Pregnancy-associated suppressor cells in mice: Functional characteristics of CD3+4−8−45R+ T cells with natural suppressor activity

Journal ~/ Reproductive hmmmology. 21 (1992) 103-125

103

Elsevier Scientific Publishers Ireland Ltd.

JRI 00751

Pregnancy-associated suppressor cells in mice: Functional characteristics of CD3+4-8-45R ÷ T cells with natural suppressor activity J a n e C. B r o o k s - K a i s e r a, R o b e r t A. M u r g i t a b a n d D a v i d W. H o s k i n a aDepartment of Microbiology, Dalhousie University, Halifax, Nova Scotia and t~Department of Microbiology and Immunology, McGill University, Montreal, Quebec (Canada) (Accepted for publication 3 October 1991)

Summary Natural suppressor (NS) cells are MHC-unrestricted regulatory cells with nonspecific inhibitory activity for immune responses. In adult mice, NS cells are characteristically found in bone marrow and in splenic tissue following total lymphoid irradiation and cyclophosphamide treatments. Recently, we have shown that the spleens of pregnant mice harbour a population of lymphocytes which resemble NS cells in terms of phenotype and inhibitory activity. In this study, we use positive and negative selection techniques to further characterize splenic pregnancy-associated NS (SPANS) cells as predominantly 'double negative' T cells (CD3+4-8 -) bearing receptors for the lectins wheat germ agglutinin and soybean agglutinin, as well as expressing CD45R and the heat-stable Jl ld.2 antigen. Taken together, these findings lead us to conclude that SPANS cells belong to an immature T cell lineage. In keeping with their T cell phenotype, SPANS ceils do not express the natural killer (NK) cell-specific markers NK2.1 and asialoGM1 and do not mediate lytic activity against NK-sensitive YAC-1 cells, although natural cytotoxic activity against WEHI-164 cells was found to co-purify with SPANS cells. Suppressive activity of SPANS cells in mixed lymphocyte reactions (MLR) is abolished by treatment with mitomycin C, suggesting that natural suppression in this system is a proliferation-dependent phenomenon. Preincubation of SPANS cells with conditioned medium from Con ACorrespondence to." D.W. Hoskin, Department of Microbiology, Sir Charles Tupper Medical Building, Dalhousie University, Halifax, Nova Scotia, Canada, B3H 4H7. 0165-0378/92/$05.00 © 1992 Elsevier Scientific Publishers Ireland Ltd. Printed and Published in Ireland

104

stimulated T cell cultures results in augmented NS activity, indicating that SPANS cells respond to T cell signals. Our data suggest that SPANS cells mediate suppression via the elaboration of a soluble suppressor factor since SPANS cells do not require cell-cell contact to mediate suppression and supernatants from short-term cultures of SPANS cell-enriched SBA + pregnancy spleen cells inhibit MLR. We believe that SPANS cells may be important in regulating hematopoiesis and maternal anti-fetal immunity during murine pregnancy.

Key words: murine syngeneic pregnancy; natural suppressor activity; doublenegative T cells

Introduction Natural suppressor (NS) cells (reviewed in Maier et al., 1989) are MHCunrestricted regulatory cells with non-specific inhibitory activity for immune responses. They are characteristically found in areas of intense hematopoiesis such as the bone marrow of adult animals (Duwe and Singhal, 1979; Corvese et al., 1980), neonatal murine spleen (Hooper et al., 1986; Jadus and Parkman, 1986) and in murine splenic tissue following induction of chronic graft-versus-host disease (Holda et al., 1985) or after total lymphoid irradiation (Oseroff et al., 1984) and after treatment with cyclophosphamide (McIntosh et al., 1982). Lymphoid cells with NS activity have also been described in the spleens of pregnant mice (Hoskin et al., 1989), as well as locally at the maternal-fetal interface (Slapsys and Clark, 1983). NS cells have been most often reported as having a 'null' phenotype, as they generally do not express cell-surface markers associated with mature T cells, B cells, or macrophages (Duwe and Singhal, 1979; McIntosh et al., 1982; Oseroff et al., 1984). However, other investigators have variously categorized these inhibitory cells as T cells (Strober et al., 1989), macrophage precursors (May et al., 1983), or pluripotent hematopoietic stem cells (Suguira et al., 1988). The precise lineage of NS cells is therefore controversial. Although it has been suggested that NS cells are members of the large granular lymphocyte family (Maier et al., 1989), they differ from NK cells in that they fail to lyse NK-sensitive targets such as YAC-1 lymphoma cells (Oseroff et al, 1984; Jadus and Parkman, 1986). However, Jadus and Parkman (1986) have reported that neonatal spleen-derived NS activity is associated with natural cytotoxic (NC) effector function against WEHI-164

105

cells. NS cells have been shown to inhibit a range of in vitro immune responses, including mixed lymphocyte reactions (MLR) (Oseroff et al., 1984; Hooper et al., 1986) and the generation of cytotoxic T lymphocytes (Slapsys and Clark, 1983; Choi et al., 1988). At least some NS cell populations have been shown to exert inhibitory activity through the elaboration of soluble suppressor factors (Duwe and Singhal, 1979; Hertel-Wulff and Strober, 1988; Clark et al., 1988). During a mammalian pregnancy, the mother harbours what is essentially a successful allograft in the form of the fetus and its surrounding fetally derived trophoblast tissues (Billington et al., 1983). Although the fetoplacental unit is in intimate contact with maternal tissues, it nevertheless escapes attack by the maternal immune system. Since pregnancy-associated NS cells exert potent immunosuppressive effects regardless of parity status (Clark, 1985), we (Hoskin et al., 1989) and others (Clark et al., 1988) have suggested that these immunoregulatory cells may play a crucial role in maintaining the maternal-fetal relationship. It is therefore important to better define the nature of NS cell populations that are unique to pregnant animals. In a previous report, we have characterized splenic NS cells associated with syngeneic murine pregnancy as being low density cells devoid of many conventional T cell, B cell and macrophage surface markers and therefore of undetermined lineage (Hoskin et al., 1989). Here we present evidence that splenic pregnancy-associated NS (SPANS) cells are in fact predominantly immature T cells with a CD3+4-8-45R ÷ phenotype. In addition, we characterize the mode of action and some of the biological properties of SPANS cells. Materials and Methods

Mice Six-eight week-old male and female CBA/J and male DBA/2J mice were purchased from the Jackson Laboratory (Bar Harbour, ME). Outbred Swiss Webster mice were obtained from Charles River Canada (St. Constant, QUE). Matings between male and female CBA/J mice were allowed to occur. The day on which a vaginal plug was observed was denoted the first day of pregnancy. Most experiments employed pregnant CBA/J mice 10-20 weeks of age which were killed from day 19-21 of pregnancy. Age-matched virgin CBA/J females or males were used as experimental controls. Lymphocyte preparation and M L R Spleens were removed aseptically from groups of 3-5 animals and pooled single cell suspensions were prepared in cold phosphate buffered saline (PBS, pH 7.2). The cells were washed twice with PBS and assayed for viability by

