B220⁺ Double-Negative T Cells Suppress Polyclonal T Cell Activation by a Fas-Independent Mechanism That Involves Inhibition of IL-2 Production

Fas and its specific ligand (FasL) ² are cell surface proteins whose interaction triggers a cascade of subcellular events that result in apoptosis (1, 2). This mechanism is critical for elimination of autoreactive T cells that escape negative selection in the thymus. However, loss of function mutations in the gene encoding Fas (Fas^lpr) or FasL (FasL^gld) does not lead to T cell-mediated disease (3), despite persistence of autoreactive T cells in abnormally high numbers (4). To the contrary, lack of a functional Fas pathway is associated with an Ab-mediated lupus-like disease and impaired delayed-type hypersensitivity (5, 6). More important, Fas^lpr and FasL^gld mutations prevent, rather than exacerbate T cell-mediated diseases, including autoimmune diabetes (7, 8, 9). Thus, the consequences of loss of function mutations in the Fas signaling pathway are surprisingly inconsistent with the role of the Fas pathway in regulating peripheral tolerance. The underlying mechanism of this paradox is currently not well understood.

A second hallmark of mice (3) and humans (10, 11) with defective Fas signaling pathways is the accumulation of αβ T cells that lack both CD4 and CD8 coreceptors and express B220 molecules (hereafter referred to as B220⁺ double-negative (DN) T cells). The prevailing view is that B220⁺ DN T cells are aberrant cells of no or limited physiological relevance. However, recent studies show that lpr B220⁺ DN T cells, which constitutively express FasL, can kill wild-type (wt) Fas-expressing CD8 T cells and prolong allograft survival in a Fas-dependent manner (12). Although these studies show the ability of B220⁺ DN T cells to regulate Fas-expressing T cells, the role of B220⁺ DN T cells in regulating autologous T cells with impaired Fas pathway is currently not fully understood. Furthermore, the mechanism by which B220⁺ DN T cells suppress other T cells is unclear.

In this study, we show that B220⁺ DN T cells, whether generated in vitro or isolated from mutant mice, can suppress activation of polyclonally activated autologous and syngeneic T cells in Fas-independent fashion. In addition, Fas-deficient CD4 T cells and Fas-expressing CD4 T cells were suppressed to the same extent by FasL-expressing lpr B220⁺ DN T cells showing that suppression of CD4 T cells is not modulated by a Fas-dependent mechanism. Suppression mechanism involves inhibition of IL-2 production and its high affinity IL-2R α-chain (CD25). Suppression also required cell/cell contact and TCR activation of B220⁺ DN T cells, but was cytokine independent. These findings suggest a potential role for B220⁺ DN T cells in immunoregulating autoreactive T cells in mice and humans with defects in the Fas signaling.

AND TCR transgenic mice (13) specific for moth cytochrome c (MCC) in the context of I-E^k on B10.BR (H-2^k) background (14) were a generous gift from S. Swain (Trudeau Institute, Saranac, NY). C3H-lpr/lpr and C3H-gld/gld mice with homozygous loss of function mutations in Fas (Fas^lpr), FasL (FasL^gld) molecules and wt (Fas^wt) C3H/HeJ (H-2^k) mice were purchased from Jackson ImmunoResearch Laboratories (West Grove, PA). All mice were bred and maintained at the Cancer Center of the Johns Hopkins School of Medicine.

Anti-CD3 beads were prepared, as previously described (15). Briefly, 75 mg of anti-CD3 (2C11) mAb in borate buffer was mixed with tosyl-activated 450 Dynabeads (Dynal, Lake Success, NY) and tumbled overnight at 40°C. The beads were then soaked in 5% serum at 40°C to block remaining active sites and stored in PBS until used. Anti-CD3/CD28 beads were prepared by mixing anti-CD3 (2C11) and anti-CD28 (37.51) mAbs at a 1:1 ratio and coupled to tosyl-activated 450 Dynabeads, as described above.

We have used negative selection to isolate various subsets of T cells. MCC-specific AND TCR CD4 T cells bearing Vα11Vβ3 segments were isolated from spleens of AND/B10.BR mice by negative selection with Dynabeads. Briefly, splenocytes were incubated with a cocktail of biotin-conjugated rat mAbs specific for murine CD8 (53-5.8 mAb), I-E^k (14-4-4S mAb), CD16/CD32 (2.4G2 mAb), and B220 (RA3-6B2 mAb) surface molecules (all were purchased from BD PharMingen, San Diego, CA). Streptavidin-conjugated Dynabeads were used to capture and remove mAb-coated cells, according to the manufacturer’s instructions. Purity and specificity of isolated T cells were determined by three-color flow cytometric analysis using mAbs specific for murine Vα11, CD8, and CD4 molecules. Purity of AND TCR CD4 T cells that were used for in vitro generation of B220⁺ DN T cells was always greater than 95%.

