Cutting Edge: Virus-Activated CD8 T Cells and Lymphokine-Activated NK Cells Express the Mast Cell Function-Associated Antigen, An Inhibitory C-Type Lectin¹

The mast cell secretory response is triggered by clustering of Fcε receptors (FcεRI) by IgE immune complexes. A cell surface protein has been identified on the rat mucosal mast cell line RBL-2H3 that inhibits FcεRI-mediated degranulation and cytokine release after aggregation by a specific mAb (1). This molecule has been named mast cell function-associated antigen (MAFA),³ and molecular analysis revealed that MAFA is a type II membrane protein that belongs to the C-type lectin superfamily (2). MAFA exists as both a monomer and a disulfide-linked homodimer, and the intracellular domain contains a putative tyrosine-based inhibition motif that was found to be constitutively phosphorylated. These characteristics reveal a close relationship of MAFA with other members of the C-type lectin superfamily that function as inhibitory receptors (i.e., Ly49 and CD94/NKG2) in NK cells and also in T cells. However, MAFA expression in the rat analyzed with the sensitive RT-PCR technique has been reported to be mast cell specific (3).

We have used the “differential display” technique to identify genes in activated T cells, and in this context we have now cloned the mouse homologue of the rat MAFA gene. Surprisingly, we found that the MAFA gene in the mouse was strongly expressed both by effector CD8 T cells after priming in vivo by a viral infection and by lymphokine-activated NK cells. The present study represents the first characterization of this gene in the mouse; its product exhibits the molecular characteristics of a regulatory molecule for CD8-effector T cells.

C57BL/6 (B6) mice were purchased from Harlan Winkelmann (Borchen, Germany), and B6.PL-Thy1^a (B6.Thy-1.1) mice were obtained from Dr. H. Mossmann (Max-Planck Institute for Immunobiology, Freiburg, Germany). CD8 lymphocytic choriomeningitis virus (LCMV) TCR transgenic mice (4) and CD4 LCMV TCR transgenic mice (5) were bred locally.

LCMV-WE was propagated on L929 cells, vesicular stomatitis virus (VSV) Indiana (Mudd-Summers isolate) was grown on baby hamster kidney cells and vaccinia virus (VV) strain WR was produced by infecting BSC 40 cells. Mice were infected i.v. with 200 plaque-forming units (PFU) of LCMV-WE, 10⁶ PFU LCMV-docile, 2 × 10⁶ PFU VSV, and 2 × 10⁶ PFU VV WR.

T cells from CD8 LCMV TCR transgenic mice were stimulated weekly with LCMV-infected, irradiated (2500 rad) B6 peritoneal macrophages in Iscove’s modified Dulbecco’s medium (IMDM, Life Technologies, Paisley, U.K.) supplemented with 10% FCS and 20 U/ml IL-2 (PharMingen, San Diego, CA). CD8 memory T cells isolated from LCMV-immune B6 mice (4–6 wk after infection) were cultured and stimulated in the same manner. Bone marrow-derived mast cells were isolated from B6 mice as described (6) and stimulated for 8 h with 0.5 μM ionomycine (Sigma-Aldrich, Deisenhofen, Germany). To generate lymphokine-activated killer cells, B6 spleen cells (SC) were depleted after lysis of RBC from T and B cells by negative selection with Dynabeads (Dynal, Hamburg, Germany). The remaining cells were cultured for 6 days in IMDM supplemented with 10% FCS and 1000 U/ml IL-2. After 6 days, the cells (90–95% NK1.1⁺) were harvested.

CD4 and CD8 effector and memory T cells were generated in vivo using an adoptive transfer system as described (7). Uninfected LCMVTCR transgenic mice were used as a source of naive CD4 or CD8 T cells. CD8 T cells were purified from SC by a two step negative selection procedure with Dynabeads as described (8). CD4-effector T cells were purified by depletion of CD8 T cells and B cells by negative selection with Dynabeads.

The differential display analysis was performed as described (8) using the DELTA RNA Fingerprinting Kit (Clontech Laboratories, Palo Alto, CA). Full-length cDNA was obtained using long-distance RT-PCR and the SMART-PCR cDNA Synthesis Kit (Clontech) according to the manufacturer’s instructions.

Total RNA was isolated from SC or purified cell populations with an RNA Isolation Kit (Fluka Chemie, Buchs, Switzerland). Northern blot analysis was performed as described (8) using 5–10 μg of total RNA per lane. The 600-bp fragment of the mouse MAFA (mMAFA) cDNA representing the entire coding region was used as mMAFA probe. The mouse cytoskeletal β-actin probe (560 bp) was amplified from cDNA by PCR using 5′ (ATGGATGACGATATCGCT) and 3′ (ATGAGGTAGTCTGTCAGGT) primers.