106

trypan blue dye exclusion. Cell preparations used in experiments were never less than 95% viable. Following experimental manipulations, lymphocytes were adjusted to the desired concentration in R P M I 1640 medium (Flow Laboratories, Mississauga, ONT) supplemented with 5 x 10 -5 M 2-mercaptoethanol (2-ME) (Sigma, St. Louis, MO), 10 mM L-glutamine, 100/zg/ml streptomycin, 100 units/ml penicillin and 10 mM N-2-hydroxyethylpiperazine-N-2-ethanesulfonic acid (HEPES), all Flow Laboratories, hereafter referred to as complete RPMI. MLR consisted of 2.5 x 105 CBA/J spleen cells responding to an equal number of mitomycin C-inactivated DBA/2J or Swiss Webster spleen cells in a 0.2 ml volume of complete RPMI plus 0.5% fresh normal mouse serum from virgin female or male CBA/J mice. Medium was routinely supplemented with normal mouse serum instead of fetal calf serum (FCS) to eliminate possible mitogenic effects by heterologous serum additives (Shustik et al., 1976). Mitomycin C inactivation of spleen cells was performed by a 30-min incubation at 37°C in 25 #g/ml mitomycin C (Sigma) followed by extensive washing in PBS. Regulatory cells from pregnant CBA/J mice were added to MLR in a 1:1 ratio with responder cells. All cultures were performed in quadruplicate in 96-well round-bottom Linbro microtiter plates (Flow Laboratories) and were maintained at 37°C in a 5% CO~ humidified atmosphere for 96-120 h. Six hours prior to harvesting, the cultures were pulsed with 0.5 ~Ci per well of tritiated thymidine ([3H]TdR): (spec. act. 65 Ci/mmol; ICN Radiochemicals, Irvine, CA). Cells were harvested onto glass fibre mats with a Titertek multiple sample harvester (Flow Laboratories). [3H]TdR incorporation was measured in an lsocap/300 automated counter (NCS Instrumentation Maritimes, Halifax, NS) using standard liquid scintillation counting procedures. Peak [3H]TdR incorporation routinely occurred between 96 and 120 h of culture. Data from replicate cultures are expressed as mean c.p.m. 4- S.E.M.

Soybean agglutinin (SBA ) fractionation Equal volumes of a spleen cell suspension at 4 x 108 cells/ml and SBA at 2 mg/ml (Vector Labs., Burlingame, CA) were mixed and incubated at room temperature for 30 min. The cell suspension was then gently layered onto 50% heat-inactivated FCS (Gibco Labs., Grand Island, NY) in PBS in a 15-ml polystyrene tube. The agglutinated cells (SBA ÷) were allowed to sediment to the bottom of the tube for 30 min before the top fraction containing the SBA- cells was removed. The middle fraction containing a mixture of cell types was discarded. The cell fractions were washed three times with a 0.2 M solution of the competing sugar, galactose (Sigma) in PBS, in order to dissociate lectin-aggregated lymphocytes and then three times with PBS prior to use. With this procedure, 45-60% of the original spleen cell population was recovered with a cell viability of greater than 95%. The yield of SBA + cells ranged from 40-75% while that of SBA- cells was 20-40%.

107

Wheat germ agglutin& (WGA) fractionation The procedure used for W G A fractionation of murine spleen cells was essentially the same as that used in SBA fractionation with the following exceptions. WGA (Vector Labs.) was used at a concentration of 50/~g/ml while the competing sugar used was N-acetyl-D-glucosamine (Sigma). This procedure allowed the recovery of 30-40% of the original spleen cell population with a cell viability of over 95%. The yield of WGA + cells was in the order of 65-75% whereas that of the W G A - cells ranged from 25-35%.

Negative and positive selection of lymphocytes Hybridomas producing monoclonal antibodies (mAb) to the heat-stable J1 ld.2 antigen (Bruce et al., 1981) and CD4 (anti-L3T4) (Wilde et al., 1983), which were used as undiluted culture supernatants, were obtained from the American Type Culture Collection (Rockville, MD). Anti-CD8 (anti-Lyt2.1) mAb was purchased from CedarLane Labs. (London, ONT). Anti-asialoGM! antiserum was obtained from Wako Chemicals (Dallas, TX). AntiCD45 and anti-Thy-l.2 mAb were obtained from New England Nuclear (Lachine, QUE) while mAb to CD45R was from Gibco BRL (Burlington, ONT). Anti-CD3 (epsilon) mAb was a kind gift from Dr. J. Bluestone (Leo et al., 1987). Rabbit antiserum to macromolecular insoluble cold globulin (MICG) was generously provided by Dr. Stephen Hauptman (1978). AntiNK2.1 antiserum was kindly supplied by Dr. Sylvia Pollack (1982). AntiSPANS cell m A b 2C1.1 was employed as previously described (Hoskin et al., 1989). Anti-asialoGMl, anti-MICG, anti-NK2.1, anti-CD45R, and antiCD8 antibodies were used at 1:20 dilution. Anti-CD3 ascites was used at 1:100 dilution while anti-Thy-l.2 and anti-CD45 were used at 1:1000 dilution. Negative selection of spleen cells was carried out using a two-step killing protocol in which spleen cells at 10 7 cells/ml were first incubated with antibodies for 30 min at room temperature, washed twice with PBS and resuspended in a 1/12 dilution of Low-Tox rabbit complement (RC) (CedarLane) in PBS. Following incubation at 37°C for 45 min, the cells were washed three times with PBS, counted and adjusted to the desired concentration according to the RC control. Cell viability was determined using trypan blue dye exclusion. Lymphocyte panning (Wysocki and Sato, 1978) was used to positively select for cell populations. Plastic 10 x 100 mm Petri plates were coated for 1 h at room temperature with 5 ml anti-hamster IgG or antirat IgG (5 ~g/ml) (Jackson I m m u n o Research, Bio/Can Scientific, Mississauga, ONT), washed extensively with PBS and blocked for 1 h at room temperature with 5 ml PBS containing 5% FCS. Spleen cells at 10 7 cells/ml were incubated with hamster anti-mouse CD3, rat anti-mouse CD4 or rat anti-mouse CD8 mAb for 1 h at room temperature, washed twice with PBS and incubated for 1 h at 4°C on plates coated with polyclonal antibodies specific for the cell-bound mAb. Non-adherent cells were collected by decant-

108

ing and gentle washing with PBS. Following extensive washing of the plates with PBS, adherent cells were recovered by vigorous pipetting. Con A-stimulated supernatants ( C A S ) CAS were prepared by incubating spleen cells (107 cells/ml) obtained aseptically from adult Wistar rats (Charles River, Canada) with 10 ~g/ml Con A (Pharmacia Canada, Baie d'Urfe, QUE) in complete RPMI supplemented with 5% FCS for 2 h at 37°C. The cells were then washed three times with PBS, resuspended in culture medium at a concentration of 5 × 106 cells/ml and incubated for 24 h at 37°C in a 5% CO~ humidified atmosphere. Following incubation, cell-free supernatants were filter-sterilized and stored at -20°C. Natural cell-mediated ~3'toto.viciO' assays 5tCr-release assays were employed to measure NK and NC activity of SBA + spleen cells using YAC-I and WEHI-164 cell lines, respectively, as target cells. Spleen cells were obtained from pregnant mice (day 19-21 of gestation) injected with 100 ~g polyinosinic polycytidilic acid (Sigma) i.p. 18 h prior to the experiment. Target cells were labelled by incubation with 100 #Ci NaeSICrO4 (ICN Radiochemicals) for 60 rain at 37°C. All ceils were washed three times with PBS prior to use. Assays were performed in 96-well V-bottom, Linbro microtiter plates (Flow Laboratories) using various E/T ratios (5000 target cells per well) in a total volume of 0.2 ml complete RPMI supplemented with 5% FCS. The plates were incubated for 4 h (NK assay) or 18 h (NC assay) at 37°C in a 5% CO~ humidified atmosphere. The plates were subsequently centrifuged at 200 × g for 4 rain and 100 ~1 of supernatant was then removed and assayed for 51Cr content using a Beckman Gamma 8000 sample counter. Percent lysis were determined as follows: % lysis = (experimental 51Cr release - spontaneous S~Cr release)/(maximum 51Cr release spontaneous 5~Cr release). Spontaneous 5~Cr release was determined by incubating 5 x 103 labelled target cells with 100 ~1 of complete RPMI supplemented with 5% FCS. Maximum 5LCr release was determined by lysing 5 x 103 labelled target cells with 100 ~1 of 10% sodium dodecyl sulphate. -

