www.jimmunol.org/ tic" (for example,two incubations on nylon-wool columns sor cells activeagainst GVHR expressthe Thy- instead of one), we lose a fraction of preimmunized T l'CD3'CD5+CD4-CD8.- phenotype and aretherefore DN cells, and along with thempart of the suppressive activ- T cells. ity, probably because activated T cells are mildly nylon- Suppressorcells active against proltferative re- adherent: and 2) afterdepletion of CD4' and CD8' cells, sponses in vitro also express the DN phenotype and together representingmore than 80%of T cells, the per- the CD3 molecule. As shown previously (4), theprolifer- centage of contaminating non-T (Thy-1-, CD3-, B, and ation of B10.D2 cells in response to irradiated DBA/2 NK) cells obviously rises. In fact. the contamination by stimulators is significantly decreased after immunization by guest on October 2, 2021 non-T cells is not critical because we know from in vivo of B10.D2 mice against Mls" determinants expressed on experiments that suppressor cells are Thy-l+ and that DBA/2 immunizing cells. In the present work, the phe- they do not exhibit NK activity. notype of the suppressorcells was furtherstudied in vitro DN cells from spleens of normal and preimmunized in mixing experiments by adding various T cell subsets mice were then labeled with FITC-conjugated anti-CD3 from preimmunized or normal B10.D2 mice to the MLR mAb (145.2Cll)together with biotinylated anti-cr/P TcR between normal B10.D2 responding cells and DBA/2 mAb (H57-157) plus avidin-PE. Figure 3 represents an stimulating cells. The suppression is detected in unsep- example of two-color analysis of immunofluorescence arated and T-enriched splenic cells from preimmunized mice (Fig. 4, experiment 1);it is abolished by anti-Thy-] or anti-CD5 mAb + C treatments but not by anti-CD4 or

TABLE V CD3+CD4TD8- phenotypeof the suppressor cellseflectfue against

GVHR- . ""

Transfer of T Cells from PI Mice' No' Of Dead Mice/Tota' No. of Grafted M~CP 'GVHR W CD8 CD3 CD8 Exp. 1 13/13 100None 13/13 Untreated (5 X 10') 30 4/13 CD4 and CD8 depletedb 0/10 0 13 x 106)

Exp. 2 11/12 92 None 11/12 Untreated (4 X lo6) 1/13 8 CD4 and CD8 depletedb 2/10 20 (3 x 106) CD8 CD3 CD3 depletedb (3 X 10'1 14/14 -"1 nn FLUORESCENCE (LOG) a Preimmunized T cells were obtained by nylon wool purification of Ffgure 3. TcR phenotype of splenic DN cells isolated from normal (a) spleen cells derived from B10.D2 mice injected 3 days before transfer or preimmunized (b]mice. The cellswere prepared as described in Mate- with 2 X 10' DBA/2 spleen cells and grafted along with bone marrowand rials and Methods. Double staining was performed using FlTC-conju- spleen cells from normal donors.In all groups, lethally irradiated(DBA/2 gated anti-CD8mAb together withPE-conjugated anti-CD4 mAb or FITC- x B1O.DZ)Fl recipients were grafted i.v. with 10' bone marrow plus 8 X conjugated anti-CD3 mAb together with biotinylated anti ala TcR mAb 10' spleen cells from normalBlO.D2 donors. and avidin PE; 5000 cells were analyzed onan EPICS V cell sorter. See Table111, footnote c. 4050 M1S"-GENERATED DOUBLE-NEGATIVE SUPPRESSOR CELLS

NETTHYMIDINE INCORPORATION cells may contain afew DN T cells that areless sensitive PROLIFERATIVE RESPONSE OF B10.02CELLS (CPM x 10-3 c s.E.) TO 15 DBA12 STIMULATORS IN THEPRESENCE OF : 5 10 to this treatment because of a low Thy-1 Ag density. As 4x105 UNSEPARATED N generally described in the literature, theDN cells repre- PI sent about 2 to 3%of the splenic cells and contain both 2x105THYI-DEPLETED N PI CD3+ and CD3- cells. The high sensitivity of suppressor cells to anti-CD3 treatment excludes the possibility that contaminating B cells might contribute to the suppres- sion. The suppressor cells are low asialo GMl+ (11). but 2x105Co8-DEPLETEU the possibility that theymight be NK cells can be excluded for thefollowing reasons: 1) theDN suppressor cells bear 2x105 CD4-DEPLETED N PI the CD3 molecule:2) anti-CD5treatment completely elim- inated suppressiveactivity both in vitro and in vivo with- 4x105 UNSEPARATED PI out eliminating NK activity in vitro; 3) DN Thy-1' cells (selected by positive panning) are suppressive whereas DN Thy-1- cells are