We used the same protocol to isolate CD4 T cells from the lymph nodes of C3H-lpr/lpr and C3H-gld/gld mice. B220⁺ DN T cells were isolated by using a cocktail of biotinylated mAbs specific for the following molecules: CD4 (GK1.5), CD8, I-E^k, and CD16/CD32. Purity of CD4 and B220⁺ DN T cells from lpr and gld mice that were used in these studies was always greater than 95%, as determined by flow cytometry.

Fluorochrome-conjugated mAbs specific for TCR (clone H57-597), CD4 (clone H129.19), CD8 (clone 53-6.7), CD45R/B220 (clone RA3-6B2), CD25 (clone 3C7), CD44 (clone 30-F11), CD45RB (clone 16-A), and CD62L (clone MEL-14) were purchased from BD PharMingen. Cells stained with fluorescent Abs were analyzed with a FACSCaliber flow cytometer and CellQuest software (BD Biosciences, San Jose, CA).

Typically, AND T cells (1 × 10⁵) were incubated with anti-CD3/CD28 beads (2 × 10⁶) in wells of a 24-well microtiter plate. Recombinant IL-2 (20 U/ml) and IL-7 (2–10 ng/ml) were added on day 2 and thereafter every 4–5 days to enhance DN T cell survival (16). Media were changed and cells were split, as required. Fresh anti-CD3/CD28 beads were added at frequent intervals of 4–5 days to keep T cells continuously stimulated. B220⁺ DN T cells that were ≥85% pure were used in the suppression assay.

This assay was performed according to the protocol described by Thornton and Shevach (17). Briefly, B220⁺ DN T cells were mixed with CD4 T cells in the presence of anti-CD3 or anti-CD3/CD28 beads in 96-well round-bottom plates in triplicate cultures. After 48 h, 100 ml of supernatant fractions was collected for cytokine analysis and wells were replenished with fresh medium. Proliferation was determined after 72 h with [³H]thymidine that was added during the last 18 h of incubation. Supernatant fractions were assayed for indicated cytokines using capture ELISA (Endogen, Woburn, MA). Each experiment represents at least three independent experiments with similar results.

For neutralization assays, purified Abs specific for IL-2, IFN-γ, and IL-10 cytokines were purchased from BD PharMingen. Pan-specific TGF-β-neutralizing Ab was obtained from R&D Systems (Minneapolis, MN).

Transwell experiments were conducted in 24 wells in 0.8 ml of complete tissue culture medium. To measure the effect of soluble cytokines, CD4 T cells (5 × 10⁵) were stimulated in the lower chamber and DN T cells (2.5 × 10⁶) in the upper chamber of Transwell. The semipermeable membrane that separates the upper and lower chamber of Transwell (Corning, Costar, Cambridge, MA) allows diffusion of soluble materials, but not particulate cells. As controls, CD4 T cells were cultured in the lower chamber either alone or mixed with DN T cells. Anti-CD3 beads (5 × 10⁵) were used for stimulation. To study the role of soluble cytokines in the suppression mechanism, supernatants collected from cultures of activated CD4 T cells, B220⁺ DN T cells, or cocultures of CD4 and B220⁺ DN T cells were used to inhibit activation of CD4 T cells.

Real-time PCR was performed to measure IL-2 mRNA transcripts in CD4 T cells (5 × 10⁵) that were activated in the presence or absence of B220⁺ DN T cells (2.5 × 10⁶). IL-2 transcripts were also measured in B220⁺ DN T cells stimulated alone. After 48-h stimulation with anti-CD3 beads, total RNA was isolated from different cultures with the RNeasy Kit (Qiagen, Valencia, CA). Briefly, cells from each sample were disrupted, homogenized, and resuspended in an equal volume of 70% ethanol. Samples were loaded into minicolumns and washed three times. Total RNAs were eluted with water, and concentrations were determined. First strand cDNAs were synthesized from total RNAs using Superscript II (Invitrogen, Carlsbad, CA). Briefly, 1 mg of total RNAs was reverse transcribed with 50 U of Superscript II reverse transcriptase and 150 ng of random hexamers as primers.

Quantitative real-time PCR was performed by adding cDNA to the reaction mixture containing IL-2 primers (Applied Biosystems, Foster City, CA). After 40 cycles, the relative copy numbers of IL-2 mRNA were determined by normalizing the amount of total RNA added to each reaction with 18S rRNA.