Soluble mediator assays Two different approaches were taken to determine whether SPANS cells mediate suppression via a soluble factor. Firstly, MLR consisting of 6 × l 0 6 CBA/J spleen cells responding to an equal number of mitomycin Cinactivated DBA/2J spleen cells were performed in 24-well cluster plates (Costar, Cambridge, MA) in a volume of 0.6 ml of complete RPMI supplemented with 0.5% normal mouse serum. A volume of 0.2 ml of culture medium alone, or containing 6 x 106 virgin, or pregnancy WGA + spleen cells

109

was added to Transwell chambers (Costar) which were then inserted into MLR-containing wells. After 96 h, 120 h and 144 h of incubation, triplicate 100 t~l aliquots of cells responding in M L R were removed from wells of the 24-well plate, placed in wells of a 96-well round-bottom Linbro microtiter plate (Flow Laboratories) and pulsed with 0.5/zCi [3H]TdR to measure cell proliferation. Samples were harvested and processed as described above. Secondly, cell-free supernatants were collected from virgin and pregnancy SBA + spleen cells cultured for 24 h in complete RPMI supplemented with 0.5% normal mouse serum. Supernatants were stored overnight at 4°C before being assayed for NS activity in MLR. Statistics Experiments were performed at least three times and representative data are presented. Statistical significance between the means of different culture replicates was tested using the two-tailed Student's t-test. P values less than 0.05 were considered to be significant. t.

Results

S P A N S cells possess receptors for SBA and WGA lectins In a previous study (Hoskin et al., 1989), we have shown that the NS activity present in the spleens of pregnant mice resides in a population of lymphocytes possessing receptors for the lectin, SBA. Sugiura et al. (1988) have recently reported that WGA ÷ murine BM cells possess potent NS activity. This prompted us to determine whether SPANS cells also express receptors for WGA, as well as for SBA. Table 1 summarizes the results of typical mixing experiments in which spleen cells from CBA/J mice pregnant by syngeneic matings (day 19-21 of gestation) and from virgin female CBA/J mice were fractionated by agglutination with SBA and/or W G A prior to addition to M L R to assay for NS activity. Unselected and lectin-fractionated spleen cells were added to M L R (consisting of 2.5 x 105 CBA/J spleen cells responding to 2.5 x 105 mitomycin C-inactivated allogeneic spleen cells) at initiation of culture in a 1:1 ratio with responder cells. In agreement with our earlier findings, the addition of SBA ÷ pregnancy spleen cells to M L R was found to result in a markedly suppressed proliferative response (expt. 1). No inhibitory effect was seen when SBA + spleen cells from virgin mice were added to MLR, indicating that the observed suppressive effects are pregnancy-associated. Similarly, the WGA + fraction of pregnancy spleen was observed to contain cells with strong NS activity while the WGA + fraction of virgin spleen lacked cells with substantial inhibitory activity (expt. 2). SBA ÷ and W G A ÷ spleen cells from primiparous and multiparous pregnant CBA/J mice were equally suppressive in M L R (data not shown). Moreover,

P r e g n a n t female spleen

N o cells a d d e d

P r e g n a n t female spleen

N o cells a d d e d

SBA + SBA +

WGA + WGA +

WGA + SBA + SBA-

-

SBA + WGA + WGA

-

29 420 ± 2002 27 001 ± 4107 48 105 ± 1957

40 789 ± 3845

21 998 ± 1332 26 731 ± 453 40 248 ± 3604

39 547 ± 4848

59 017 .4- 8509 146 929 ± 9021

t58 133 ± 5465

WGA WGA + WGA-

105 483 ± 5637

WGA +

" P e a k response m e a s u r e d at 96 h o f c u h u r c in expt. I a n d 2 a n d at 120 h in cxpt. 3 a n d 4. b'V,, S u p p r e s s i o n = (1 -c.p.m.cxpt/c.p.m.ctl) x 100. CDetermined b y S t u d e n t ' s t test.

4

3

P r e g n a n t female spleen

Virgin female spleen

-

76 737 ± 1097 120 182 ± 1009

SBA-

N o cells a d d e d

45 922 ± 3645

SBA +

72 873 ± 2966 60 659 -4- 6161

69 184 ± 3646

P r e g n a n t female spleen

-

[ 3 H ] T d R i n c o r p o r a t i o n ;' c.p.m. ± S . E . M .

SBA + SBA

N o cells a d d e d

1

Cell fraction

Virgin female spleen

Cells a d d e d to M L R

Expt.

Positive selection o f S P A N S cells u s i n g S B A a n d W G A lectins.

TABLE 1

28 34 0

44 32 0

51 0

0

12

-

0

34

0 12

-

".'i, S u p p r e s s i o n b

P < 0.025 P < 0.025 P > 0.05

P < 0.005 P < (I.025 P > 0.05

P < 0.001 P > 0.05

P > 0.05

P < 0.05

P > 0.05

P < 0.01

P > 0.05 P > 0.05

S i g n i f i c a n c e ~"

111

neither SBA- nor W G A - spleen cells from pregnant or virgin animals were suppressive in M L R (expt. 1 and 2). Inhibition of MLR by lectinfractionated pregnancy spleen cells was not due to overcrowded culture conditions since the addition o f an equal number of splenocytes from virgin CBA/J mice to M L R did not significantly inhibit the proliferative response. Serial agglutination o f pregnancy spleen cells with SBA and WGA (expt. 3), or vice versa (expt. 4), revealed that pregnancy spleen cells which co-express SBA and W G A receptors were suppressive in MLR while SBA+WGA - and WGA+SBA - pregnancy spleen cells lacked suppressor activity. From these data, we conclude that SPANS cells bear receptors for WGA, as well as for SBA. WGA or SBA fractionation was therefore employed to prepare SPANS cell-enriched preparations from pregnancy spleen for use in all subsequent experiments.

Kinetics of SPANS activity development during syngeneic pregnancy Figure 1 depicts a representative experiment comparing the ability of unfractionated, SBA + and S B A - pregnancy spleen cells isolated at different

180

160

X u Z 0

o

140 120

100

80

o zH

6O

~

40

I

20

VG

P7 P14 P21 PP7

WHOLE

SPLEEN

VG

P7 P14 P21 PP7

SBA*

VG

P7 P14 P21 PP7

SBA-

Fig. 1. Kinetics of S P A N S activity during the course of murine syngeneic pregnancy. Spleen cells were isolated from virgin female and pregnant mice at days 7 ( P 7 ) , 14 ( P I 4 ) a n d 21 ( P 2 1 ) of pregnancy as w e l l as 7 days post partum (PP7), fractionated by SBA agglutination and added to MLR at a 1:1 ratio with responder cells. The hatched area denotes the response range of control MLR with no cell additions. *Indicates statistical significance as determined by Student's t-test.