not (our unpublished observations); 4) anti-asialo GM1 treatment removed all NK activity in vitro, although only slightly affecting thesuppression of N GVHR. The presenceof immunizing (DBA/2)cells among 4x105 UNSEPARATED +IL-2 immunized (BlO.D2)DN T cells we describe hereis prob- Downloaded from 2x105 CD4 + Co6-DEPLETED Fj able, but the suppressive activity observed cannot be attributed to Tcells of DBA/2 origin because depletion of Flgure 4. Thy-l+CD3+CD5+CD4-CD8-phenotype of suppressor cells Lyt-1.2+ (B10.D2 allele) but not of Lyt-1.1' (DBA/2 allele) effective against proliferative response in vitro. Spleen cells (4 X lo5) from normal BlO.D2 mice were stimulated with irradiated spleen cells (4 cells removed this activity in vivo (7). X 10') from DBA/P mice together with irradiated unseparated (4 x 10'1 or Relatively little is known about the function, antigenic depleted (2x 10') spleen cells from normal [N) or from preimmunized (PI] mice or from mice stimulated in vivo with 30 pg of Con A. The cultures repertoire, and lymphokine production (or responsive- were performed either in the absence or in the presence of 5%exogenous ness) of DN T cells present in theperiphery. In the mouse, http://www.jimmunol.org/ IL-2. several reports have described a distinct population of CD3+ DN T cells that appear in the fetal thymus and are anti-CD8 mAb C treatments (Fig. 4, experiment 1).In + still present in the mature thymus and the peripheral addition, the DN cells (depleted of both CD4' and CD8' lymphoid organs throughout adult life. The majority of populations) from preimmunized mice are also active in normal splenic DN T cells bear y/6 TcR (1 2, 13). andDN mediating suppression; in contrast, CD3-depleted cells T cells activated in mixed lymphocyte reaction express from preimmunized mice lose their capacity to suppress functional y-transcripts (14). This type of cell was also mixed lymphocyte reaction (Fig. 4, experiment 2). To found in murine skin (15) and intestinal epithelium 6).(1 exclude the possibility that suppression simply reflects by guest on October 2, 2021 A minor set of DN T cells bearing a/P-TcR was identified IL-2 absorption by activated (IL-2 receptor-positive) cells, in the thymus (17, 18).Our double-labeling experiments we evaluated the suppressive effect of in vivo Con A- clearly indicated that, contrastingwith resultsdescribed activated spleencells and of DN cells from preimmunized by others, the vastmajority of CD3' DN cells in the spleen mice in thepresence of exogenous IL-2. Results presented of normal and preimmunized B10.D2 miceexpress an a/ in Figure 4, experiment 2 demonstrate that when exoge- P-TcR. The question arises as to whether theproportion nous IL-2 is added to the cultures, DN cells from immu- of yl6-v~a/p-peripheral DN cells may vary among mouse nized mice retain full suppressive capacity, whereasCon strains. The phenotype of the DN Ts cells we describe A-activated cells (negative control) do not suppress. These here strongly resembles the CD5+ a/P-TcR DN T cells data demonstrate that the suppressor cells effective in found in the thymus and in theperiphery of C57BL/Ka vitro, like those effective in vivo, expressthe Thy- mice, the function and specificity of which are still un- l+CD3+CD5'CD4"2D8" phenotype (DN T cells) and do known (19). Cells with a similar phenotype behave as not act by absorbing out the IL-2. autoreactive T and recognize self peptides bound to I-A molecules on the surface of dendritic cells DISCUSSION (20). In this paper, we characterized the Ts cells generated DN Ts cells were found in human thymus (21) and in by immunization against Mls" determinants. These sup- the blood of bone marrow grafted patients (22).In mice, pressor cells are responsible for the inhibition of both DN NS cells have been identified in thespleen of neonatal the GVHR developed to DBA/2 non-H-2 Ag after bone mice (23, 24). in thebone marrow of adult mice (25, 26). marrow transplantation and the primary proliferative as well as after total lymphoid irradiation (24, 27) or response invitro to Mls" determinants expressed on host- cyclophosphamide treatment (28) or during GVHR (26, specific