We have recently described a system to enrich B220⁺ DN T cell population in vitro by chronic anti-CD3/CD28 stimulation of AND TCR transgenic CD4 T cells (15). Using this system, we were able to generate B220⁺ DN T cells that were at least 85% pure (Fig. 1,A). We explored the suppressive properties of B220⁺ DN T cells by determining whether they could inhibit proliferation of syngeneic CD4 T cells using an in vitro model system (18, 19, 20, 21, 22). Freshly isolated AND TCR CD4 T cells were cultured either alone or with increasing numbers of B220⁺ DN T cells. The different cultures were stimulated with anti-CD3 beads, and proliferation was measured after 72 h by using [³H]thymidine. CD4 T cells proliferated significantly in the absence of B220⁺ DN T cells (Fig. 1,B, left panel), but their proliferation was inhibited in the presence of B220⁺ DN T cells in a dose-dependent manner (Fig. 1 B, right panel).

Because of the modest proliferative response of naive CD4 T cells stimulated with anti-CD3 in the absence of costimulation (Fig. 1,B), it was important to determine whether B220⁺ DN T cells could inhibit more robust responses induced by TCR and CD28 stimulation. Stimulation of AND TCR transgenic CD4 T cells with anti-CD3/CD28 beads, as expected, enhanced proliferation by ∼5- to 7-fold and led to IL-2 production, but did not lead to B220⁺ DN T cell proliferation (Fig. 2, A and B). Importantly, B220⁺ DN T cells completely suppressed proliferation and IL-2 production of CD28-costimulated CD4 T cells (Fig. 2, C and D). Thus, whereas CD28 costimulation augmented the overall responses of CD4 T cells, it did not abrogate the suppressive ability of the in vitro generated B220⁺ DN T cells.

To determine the pathophysiologic relevance of the suppressive ability of in vitro generated B220⁺ DN T cells, we determined whether B220⁺ DN T cells that accumulate in Fas-deficient lpr mice also have suppressive properties. We isolated B220⁺ DN T cells and CD4 T cells from 16- to 20-wk-old lpr mice by negative selection, as described in Materials and Methods. Purity of B220⁺ DN and CD4 T cells was greater than 95% (Fig. 3, A and B). All DN T cells expressed B220 and manifested a phenotype of activated T cells, as indicated by high levels of CD44 and low levels of CD62L. CD45RB molecules, which are usually down-regulated on activated T cells, remained highly expressed on B220⁺ DN T cells. The majority of CD4 T cells also exhibited activated phenotype (CD44^high; CD62L^low; CD45RB^low) consistent with the high frequencies of autoreactive T cells in mice with impaired Fas-mediated deletion (Fig. 3,B). Functionally, CD4 T cells proliferated in response to anti-CD3 stimulation, whereas B220⁺ DN T cells did not proliferate (Fig. 3 C), as described (3).

We next determined whether lpr B220⁺ DN T cells could suppress proliferation of autologous CD4 T cells. Our initial experiments showed that lpr B220⁺ DN T cells could suppress CD4 T cells, but they were less potent than the in vitro generated B220⁺ DN T cells (data not shown). Robust suppression, however, was seen at higher DN to CD4 T cell ratio of 5:1 or 2.5:1. Suppression was not significantly affected by the high numbers of cells in the cocultures, as only minimal effect was seen in corresponding cultures that contained equal number of CD4 T cells (Fig. 3,D). B220⁺ DN T cells also suppressed proliferation of CD8 T cells in a dose-dependent manner, and no suppression was seen in control cultures that contain equivalent numbers of CD8 T cells (Fig. 3 E).

Although stimulation of CD4 T cells with anti-CD3/28 beads resulted in vigorous proliferation and cytokine production, it did not alter the suppressive ability of B220⁺ DN T cells. Proliferation, IL-2, IL-10, and IFN-γ production of anti-CD3/28-stimulated CD4 T cells were inhibited by B220⁺ DN T cells in a dose-dependent manner (Fig. 4). These results show that lpr B220⁺ DN T cells possess a significant ability to suppress activation of autologous CD4 and CD8 T cells.

Significant accumulation of B220⁺ DN T cells also occurs in gld mice with homozygous loss of function mutations in the FasL gene. We determined whether B220⁺ DN T cells from gld mice share the suppressive properties of B220⁺ DN T cells that accumulate in lpr mice. We purified B220⁺ DN T cells from 16- to 20-wk-old C3H-gld/gld mice and tested their ability to suppress activation of autologous (FasL^gld) or syngeneic (Fas^lpr) CD4 T cells. All cell preparations were >95% pure (data not shown). B220⁺ DN T cells from gld mice potently suppressed proliferation of CD4 T cells from lpr and gld mice in a dose-dependent fashion (Fig. 5 B). They also inhibited IL-10 and IFN-γ production (data not shown). Therefore, it appears that suppression is a general function of murine B220⁺ DN T cells associated with impaired Fas-mediated apoptosis.