Pan T cell

Pan T cell

Helper T cells

Cytotoxic/suppressor T cells Immature T cells, B cells NK cells, some cytotoxic T cells

anti-MICG

anti-Thyl.2

anti-CD4

anti-CD8

36 46 38 36 21 38 58 56 49 56 49 36 56 36 32 53 38 32 53

Untreated 3 17 17 0 2 38 65 44 48 44 42 10 15 24 40 58 31 25 48

Treated

% Suppression of MLR by SBA+pregnancy spleen cells a

P P P P P P P P P P P P P P P P P P P

< < < < < > > > > < > < < > :> > < < >

0.005 0.005 0.005 0.001 0.05 0.05 0.05 0.05 0.05 0.01 0.05 0.01 0.001 0.05 0.05 0.05 0.05 0.05 0.05

Significance b

No

No

Yes

No

No

No

Yes

Yes

Marker present

aThe results of replicate experiments are shown. bMean c.p.m, of replicate cultures containing treated (Ab plus RC) and untreated (RC alone) regulatory cells were compared by Student's t test.

anti-NK2.1

anti-asialoGMl

NK cells

Pan leukocyte

anti-CD45

anti-Jlld.2

Specificity

Antibody

E f f e c t o f c y t o t o x i c p r e t r e a t m e n t w i t h a n t i b o d i e s t o c e l l - s u r ~ c e m a r k e r s o n S P A N S a c t i v i t y i n MLR.

TABLE2

t~

113

o)

3H-TdR INCORPORATION(cOrnX 10-3) 20 40 60 80 I

ADDITIONSTO MLR MEDIA

UNTREATED PG. SPL. CELLS

1 I

PG. CD3÷ SPL. CELLS

PG. CD3- SPL. CELLS

b)

3H-TdR INCORPORATION(cpm X 10-3) ADDITIONSTO MLR

0

20

40

60

80

100

120

MEDIA

UNTREATED PG. SPL. CELLS

PG. CD4÷ SPL. CELLS

PG. CO4- SPL. CELLS

¢) ADDITIONSTO MLR

I 3H-TdR INCORPORATION(cpm X 10-3) 50 100 150

MEDIA

UNTREATED PG. SPL. CELLS

PG. CD8÷ SPL. CELLS

PG. CD8- SPL. CELLS Fig. 2. NS activity of positively selected spleen cell populations from pregnant mice. Unselected and adherent and nonadherent pregnancy spleen (PG.SPL.) cell populations obtained by panning with (a) antiCD3; (b) anti-CD4; and (c) anti-CD8 mAb were added to MLR at a 1:1 ratio with responder cells. *Denotes statistically significant results compared to control MLR with no cell additions as determined by Student's t-test.

114

time points in gestation to inhibit MLR. Suppressive activity was absent in all virgin control fractions tested. Significant inhibition o f M L R by unfractionated pregnancy spleen cells was seen by day 14 of pregnancy and the suppressive activity peaked at day 21 of gestation. Subsequent experiments therefore employed pregnant CBA/J mice at day 19-21 of gestation. In the SPANS cell-enriched SBA + fraction, similar trends were seen but significant inhibitory activity was detectable by day 7 of gestation. The suppressive activity of both unfractionated and SBA + pregnancy spleen cells was markedly diminished by 7 days post-partum. In agreement with the data reported in Table 1, inhibitory activity was not observed at any time point in the SBA- fraction of pregnancy spleen. Evidence that S P A N S cells are predominantly T cells Our earlier findings suggested that SPANS cells do not express cell-surface markers associated with mature T cells, B cells, or macrophages (Hoskin et al., 1989). However, given that the range of antibodies employed in previous negative selection studies was rather modest, we decided to confirm and extend our initial results using a larger panel of antibodies. SBA + spleen cells from pregnant mice were subjected to treatment with an array of polyclonal and monoclonal antibodies plus RC prior to addition to culture. The data shown in Table 2 indicate that cytotoxic pretreatment with polyclonal antiserum to MICG, which is selectively expressed by T cells (Hauptman et al., 1978), eliminates the suppressive activity of the SPANS cell-enriched SBA + pregnancy spleen cells, suggesting that SPANS cells are in fact members of the T cell lineage. Interestingly, SPANS cells also express J1 ld.2, an antigen that is present on immature T cells and most B cells (Bruce et al., 1981), since anti-Jlld.2 plus RC treatment of SBA + pregnancy spleen cells removes most NS activity. SPANS activity was also reduced by cytotoxic pretreatment with mAb to CD45. However, cytotoxic pretreatment of SBA + pregnancy spleen cells with m A b to the T cell markers Thy-l.2, CD4 and CD8 did not diminish the ability of SPANS cells to inhibit MLR, indicating that these structures are absent or expressed at very low density on SPANS cells. Treatment of SBA + pregnancy spleen cells with NK-specific antisera (anti-asialoGMl, anti-NK2.1) plus RC had no effect on NS activity, suggesting that SPANS cells are not N K cells. Lymphocyte panning of pregnancy spleen cells with mAb to CD3, CD4 and CD8 was used in the next series of experiments to confirm and extend the above characterization data regarding the T cell nature of SPANS cells. We found that pregnancy spleen cells with NS activity in M L R did not adhere to anti-CD4 or anti-CD8-coated plates, confirming that SPANS cells express neither CD4 nor CD8 markers (Fig. 2). Surprisingly, NS activity was evident in both the adherent CD3 + and non-adherent CD3 fraction of

115

3H-TdR INCORPORATION(cprn X 10-3) ADDITIONS TO MLR MEDIA

TREATMENT 0

15

30

45

60

75

MEDIA

RC CD3+ PG. SPL.

CD3PG. SPL.

°°

ANTI-J1 ld +RC ANTI~3D45R+ RC

ANTI~I1ld + RC ANTI~3D45R+ RC

---'1

Fig. 3. Effect of cytotoxic pretreatment with anti-Jlld.2 and anti-CD45R m A b on CD3 + and C D 3 SPANS cells. Spleen cells from pregnant mice were separated into CD3 ÷ and C D 3 - fractions by panning with anti-CD3 m A b and cells were then treated with anti-J1 ld.2 and anti-CD45R m A b plus RC prior to addition to M L R in a 1:1 ratio with responder cells. *Indicates statistical significance; ** lack of significance compared to control M L R with no cell additions as determined by Student's t-test.

pregnancy spleen, suggesting that SPANS cells may consist of T cell and non-T cell components. To further clarify this observation, we treated CD3 ÷ and CD3- pregnancy spleen cells with anti-J1 ld.2 and anti-CD45R mAb plus RC prior to adding the regulatory cells to MLR. As shown in Fig. 3, CD3 + SPANS cells were sensitive to cytotoxic pretreatment with both anti-J1 ld.2 and anti-CD45R mAb while CD3- SPANS cells were resistant. Since both anti-J1 ld.2 and anti-MICG plus RC treatments relieved most of the natural suppression mediated by SBA + pregnancy spleen cells (Table 2), we conclude that SPANS cells are predominantly SBA receptor-bearing J1 ld.2÷CD3+4-8-45R + T cells. However, the SBA + fraction of pregnancy spleen also appears to contain a minor population of J1 ld.2-CD3-4-8-45 R- 'null' cells with suppressor activity.