DBA/2 stimulators (3).Considered together, the 29). However, it is interesting to note that M1s"-generated results in vivo and in vitro indicated that thesuppressor suppressor cells are not sensitive to leucyl-methyl-ester cells have a Thy- l',CD3',CD5+,CD8-,CD4- phenotype treatment (datanot shown)as opposed to those found in and are therefore DN T cells. The DN T suppressor cells pathologic situations suchas GVHR (29).suggesting that originate from a post-thymiclong-lived T cell that is most they belong to different cell populations. The Tcell origin likely also DN. A slight, although not significant suppres-of NS cells is controversial since in some reports they sion is inducible by immunization of "Bmice, possibly were found to be Thy-1 positive (29, 30) and in others for two reasons: 1) some T cells may escape the 10-Gy Thy-1 negative (24, 25, 27,31). DN, Thy-1' NS cells were irradiation; and 2) anti-Thy-1+ C'-treated bone marrow derived from fresh spleen and thymus of normal adult MW-GENERATED DOUBLE-NEGATIVE SUPPRESSORCELLS 405 1 mice (30).Also dealing with normal mice, Strober et al. Ginette Debotte and Veronique Gay-Bellile for FACS anal- (32) described a cloned splenic Thy-1' NS cell that is yses. asialo GM l+CD5-CD4-CD8-, and expresses thea/P-TcR. Therefore, except for the CD5 marker, these cloned NS REFERENCES cells resemble the M1s"-generated suppressor cells we 1. Halle-Pannenko. 0..L. L. Pritchard, R. Motta, and G. Mathe. 1978. describe here; at thepresent stage of our knowledge, we Non-H-2 antigens can induce highGVH mortality in adult recipients of normal cells. Biornedlclne29:253. cannot conclude whether they represent distinctlineages 2. Halle-Pannenko. 0..L. L. Pritchard, and H. Rappaport. 1983. Al- or the samecell at different stages of activation. loimmunization-activated suppressor cells. 1. Abrogation of lethal Thesuppressor properties of the Mls determinants graft-versus-host reactiondirected against non-H-2 antigens.Trans- plantation 36:60. were first observed by Lilliehook et al.(33) and, asin our 3. Halle-Pannenko, 0.. L. L. Pritchard. H. Festenstein. and L. Beru- experiments (6).directa relationship between the men. 1986. Abrogation of lethal graft-versus-host reactiondirected strength of stimulation and thedevelopment of suppres- against non-H-2 antigens:role of MIS" and K/l region antigens in the induction of unresponsiveness by alloimmunization. J. Irnrnunoge- sion by in vivo immunization was demonstrated. In our net. 13:450. experiments, the induction of suppression was highly 4. Pritchard, L. L., and 0.Halle-Pannenko. 1983. Alloimmunization- activated suppressor cells. 11. In vitro activity of suppressor cells specific for Mls" determinants and was absent or negli- implicatedin the abrogation of lethal graft-versus-host reaction. gible after immunization with other non-H-2 Ag or with Transplantation 36:310. H-2 Ag. This conclusion is further confirmed inthe pres- 5. Halle-Pannenko, 0..L. L. Pritchard. M. Bruley-Rosset.L. Berumen. and R. Motta. 1985. Parameters involved in the induction and ab- ent experiments by demonstrating that polyclonal acti- rogation of the lethalgraft-versus-host reactiondirected against non- vation with Con A does not give rise to cells exhibiting H-2 antigens. Irnrnunol. Rev. 88:59. suppressive activity either invivo or in vitro. In addition, 6. Halle-Pannenko. 0..and H. Festenstein. 1981. Incompatibility for Downloaded from or pre-immunizationagainst Mls determinantsdecreases lethal Jadus andPeck (23) have reported experiments suggest- graft-versus-host reactiondeveloped across non-H-2 and/orH-2 bar- ing a role of Mls in suppressing lethalGVHR via NS cells. riers. J. Irnmunogenet. 8:443. 7. Halle-Pannenko, 0..L. L. Pritchard, and M. Bruley-Rosset. 1987. Our results arealso consistent with the data (34) indicat- Abrogation of the lethal graft-vs-host reactiondeveloped to non-H-2 ing that Thy- l'CD5TD8- suppressor cells are generated antigens: involvement of T suppressor cells distinctfrom veto cells. in an anti-Mls" mixed lymphocyte reaction;since the Eur. J. Irnrnunol. 17:I 751. 8. Rappaport, H.. A. Khalil, 0. Halle-Pannenko, L.L. Pritchard, D. mechanism of this in vitro-generated suppressionre- Dantchev. and G. Mathe. 