The ability of B220⁺ DN T cells from lpr and gld mice to suppress activation of autologous CD4 T cells shows that suppression is Fas independent. However, B220⁺ DN T cells from lpr mice constitutively express FasL (Fig. 5,A, left dot plot) and are reported to induce apoptosis of wt (Fas^wt) CD8 T cells in a Fas-dependent fashion (12). We therefore determined whether expression of Fas on CD4 T cells augments their suppression. lpr B220⁺ DN T cells inhibited proliferation of Fas-expressing wt and Fas-deficient lpr CD4 T cells to similar extent (Fig. 5, C and D), showing that offered Fas/FasL interaction does not modulate B220⁺ DN T cell-mediated suppression of CD4 T cells. We further confirmed these results by CFSE analysis. Proliferation of CFSE-labeled CD4 T cells from lpr and wt mice was inhibited by lpr B220⁺ DN T cells to similar extent, as determined by the dilution of CFSE (data not shown). In addition, the level of apoptosis, as measured by Annexin V, among lpr and wt CD4 T cells cocultured with lpr B220⁺ DN T cells was similar (data not shown). Thus, B220⁺ DN T cell-mediated suppression appears to be Fas independent and is not modulated by Fas/FasL interactions.

T cell-mediated suppression can be mediated via direct contact, soluble cytokines, or both (18, 23). To determine whether suppression is contact dependent or mediated by cytokines, we performed a Transwell experiment. Separation of simultaneously activated B220⁺ DN and CD4 T cells by a semipermeable membrane of Transwell abrogated suppression, whereas significant suppression occurred in mixed cultures (Fig. 6,A). Furthermore, supernatant fractions from mixed or separate cultures of activated CD4 T cells or B220⁺ DN T cells did not suppress proliferation of CD4 T cells, confirming that suppression is cytokine independent (data not shown). Neutralization of potentially suppressive cytokine IL-10 and TGF-β and other cytokines also did not abrogate suppression (data not shown). Suppression required TCR activation because lpr B220⁺ DN T cells were unable to suppress Ag-specific proliferation of AND TCR transgenic CD4 T cells stimulated with MCC/I-E^k beads (Fig. 6,B, left panel). In contrast, in vitro generated AND Tg B220⁺ DN T cells readily suppressed proliferation of syngeneic AND Tg CD4 T cells (Fig. 6, right panel). Thus, it appears that suppression of CD4 T cells requires direct contact and TCR activation, but is cytokine independent.

Because very small amounts of IL-2 were detected in the coculture of CD4 T cells and B220⁺ DN T cells (Fig. 4,B), lack of IL-2 could be the limiting factor that impaired CD4 T cell proliferation. Addition of exogenous IL-2 to the cocultures of lpr B220⁺ DN and CD4 T cells significantly abrogated suppression at low DN to CD4 T cell ratio (2.5:1) and partly reversed suppression at a higher ratio of 5:1. These results show IL-2 can overcome suppression and that lack of IL-2 plays a role in the suppression mechanism (Fig. 6 C). Abrogation of suppression was not due to proliferation of B220⁺ DN T cells, which did not proliferate in response to anti-CD3 stimulation even in the presence of IL-2 (data not shown).

To further understand the mechanism of suppression, it was important to determine whether IL-2 was produced by CD4 T cells, but consumed by lpr B220⁺ DN T cells (IL-2 sink) or, alternatively, lpr B220⁺ DN T cells impaired the ability of CD4 T cells to produce IL-2. Differentiation between these mechanisms was achieved by the analysis of IL-2 mRNA levels in different cultures using quantitative real-time PCR. There was almost complete inhibition of IL-2 transcription by CD4 T cells stimulated in the presence of lpr B220⁺ DN T cells, whereas high numbers of IL-2 mRNA transcripts were detected in CD4 T cells stimulated in the absence of B220⁺ DN T cells (Fig. 7,A). No IL-2 mRNAs were detected in B220⁺ DN T cell stimulated alone consistent with their failure to produce IL-2. Because the ratio of DN to CD4 cells was 5:1 and DN cells do not express IL-2 mRNA, the decreased transcript levels in the cocultured cells might have resulted from a cell dilutional effect. However, the inhibition of IL-2 mRNA in DN/CD4 cocultures was so potent and the differences in the levels of IL-2 message between CD4 cultures and CD4 and DN cocultures were in the range of 15- to 20-fold and thus could not be accounted for by the dilutional effect alone. In addition, inhibition of IL-2 mRNA transcription provides an explanation for the suppression of IL-2 production in the DN/CD4 T cell cocultures (Figs. 2,D and 4B). The alternative possibility that IL-2 was produced by CD4 T cells, but consumed by B220⁺ DN T cells is unlikely, as B220⁺ DN T cells (Fig. 7 B) did not up-regulate CD25 (IL-2R α-chain), and thus their ability to act as IL-2 sink is significantly impaired. Thus, the failure of suppressed CD4 T cells to produce IL-2 implicates inhibition of IL-2 transcription as mechanism of suppression and rules out consumption of IL-2 by lpr B220⁺ DN T cells as a cause of suppression.