Mitomycin C treatment abrogates inhibition of MLR by SPANS cells Mitomycin C is a drug which chemically crosslinks DNA, thereby inhibiting DNA replication (Franklin and Snow, 1981). In the next experiments, SBA + pregnancy spleen cells were treated with mitomycin C prior to addition to MLR to determine whether DNA replication is required for the generation of SPANS activity. As shown in Fig. 4, mitomycin Ctreated SBA + pregnancy spleen cells were unable to inhibit MLR while the addition of untreated SBA ÷ cells to MLR resulted in a 72% inhibition of

8O z

_o

6° rr~

0 x

"+

20

175

-

150

"

125

-

100

-

75

-

"1--

5025-

Fig. 4. Effect of pretreatment with mitomycin C on NS activity of SBA + pregnancy spleen cells. SPANS cell-enriched SBA ÷ spleen cells obtained from pregnant mice were left untreated (hatched bar) or treated with mitomycin C (solid bar) prior to addition to MLR in a 1:1 ratio with responder cells. The open bar denotes the response of control MLR with no cell additions: *indicates statistically significant results compared to control MLR as determined by Student's t-test. Fig. 5. Comparison of NS activity by unstimulated and CAS-stimulated SPANS cells in MLR. SPANS cell-enriched WGA + spleen cells obtained from pregnant mice were incubated for 18 h in the presence (solid bar) or absence (hatched bar) of 25% CAS prior to addition to MLR in a 1:1 ratio with responder cells. The open bar denotes the response of control MLR with no cell additions: *indicates statistically significant results compared to control MLR as determined by Student's t-test.

the response. These data indicate that S P A N S cells must proliferate in order to mediate suppression.

Suppression by SPA NS cells is enhanced by CA S Several investigators working with NS cells in other experimental systems have reported that NS activity is enhanced by CAS (Holda et al., 1986: Subiza et al., 1989). We were therefore interested to determine whether CAS has a similar effect on S P A N S cells. W G A + cells obtained from pregnancy spleen were cultured for 18 h in the presence or absence of 25% CAS. Loss of viability was minimal and equivalent numbers of cells were recovered from CAS-supplemented and control cultures. Following extensive washing, the W G A ÷ pregnancy spleen cells were added to M L R to determine what, if any, effect pretreatment with CAS had on their inhibitory capacity. As shown in Fig. 5, W G A ÷ pregnancy spleen cells which had been preincubated with CAS were approximately twice as suppressive as the untreated controls (66% vs. 32% suppression), indicating that SPANS activity is enhanced by T cell signals. NS and NC activity copurifv in SPA NS cell-enriched SBA +pregnancy spleen cell fractions We next examined S P A N S cells for natural cell-mediated cytotoxicity

117

TABLE 3

NS and NC activity coexist in SPANS cell-enriched SBA ÷ pregnancy spleen cell fractions. Treatments a

E/T

% Lysis

YAC-I

Untreated R C alone mAb 2C1.1 + R C

WEHI-I~

100:1

50:1

25:1

25:1

12:1

6:1

34 37

24 26

15 18

23 21

13 16

11 8

32

23

17

13

9

6

aSBA+ spleen cells from pregnant CBA/J mice were subjected to negative selection as described in Materials and Methods.

since other investigators have reported that some NS cell populations possess cytolytic potential (Jadus and Parkman, 1986). YAC-1 cells were employed to assess N K activity of SPANS cells while WEHI-164 cells were used to assay for NC activity. As shown in Table 3, both YAC- 1 and WEHI- 164 cells were lysed by SBA + pregnancy spleen cells, indicating the presence of cells with N K and NC activity in this spleen cell fraction. Anti-SPANS cell mAb 2C1.1 (Hoskin et al., 1989) was therefore used to deplete SBA + pregnancy spleen cells of cells with NS activity prior to assaying for cytolytic activity. It was found that SPANS cell-depleted SBA + pregnancy spleen cells did not differ from untreated cells in terms of their cytolytic activity against YAC- 1 targets, indicating that SPANS cells do not contribute to the observed cytolysis of YAC-1 targets and are therefore devoid of N K activity. However, a substantial decrease in NC activity was observed when SBA + pregnancy spleen cells were pretreated with mAb 2C 1.1 plus RC, suggesting a possible contribution by SPANS cells to NC activity.

40 0

~ Q O ~ ~Z X

20

H

~ ~

cu u

~

10

o

It

Fig. 6. Cell-cell contact is not required for SPANS cells to mediate suppression. Culture medium (open bar), W G A + virgin spleen cells (hatched bar), or W G A + pregnancy spleen cells (solid bar) were placed in culture well-inserts containing cell impermeable membranes which were then placed in MLR. *Denotes statistically significant results compared to control MLR with culture medium added as determined by Student's t-test.

118

S P A N S cells mediate natural suppression via a soluble suppressor factor Studies by other investigators (Duwe and Singhal, 1979; Clark et al., 1988; Hertel-Wulff and Strober, 1988) have shown that NS cell populations in some experimental systems mediate suppression through the production of soluble suppressor factors. To determine if this is also the case with SPANS cells, experiments were performed using culture well-inserts consisting of a chamber with a microporous membrane floor to physically separate SPANS cells from cells responding in MLR. As shown in Fig. 6, the addition of WGA + pregnancy spleen cells to the chamber resulted in 46% inhibition of MLR while the presence of control WGA + spleen cells from virgin mice had no effect on MLR. These data indicate that SPANS cells do not require cellcell contact to mediate suppression but rather produce a soluble inhibitory factor capable o f diffusing across the microporous culture well-insert membrane. To further substantiate this data, we next examined the ability of 24 h supernatants from unstimulated SBA + pregnancy spleen cell cultures to inhibit cellular proliferation in MLR. As shown in Fig. 7, addition of SBA + pregnancy spleen cell culture supernatants (final concentration of 50%) to MLR at initiation of culture resulted in a 30% inhibition of MLR. Note that a twofold dilution of the SBA + pregnancy spleen cell culture supernatants

150

--

o X 0 Z

~'~//ffJJfJ/JJJ///JJfJ//~ lOO

0 H (9%)

(9%) (12%)

0 0 o z

(30%) 50

I

%

i

i

i

i

i0

20

30

40

50

SUPERNATANT

ADDITIONS

FINAL

CONCENTRATION

i

Fig. 7. Supernatants from short-term cultures o f SBA + pregnancy spleen cells inhibit MLR. Supernatants from 24 h cultures o f SBA + pregnancy spleen cells (O) or SBA + virgin spleen cells ( 0 ) were added to M L R at the stated concentrations. (%) Indicates percent inhibition compared to the mean response o f control M L R denoted by the hatched area. *Denotes statistical significance as determined by Student's t-test.