1979. Histopathologic sequence of events http://www.jimmunol.org/ quired a proliferative stimulus delivered by Mls" deter- in adultmice undergoing lethal graft-versus-host reactiondeveloped minants, it is possible that it is the same as themecha- across H-2 and/or non-H-2 histocompatibility barriers. Am. J. Pa- thol. 96:121. nism of in vivo-generated suppression described here. 9. Leo, 0..M. Foo. D. H. Sachs, L. E. Samelson. and J. A. Bluestone. Among severalhypotheses, we can envisage that the 1987. Identification of a specific for a murine generation of DN Ts cells by Mls" is: 1) a "second step" T3 polypeptide. Proc. Natl. Acad. Sci. USA 84~1374. 10. Kubo, R. T., W. Born, J. W. Kappler, P. Marrack, and M. Pigeon. phenomenon resulting from classic recognition ofMls" 1989. Characterization of a monoclonal antibody which detects all determinants by CD4+cells; or 2) theconsequence of the murine a0 T cell receptors. J. Irnrnunol. 142:2736. 11. Stout, R. D., G. A. Schwarting. and J. Suttles. 1987. Evidence that recognition of Mls" (seen without class I1 molecules) by expression of asialo-GM1 may be associated with cell activation: DN T cells, in which lack of the CD4 accessory molecule correlation of asialo-GM1 expression with increased total cellular by guest on October 2, 2021 leads to activationof a "negative signal" instead of prolif- RNA and content in normal and spleen cell pop- ulations. J. Irnrnunol. 139:2123. eration. In our experiments, therecipient mice reconsti- 12. Cron. R. Q.. T. F. Gajewski, S. 0. Sharrow. F. W. Fitch. L. A. Matis, tuted with cells from M1s"-preimmunized donors became and J. A. Bluestone. 1989. Phenotypic and functional analysis of 100%chimeras: but the mechanism of tolerance estab- murine CD3TD4-CD8- TCR 76 expressing peripheral T cells. J. Irnrnunol. 1423754. lished in grafted mice is peripheral, sincethymectomized 13. Bluestone, J. A., R. Q. Cron, M. Cotterman. B. A. Houlden. and L. and nonthymectomized recipients survived equally well A. Matis. 1988. Structure and specificity of T cell receptor y/6 on major histocompatibility complex antigen-specific CD3'. CD4-. CD8- (manuscript in preparation). The next paperwill describe T lymphocytes.J. Exp. Med. 168: 1899. several other properties of M1s"-generated DN cells in- 14. Jones, B.. S. Carding, S. Kyes, S. Mjolsness, C. Janeway, and A. cluding the specificity of the suppression at theeffector Hayday. 1988. Molecular analysis of T cell receptor gamma gene expression in allo-activated splenic T cells of adult mice. Eur. J. phase in comparison with DN cells from normal mice. Irnrnunol. 18:1907. The role of the TcR in the suppressive function of these 15. Havran, W. L., S. Grell. G. Duwe. J. Kimwa. A. Wilson. A. M. cells and the modifications of the repertoire (VP usage) Kruisbeek, R. L. O'Brien. W. Born. R. E. Tigelaar, andJ. P. Allison. 1989. Limited diversity of T-cell receptor y-chain expressionof mu- within CD4+, CD8'. and DN T cell populations are cur- rine Thy-1' dendritic epidermalcells revealed by Vy3-specific mono- rently under study in the spleens of normal and M1s"- clonal antibody.Proc. Natl. Acad. Sci. USA 86:4185. immunized mice. In conclusion, M1s"-generated a/P-DN 16. Kyes. S., E. Carew. S. R. Carding, C. A. Janeway, Jr., and A. Hayday. 1989. Diversity in T-cell receptor y gene usage in intestinal suppressor cells are functionalin vitro and in vivo; they epithelium. Proc. Natl. Acad. Sci. USA 86:5527. resemble a/@,DN NS cells found in normal adult mice 17. Budd. R. C., G. C. Miescher, R. C. Howe, R. K. Lees, C. Bron. and H. R. MacDonald. 1987. Developmentally regulated expression of T and may represent a functionally important subpopula- cell receptor 0 chain variable domains in immature , J. tion involved in immunoregulation and immunosurveil- Exp. Med. 166:577. lance of peripheral T cells. 18. Fowlkes, B. J., A. M. Kruisbeek. H. Ton-That, M. A. Weston. J. E. Coligan, R. H. Schwartz. andD. M. Pardoll. 1987. A novel population of T cell receptor a@ bearing thymocytes which predominantly ex- Acknowledgments. We are grateful to Dr. C. A. Jane- presses a singleV0 gene family. Nature329:251. 