We also determined whether B220⁺ DN T cells could suppress CD25 up-regulation. Expression of CD25 is necessary for efficient IL-2 uptake and proliferation of activated T cells. B220⁺ DN T cells potently inhibited CD25 up-regulation. Very few CD4 T cells expressed CD25 when stimulated in the presence of B220⁺ DN T cells (Fig. 7,B, right top panel), whereas the CD44 expression was not inhibited (Fig. 7,B, right bottom panel). In positive control cultures of CD4 T cells that were stimulated in the absence of lpr B220⁺ DN T cells, high levels of both CD25 (Fig. 7,B, left top panel) and CD44 molecules (Fig. 7,B, left bottom panel) were detected. B220⁺ DN T cells did not express CD25 molecules when stimulated alone or in the presence of CD4 T cells (Fig. 7 B, middle and right top panels). These results show that B220⁺ DN T cell-mediated suppression involves inhibition of transcription of IL-2 and its high affinity receptor.

These studies show that B220⁺ DN T cells, whether in vitro generated or isolated from lpr or gld mice, can potently suppress proliferation of activated CD4 T cell and prevent cytokine production. Suppression is Fas independent and is not influenced by Fas/FasL interaction. Suppression mechanism involves inhibition of IL-2 production and requires cell/cell contact and TCR activation, but not soluble cytokines. These findings suggest a potential immunoregulatory role for B220⁺ DN T cells in animals and humans with impaired Fas-mediated tolerance.

Fas-induced apoptosis is critical for elimination of effector T cells and maintenance of immune homeostasis. To allow productive immune response, activated cells remain relatively resistant to Fas-mediated apoptosis during the early phase of activation, proliferation, and differentiation into effector cells (2). Production of IL-2 as well as undergoing cell cycle are among the factors that sensitize activated T cells for Fas-mediated apoptosis (2). IL-2 facilitates Fas-mediated signal by inhibition of the transcription and expression of Fas-associated death domain-like IL-1b-converting enzyme-inhibitory protein (24). Failure of suppressed cells to produce IL-2 could thus reduce the impact of Fas engagement on the intensity of suppression. Consistent with this view, we did not detect differences in the level of suppression of Fas-deficient and Fas-expressing CD4 T cells by FasL-lpr B220⁺ DN T cells (Fig. 5, C and D). In contrast, lpr B220⁺ DN T cells were reported to kill target cells in the presence of IL-2 and when preactivated CD8 T cells were used (12). Thus, it appears that the activation status and presence or absence of IL-2 are important factors that influence whether target cells are suppressed or killed by lpr B220⁺ DN T cells. The potential ability of lpr B220⁺ DN T cells to induce apoptosis of Fas-sensitive cells puts some restrains on the use of lpr B220⁺ DN T cells for adoptive immunotherapy, as it may lead to nonspecific cytotoxicity. Consistent with this view (3), syngeneic transfer of lpr/lpr (25), but not gld/gld (26) bone marrow into irradiated wt mice has been reported to result in a wasting disease. Efforts should therefore focus on using the equally potent, but potentially safe, FasL-deficient, gld B220⁺ DN T cells for adoptive immunotherapy studies.

In summary, our data show that B220⁺ DN T cells derived from various sources are capable of suppressing proliferation of autologous and syngeneic T cells by a Fas-independent mechanism that involves inhibition of IL-2 and its high affinity receptor. Suppression also requires cell/cell contact and TCR activation of B220⁺ DN T cells, but not soluble cytokines. These results offer novel insights into the relationship between the different components of T cell tolerance network. In addition, these studies may help better understanding of the pathogenesis of autoimmune lymphoproliferative syndrome in humans and have broad implications for other autoimmune disorders.

We thank Drs. Noel Rose and Patrizio Caturegli for critical review of the manuscript.