119

(final concentration of 25%) decreased suppressor activity by approximately 50%, indicating that the inhibitory effect of the putative suppressor factor on M L R is concentration-dependent. Addition of control supernatants derived from 24 h cultures of SBA ÷ virgin spleen cells did not significantly inhibit MLR. These data indicate that supernatants from short term SBA ÷ pregnancy spleen cell cultures are inhibitory for MLR, suggesting that SBA ÷ SPANS cells elaborate a soluble suppressor factor. Discussion

In this study, we present evidence that NS cells found in the spleens o f pregnant mice belong predominantly to the T cell lineage. In an earlier report, we demonstrated that SPANS cells are devoid of many cell-surface markers normally associated with mature T cells, B cells and macrophages but express receptors for the lectin SBA (Hoskin et al., 1989). The SBA ÷ nature of SPANS cells initially led us to tentatively assign them to an early B cell lineage since B cells are uniformly SBA + while mature T cells are SBA- (Reisner et al., 1976). However, we show here that the majority of SBA + SPANS cells are in fact immature T cells with a CD3+4-8-45R + J 1 l d.2 + phenotype since (1) treatment with T cell-specific anti-MICG antiserum ([Hauptman et al., 1978) plus RC eliminates virtually all NS activity in the SBA ÷ fraction of pregnancy spleen; (2) SBA ÷ SPANS cells are sensitive to cytotoxic pretreatment with m A b directed against J1 l d.2, a heatstable differentiation antigen expressed by immature T cells but not by mature T cells (Bruce et al., 1981); (3) immature thymic T cells are known to possess SBA receptors (Sharon, 1983); (4) anti-CD4 and anti-CD8 plus RC treatment has no effect on suppressor activity, which is also shown by lymphocyte panning to reside in the CD4- and CD8- fraction of pregnancy spleen and (5) CD3 + pregnancy spleen cells (positively selected by panning), which also bear J l l d . 2 and CD45R structures, are highly suppressive in MLR. Double-negative splenic T cells with NS activity that are associated with murine pregnancy are phenotypically and functionally similar to a subset of CD3+4-8-45R + T suppressor cells recently identified in human cord blood which down-regulate autoreactive T cell responses to self-Ia in autologous M L R by suppressing IL-2 production (Kawano et al., 1990). Indeed, we have found that SPANS cells also act by down-regulating IL-2 synthesis in M L R (manuscript in preparation). T suppressor cells with a CD3+4-8 - phenotype and non-specific activity have also been reported in the spleens of mice immunized against Mls a determinants (Bruley-Rosset et al., 1990) and have been cloned from the spleens of neonatal mice and mice treated with total lymphoid irradiation (Strober et al., 1989). Certain subsets of double-negative T cells may therefore function as important im-

120

munoregulatory elements in vivo. At present it is not known whether the T cell receptor is involved in the activation and function of T cell lineage NS cells, although given the apparent lack of antigen specificity associated with NS cells it seems likely that the T cell receptor would have to recognize a common ligand. Despite the T cell phenotype of most SPANS cells, we have been unable to demonstrate the presence of cell-surface Thy-1 molecules by negative selection. Similarly, NS cells freshly isolated from spleens of mice treated with total lymphoid irradiation do not express detectable levels of Thy-1 antigen (Oseroff et al., 1984), even though clonal T suppressor cell lines derived from these isolates are strongly Thy-1 + (Strober et al., 1989). One possible explanation for this is that freshly isolated T cell lineage NS cells express Thy-I but only at low levels and are therefore resistant to depletion with antiThy-1 plus RC. However, certain T cell populations are known to lack detectable Thy-1 (Harriman et al., 1990). Thy-I may therefore be an activation marker which is absent on native T cell-derived NS cells, but is acquired following prolonged culture in IL-2-containing medium. This would be consistent with the finding that lymphokine-activated killer (LAK) cells, generated by culturing N K cells in the presence of IL-2, express Thy- 1 which their progenitor cells lack (Mul6 et al., 1986). SPANS cells are characterized by co-expression of receptors for the lectins, SBA and WGA. Although NS activity has recently been associated with WGA + cycling hematopoietic stem cells (Suguira et al., 1988), the presence of the pan-T cell markers CD3 and M I C G on the majority of SPANS cells indicates that they are not stem cells. Thus, our data is more in line with a recent report by Skyes et al. (1990) demonstrating that NS cells present in bone marrow and in the spleens of irradiated, bone marrow-reconstituted mice are not pluripotent hematopoietic stem cells. However, panning pregnancy spleen cells with anti-CD3 mAb revealed the existence of a minor population of C D 3 - 4 5 R - J l l d . 2 - cells which also exhibit NS activity in MLR. The presence of these null suppressor cells accounts for the residual NS activity remaining in the SBA ÷ fraction of pregnancy spleen following anti-Jlld.2 and anti-CD45 plus RC treatments. While it is possible that CD3-45R-J1 ld.2- SPANS cells are proliferating stem cells, it is also conceivable that this small population of null NS cells is representative of an intermediate stage in SPANS cell differentiation. In support of this, although T cell development is thought to occur primarily in the thymus, extrathymic T cell differentiation from CD3- precursors has recently been documented (Sydora and Kronenberg, 1991; Benveniste et al., 1990). Like non-specific T suppressor cells activated by PHA (Clement et al., 1983), SPANS cells require proliferation-dependent activation to express suppressive activity. Thus, mitomycin C treatment of SPANS cells blocks

121

their inhibitory function. This result is in agreement with an earlier finding that suppressive activity of SPANS cells is abolished by irradiation (1500 rad, in vitro) prior to addition to culture (Hoskin and Murgita, unpublished data). This is also in line with a report by Corvese et al. (1980) that the NS activity of bone marrow cells is abolished by irradiation. Taken together, these data are consistent with a proposed induction pathway of NS cell activity suggested by Maier and coworkers (1989) wherein NS cells must proliferate in response to T cell signals in order to express their full suppressive potential. Indeed, SPANS cells which have been exposed to CAS subsequently display enhanced suppressor activity. Our results indicate that cellcell contact is not required for SPANS cells to mediate inhibitory effects, suggesting that suppression is mediated via a soluble factor. The finding that supernatants from short-term cultures of SBA + pregnancy spleen cells inhibit M L R provides additional evidence that SPANS cells produce a soluble suppressor factor. Thus, SPANS cells are similar to NS cells described m other experimental systems which produce immunosuppressive factors independently of antigenic stimulation (Duwe and Singhal, 1979; Clark et al., 1988; Hertel-Wulff and Strober, 1988). It has been suggested that NS cells are members of the large granular lymphocyte family of regulatory cells which also includes NK and NC ceils (Maier et al., 1989). However, depletion of SPANS cells by treatment with mAb 2C1.1 plus RC does not diminish the ability of SBA + pregnancy spleen cells to lyse NK-sensitive YAC-1 cells, indicating that SPANS cells do not contribute to NK activity. Moreover, NS activity is not removed by cytotoxic pretreatment with antisera to NK-associated antigens such as asialoGM1 and NK2.1. SPANS cells are therefore distinct from NK cells. This is in agreement with findings by other investigators that NS activity and NK activity are mediated by distinct populations of effector cells (Oseroff et al., 1984; Jadus and Parkman, 1986). However, when SBA + pregnancy spleen cells were treated with anti-SPANS mAb 2C1.1 plus RC, a significant decrease in cytolytic activity against WEHI-164 targets was observed, indicating a decrease in NC activity. While it is possible to interpret this data to mean that mAb 2C1.1 recognizes a common epitope shared by SPANS cells and a minor NC population which coexist in the SBA + spleen fraction, a more likely explanation is that SPANS cells also possess NC potential. This would be consistent with the recent finding that, upon activation, cloned CD3+4-8 - NS cells derived from splenic tissue of neonates and mice treated with total lymphoid irradiation secrete tumor necrosis factor-alpha which is known to mediate NC activity against WEHI-164 target cells (Van Vlasselaer and Strober, 1991). Cells with NS activity have been documented in adult bone marrow (Duwe and Singhal, 1979; Corvese et al., 1980), neonatal murine spleen (Hooper et