19. Guidos, C. J., I. L. Weissman, and B. Adkins. 1989. Developmental way for his gift of rat anti-murineCD3 hybridoma cells, potential of CD4-8- thymocytes: peripheralprogeny includes mature and to Dr. M. Papiernick for providing conjugated anti- CD4-8- T cells bearing a0 T cell receptor. J. Irnrnunol. 1423773. CD3 L. 20. Seman. M.. S. Boudaly. T. Roger. 3. Morisset, and G. Pham. 1990. and anti-a/@TcR mAb; we also thank Dr. Linda Autoreactive T cells in normalmice: unrestricted recognition of self Pritchard for helpful discussions and for reviewing the peptides on dendritic cell 1-A molecules by CD4-/CD8- TcR a/@ T cell manuscript. We are indebtedto Danielle Bonardelle, clones expressingV08. 1 gene segments. Eur.J. Irnrnunol. 20:1265. 21. Gelin, C.. L. Boumsell. and A. Bernard. 1986. Control of cell prolif- Laurence Majbruch, and Isabelle Vergnon for technical eration within the human thymus:a very limited thymocyte subpop- assistance, Annie Talia for secretarial assistance, andto ulation generatesa suppressive activity. Eur.J. Irnrnunol. 16:1209. Ml#-GENERATED DOUBLE-NEGATIVE SUPPRESSOR CELLS

22. Vilmer, E., P. Gugliemi, V. David, G. Leca. C. Rablan, L. Degos, M. 28. Segre. M., E. Tomei, and D. Segre. 1985. Cyclophosphamide-induced Boiron. and A. Bensussan. 1988. Predominant expressionof circu- suppressor cells in mice: suppression of the antibody response in lating CD3+ lymphocytes bearing gamma T cell receptor in a pro- vitro andcharacterization of the effector cells. Cell. lrnrnunol. longed immunodeficiency after allogeneic bone marrow transplan- 91:443. tation. J. Clin. Invest. 82755. 29. Wall, D. A.. S. D. Hamberg. D. S. Reynolds, S. J. Burakoff, A. K. 23. Jadus. M. R., and A.B. Peck. 1984. Naturally occurring spleen- Abbas. and J. L. Ferrara. 1988. Immunodeficiency in graft-versus- associated suppressor activity of the newborn mouse: requirement host disease. I. Mechanism of immune suppression. J. lrnrnunol. for two genetic restrictions in suppressionof lethal graft-versus-host 140:2970. disease by newborn spleen cells. Scand. J. lrnrnunol. 20:81. 30. Schwadron. R. B.. V. Palathumpjat, and S. Strober. 1989. Natural 24. Oseroff. A.. S. Okada. and S. Strober. 1984. Natural suppressor(NS) suppressor cells derived from adult spleen and thymus. Transplan- cells found in the spleenof neonatal mice and adultmice given total tatton 48: 107. lymphoid Irradiation (TLI] express the null surface phenotype. J. 31. Maier, T.. J. H. Holda, and H. N. Claman. 1986. Natural suppressor cells: members of the LGL regulatory family. lrnrnunol. Today lrnrnunol. 132:lOl. (NS) 7~312. 25. Sykes. M.. A. Eisenthal, and D.H. Sachs. 1988. Mechanism of 32. Strober. S., S. Dejbachsh-Jones. P. Van Vlasselaer. G. Duwe, S. protection from graft-vs-host disease in murine mixed allogeneic Salimi. and J. P. Allison. 1989. Cloned natural suppressorcell lines chimeras. I. Development of a null cell population suppressive of express the CD3TD4-CD8- surface phenotype and the a.0 hetero- cell-mediated lympholysisresponses and derived fromsyngeneic dimer of the T cell antigen receptor. J. lrnrnunol. I43:1118. bone marrow component.J. lrnrnunol. 140:2903. 33. LilliehWk, B., H. Jacobsson, and H. Blomgren. 1975. Specifically 26. Holda. J. H., T. Maier, and H. N. Claman. 1986. Natural suppressor decreased MLC response of lymphocytes from DBA mice injected activity in graft-vs-host spleen and normal bone marrow is aug- with cells from H-2 compatible M antigen incompatible strainC3H: mented by IL-2 and interferon.J. lrnrnunol. 137~3538. no such effect after injection of H-2 disparate C3H-hybrid cells. 27. Weigensberg, M.. S. Morecki. L. Weiss, Z. Fuks. and S. Slavin. Scand. J. lrnrnunol. 4:209. 1984. Suppression of cell-mediated immune responses after total 34. Macphail, S., and 0.Stutman. 1982. Suppressor Tcells activated in lymphoid irradiation (TLI). I. Characterization of suppressor cells of a primary in vitro response to non-major histocompatibility alloan- the mixed lymphocyte reaction. J. lrnrnunol. 132:971. tigens. J. Exp. Med. 156:1398. Downloaded from http://www.jimmunol.org/ by guest on October 2, 2021