122

al., 1986; Jadus and Parkman, 1986), in the spleens of mice undergoing chronic graft-versus-host disease (Holda et al., 1985) and in murine splenic tissue following total lymphoid irradiation (Oseroff et al., 1984) and treatment with cyclophosphamide (McIntosh et al., 1982). All of these are sites of intense hematopoiesis which has led to the suggestion that one function of NS cells is to regulate cell proliferation during the process of hematopoiesis (Maier et al., 1989). During murine pregnancy the spleen enlarges and the nucleated cell count increases (Maroni and de Sousa, 1973; Gambel et al., 1980). Increased spleen size has been attributed to the proliferation of hematopoietic stem cells in the red pulp (Maroni and de Sousa, 1973), indicating that during pregnancy the spleen becomes a site of limited hematopoiesis. It has been suggested that splenic hematopoiesis during pregnancy results from greatly elevated levels of uterine colony-stimulating factor 1 (Bartocci et al., 1986). Interestingly, maternal spleen cell counts and the actual size of the spleen are greatest at 12 days of gestation (Gambel et al., 1980) which corresponds to the time when SPANS cell activity first becomes detectable. This suggests the intriguing possibility that one function of SPANS cells is to down-regulate hematopoiesis occurring in the splenic environment during pregnancy. The survival of the fetus in a potentially hostile maternal immunologic environment is often attributed to a combination of local (Clark, 1985) and systemic (Murgita and Wigzell, 1981) immunosuppressive mechanisms. The ability of some NS cell populations to inhibit the production and/or utilization of IL-2 by cytotoxic effector cells (Clark et al., 1988; Maes et al., 1988) suggests that during pregnancy NS cells may be involved in protecting fetal tissues which are vulnerable to attack by IL-2-activated MHC-unrestricted NK cells and cytotoxic T lymphocytes, otherwise known as LAK cells (Drake and Head, 1989). Activation of LAK cells could occur either at the maternal-fetal interface or in the spleen from whence LAK cells could gain access to the fetoplacental unit via its blood supply. The nature of immunosuppression at the maternal-fetal interface remains controversial since Clark et al. (1988) report that decidual NS cells produce a transforming growth factor-beta-like molecule which prevents induction of LAK activity by inhibiting IL-2 uptake while Lala et al. (1990) indicate that inactivation of decidual NK cells occurs via a PGE2-mediated mechanism. We believe that SPANS cells may inhibit the systemic generation of LAK cells with potentially dangerous anti-fetal activity as a result of IL-2 release by maternal T cells responding to oncofetal and other autoantigens (Hoskin and Murgita, 1985; 1989). In support of this hypothesis, we have shown that in vivo administration of anti-SPANS cell mAb to mice pregnant by syngeneic matings abrogates splenic NS activity and leads to fetal resorption and/or decreased litter size (Gronvik et al., 1987).

123

In summary, the results of this study extend our previous work (Hoskin et al., 1989) and indicate that maternal spleen harbours a unique population of immature T cells with potent NS activity which is mediated through the production of a soluble suppressor factor. Ongoing studies seek to examine the physicochemical nature of this suppressor factor. Acknowledgements The authors are grateful to Mr. N. Rowley for graphics and Angelique Sherwood for expert secretarial assistance. This study was supported by grants from the Dalhousie Medical Research Foundation and the Natural Sciences and Engineering Research Council of Canada. J.C. Brooks-Kaiser is the recipient of a studentship from the Medical Research Council of Canada. References Bartocci, A., Pollard, J.W. and Stanley, E.R. (1986) Regulation of colony-stimulating factor 1 during pregnancy. J. Exp. Med. 164, 956-961. Benveniste, P., Chadwick, B.S. and Miller, R.G. (1990) Development of T cells in vitro from precursors in mouse bone marrow. Cell. lmmunol. 127, 92-104. Billington, W.D., Bell. S.C. and Smith, G. (1983) Histocompatibility antigens of mouse trophoblast of significance in maternal-fetal immunological interactions. In: Immunology of Reproduction (Wegmann, T.G. and Gill, T.J., eds.), pp. 205-227. Oxford University Press, London and New York. Bruce, J., Symington, F.W., McKearn, T.J. and Sprent, J. (1981) A monoclonal antibody discriminating between subsets of T and B cells. J. Immunol. 127, 2496-2501. Bruley-Rosset, M., Miconnet, 1., Canon, L. and Halle-Pannenko, O. (1990) Mls a generated suppressor cells. 1. Suppression is mediated by double-negative (CD3+CD5+CD4-CD8-) ~/~ T cell receptorbearing cells. J, Immunol. 145, 4046-4052. Choi, K.L., Maier, T., Holda, J.H. and Claman, H.N. (1988) Suppression of cytotoxic T cell generation by natural suppressor cells from mice with GVHD is partially reversed by indomethacin. Cell. Immunol. 112, 271-278. Clark, D.A. (1985) Materno-fetal relations, lmmunol. Lett. 9, 239-247. Clark, D.A., Falbo, M., Rowley, R.B., Banwatt, D. and Stedronska-Clark, J. (1988) Active suppression of host-vs-graft reaction in pregnant mice. IX. Soluble suppressor activity obtained from allopregnant mouse decidua that blocks the cytolytic effector response to IL-2 is related to transforming growth factor-/3. J. lmmunol. 141, 3833-3839. Clement, L.T., Dagg, M.L., Lehmeyer, J.E. and Kiyotaki, M, (1983) Two phenotypically distinct suppressor T cell subpopulations inhibit the induction of B cell differentiation by phytohemagglutinin. J. Immunol. 131, 1214-1217. Corvese, J.S., Levy, E.M., Bennett, M. and Cooperband, S.R. (1980) Inhibition of an in vitro antibody response by a suppressor cell in normal bone marrow. Cell. lmmunol. 49. 293-306. Drake, B.L. and Head, J.R. (1989) Murine trophoblast can be killed by lymphokine-activated killer cells. J. Immunol. 143, 9-14. Duwe, A.K. and Singhal, S.K. (1979) The immunoregulatory role of bone marrow, il. Characterization of a suppressor cell inhibiting the in vitro antibody response. Cell. lmmunol. 43, 372-381. Franklin, T.J. and Snow, G.A. (1981) Biochemistry of Antimicrobial Action, Chapman and Hall, London and New York.

124 Gambel, P,I,, Cleland, A.W. and Ferguson, F.G. (1980) Alterations in thymus and spleen cell populations and immune reactivity during syngeneic pregnancy and lactation. J. Clin. Lab. lmmunol. 3, 115-119. Gr6nvik, K.-O., Hoskin, D.W. and Murgita, R.A. (1987) Monoclonal antibodies against murine neonatal and pregnancy-associated natural suppressor (NS) cells induce resorption of the fetus. Scand. J. Immunol. 25, 533-540. Harriman, G.R., Lycke, N.Y., Elwood, L.J. and Strober, W. (1990) T lymphocytes that express CD4 and the ~13-T cell receptor but lack Thy-l. Preferential localization in Peyer's patches. J. lmmunol. 145, 2406- 2414. Hauptman, S.P., Sobczak, G. and Guttman, I.A. (1978) Macromolecular insoluble cold globulin (MICG): a novel protein from mouse lymphocytes. 11. T-cell origin of MICG and response to mitogens. Immunochemistry 15, 423-428. Hertel-Wulff, B. and Strober, S. (1988) immunosuppressive lymphokine derived from natural suppressor cells. J. Immunol. 140, 2633-2638. Holda, J.H., Maier, T. and Claman, H.N. (1985) Graft-vs-host reactions (GVHR) across minor histocompatability barriers. 1. Impairment of mitogen responses and suppressor phenomena. J. lmmunol. 134, 1397-1402. Holda, J.H., Maier, T. and Claman, H.N. (1986) Natural suppressor activity in graft-versus-host spleen and normal bone marrow is augmented by IL-2 and interferon-r, J. lmmunol. 137, 3538-3543. Hooper, D.C., Hoskin, D,W., Gr6nvik, K.-O. and Murgita, R.A. (1986) Murine neonatal spleen contains both T and non-T suppressor cells capable of inhibiting adult alloreactive and newborn autoreactive T cell proliferation. Cell. lmmunol. 99, 461-475. Hoskin, D.W. and Murgita, R.A. (1985) Increased maternal T cell autoreactivity associated with syngeneic murine pregnancy. J. Reprod. lmmunol. 8, 187-196. Hoskin, D.W., Gr6nvik, K.-O., Hooper, D.C., Reilly, B.D. and Murgita, R.A. (1989) Altered immune response patterns in murine syngeneic pregnancy: Presence of natural null suppressor cells in maternal spleen identifiable by monoclonal antibodies. Cell. Immunol. 120, 42-60. Hoskin, D.W. and Murgita, R.A. (1989) Specific maternal anti-fetal lymphoproliferative responses and their regulation by natural immunosuppressive factors. Clin. Exp. Immunol. 76, 262-267. Jadus, M.R. and Parkman, R. (1986) The selective growth of murine newborn-derived suppressor cells and their probable mode of action. J. Immunol. 136, 783-792. Kawano, Y., Noma, T. and Yata, J. (1990) Identification of a cord blood T cell subset of CD3+4-8-45 R ÷ suppressing interleukin 2 production in the autologous mixed lymphocyte reaction and the mode of action of exogenous IL-2 in the induction of IL2 production. Cell. Immunol. 131, 27-40. Lala, P.H., Scodras, J.M., Graham, C.H., Lysiak, J.J. and Parhar, R.S. (1990) Activation of maternal killer cells in the pregnant uterus with chronic indomethacin therapy, IL-2 therapy, or a combination therapy is associated with embryonic demise. Cell. lmmunol. 127, 368-381. Leo, O., Foo, M., Sachs, D.H., Samelson, L.E. and Bluestone, J. (1987) Identification of a monoclonal antibody specific for a murine T3 polypeptide. Proc. Natl. Acad. Sci. U.S.A. 84, 1374-1378. Maes, L.Y., York, J.L. and Soderberg, L.S.F. (1988) A soluble factor produced by bone marrow natural suppressor cells blocks interleukin 2 production and activity. Cell. Immunol. 116, 35-43. Maier, T., Holda, J.H., Choi, K.L. and Claman, H.N. (1989) Characterization and functions of the natural suppressor cell systems. In: Functions of the Natural Immune System, (Reynolds, C.W. and Wiltrout, R.H., eds.), pp. 267-298. Plenum Publishing Corp., London and New York. Maroni, E.S. and de Sousa, M.A.B. (1973) The lymphoid organs during pregnancy in the mouse. A comparison between a syngeneic and an allogeneic mating. Clin. Exp. lmmunol. 13, 107-124. May, R.D., Slavin, S. and Vitetta, E.S. (1983) A partial characterization of suppressor cells in the spleens of mice conditioned with fractionated total lymphoid irradiation {TLI). J. lmmunol. 131, 1108-1114. Mclntosh, H.R., Segre, M. and Segre, D. (1982) Characterization of cyclophosphamide-induced suppressor cells. Immunopharmacology 4, 279-289. Mul6, J.J., Yang, J., Shu, S. and Rosenberg, S.A. (1986} The anti-tumour efficacy of lymphokineactivated killer and recombinant interleukin 2 in vivo: direct correlation between reduction of experimental metastases and cytolytic activity of lymphokine-activated killer cells. J. Immunol. 136, 3899-3909. Murgita, R.A. and Wigzell, H. (1981) Regulation of immune functions in the fetus and newborn. Prog. Allergy 29, 54-133.

125 Oseroff, A., Okada, S. and Strober~ S. (1984) Natural suppressor (NS) cells found in the spleen of neonatal mice and adult mice given total lymphoid irradiation express the null surface phenotype. J. Immunol. 132, 101-110. Pollack, S.B. and Emmons, S.L. (1982) NK-2. l: An NK-associated antigen detected with NZB anti-Balb/c serum. J. lmmunol. 129, 2277-2281. Reisner, Y., Ravid, A. and Sharon, N. (1976) Use of soybean agglutinin for the separation of mouse B and T lymphocytes. Biochem. Biophys. Res. Commun. 72, 1585-1591. Sharon, N. (1983) Lectin receptors as lymphocyte surface markers. Adv. lmmunol, 34, 213-298. Shustik, C., Cohen, I.R., Schwartz, R.S. and Lathan-Griffin, E. (1976) T-lymphocytes with promiscuous cytotoxicity. Nature (Lond.) 263, 699-701. Skyes, M., Sharabi, Y. and Sachs, D.H. (1990) Natural suppressor cells in spleens of irradiated bone marrow-reconstituted mice and normal bone marrow: Lack of SCA-I expression and enrichment by depletion of MAC-l-positive cells. Cell. Immunol. 127, 260-274. Slapsys, R.M. and Clark, D.A. (1983) Active suppression of host-versus-graft reaction in pregnant mice. V. Kinetics, specificity and in vivo activity of non-T suppressor cells localized to the genital tract of mice during first pregnancy. Am. J. Reprod. Immunol. 3, 65-71. Strober, S., Dejbachsh-Jones, S., van Vlasselaer, P., Duwe, G., Salimi, S. and Allison, J.P. (1989) Cloned natural suppressor cell lines express the CD3+CD4 - CD8- surface phenotype and the o~,/3heterodimer of the T cell antigen receptor. J. Immunol. 143, 1118-1122. Subiza, J.L., Vinuela, J.E., Rodriguez, R., Gil, J., Figueredo, M.A. and de la Concha, E.G. (1989) Development of splenic natural suppressor (NS) cells in Ehrlich tumour-bearing mice. Int. J. Cancer 44, 307-314. Sugiura, H., lnaba, M,, Ogata, H., Yasumizu, R., lnaba, H., Good, R.A. and Ikehara, S. (1988) Wheat germ agglutinin-positive cells in a stem cell-enriched fraction of mouse bone marrow have potent natural suppressor activity. Proc. Natl. Acad. Sci. U.S.A. 85, 4824-4826. Sydora, B.C. and Kronenberg, M. (1991) Characterization of a CD4-positive T-cell line derived from an athymic (nu/nu) mouse. Cell. Immunol. 134, 54-64. Van Vlasselaer, P. and Strober, S. (1991) 043 TCR+CD3+CD4 CD8- cloned natural suppressor (NS) cells produce an immunosuppressive factor which is different from IFN-r and TGF/3. Transplant. Proc. 23, 200-202. Wilde, D.B., Marrack, P., Kappler, J., Dialynas, D.P. and Fitch, F.W. (1983) Evidence implicating L3T4 in class I1 MHC antigen reactivity; monoclonal antibody GKI.5 (anti-L3T4a) blocks class 11 MHC antigen-specific proliferation, release of lymphokines, and binding by cloned murine helper T lymphocyte lines. J. Immunol. 131, 2178-2183. Wysocki, J.L. and Sato, V.L. (1978) "Panning" for lymphocytes: A method for cell selection. Proc. Natl. Acad. Sci. U.S.A. 75, 2844-2848.