We performed mRNA differential display PCR to identify genes induced in effector CD8 T cells and isolated a PCR product that revealed significant sequence homology to the rat MAFA gene (2). Full-length cDNA was obtained by RT-PCR with mRNA from CD8 effector T cells and the full primary sequence was determined. The primary sequence was aligned with that of the rat MAFA, and sequence comparison revealed 80% identity and 89% similarity of the mouse to the rat sequence (Fig. 1). This result suggested that the mouse homologue of the rat MAFA gene had been isolated. Southern blot analysis of genomic DNA using a full-length mMAFA cDNA as a probe revealed a hybridization pattern suggesting that MAFA is a single gene in the mouse (not shown). The same conclusion has been drawn for the MAFA gene in the rat (3).

Northern blot analysis using the mMAFA cDNA as a probe demonstrated that CD8 T cells from LCMV TCR transgenic mice isolated on both day 8 and day 14 after LCMV infection expressed high levels of mMAFA mRNA (Fig. 2,A, lanes 2 and 3). In contrast, naive CD8 T cells from these mice did not express mMAFA at detectable levels (Fig. 2,A, lane 1). In CD8 memory T cells, mMAFA gene expression was still detectable 3–4 wk after infection, but to a lower extent (Fig. 2,A, lanes 4 and 5). Importantly, mMAFA gene expression was also induced in CD8 effector T cells from normal C57BL/6 mice isolated 8 days after LCMV infection (Fig. 2,B, lane 2). To test whether SC other than CD8 effector T cells expressed mMAFA, SC from B6 mice taken on day 8 after LCMV infection were depleted of CD8 T cells by negative selection in vitro. As shown in Fig. 2,B, lane 3, these CD8-depleted immune SC did not express mMAFA at significant levels, indicating that mMAFA was virtually exclusively expressed in activated CD8 T cells. To directly compare mMAFA expression in activated CD4 and CD8 T cells, effector T cells were generated using CD4 and CD8 TCR transgenic mice, which both express an Ag receptor specific for LCMV glycoprotein (GP). T cells from CD8 LCMV TCR transgenic mice are specific for LCMV GP amino acid 33–41 in the context of H-2D^b (4), and CD4 LCMV TCR transgenic mice express an Ag receptor specific for LCMV GP amino acid 61–80 together with I-A^b (5). CD4 and CD8 T cells from both TCR transgenic lines were activated in vivo using the same adoptive transfer system, which allows the same vigorous expansion of the transferred transgenic T cells (5, 7). As shown in Fig. 2 B, lane 5, effector cells from CD4 LCMV TCR transgenic mice expressed mMAFA at low levels when compared with effector cells from CD8 LCMV TCR transgenic mice (lane 4). Thus, these results indicated that mMAFA expression on effector T cells was CD8 specific.

To examine whether viral infections other than LCMV were also capable of inducing mMAFA expression, SC from B6 mice acutely infected with LCMV, VV, or VSV were examined. Fig. 3,A shows that mMAFA expression levels were comparable in splenocytes from mice infected with either LCMV, VSV, or VV. It is noteworthy that mMAFA expression was lower in splenocytes from mice infected with a high dose (10⁶ PFU) of the docile strain of LCMV, a virus strain known to induce clonal exhaustion of CD8 effector T cells (9). To test whether Ag stimulation in vitro also induced mMAFA expression, CD8 T cells from LCMV TCR transgenic mice were stimulated with LCMV-infected APC. Ag stimulation of CD8 T cells from LCMV TCR transgenic mice in vitro is known to lead to vigorous proliferation and induction of CTL activity (10). However, these in vitro conditions were not sufficient to induce significant mMAFA expression (Fig. 3,B, lanes 2–4). Only after repeated restimulation and prolonged in vitro culture (3–4 wk), low level mMAFA expression was apparent in these cultures (Fig. 3,B, lanes 5 and 6). In striking contrast, high levels of mMAFA expression were induced in memory CD8 T cells within 3 days of in vitro restimulation (Fig. 3 B, lane 7); these memory CD8 T cells had been primed in vivo with LCMV.

Expression of mMAFA mRNA was not detected by Northern blotting in brain, heart, kidney, liver, lymph node, lung, skeletal muscle, spleen, thymus, and testis (not shown). As depicted in Fig. 4,A, mMAFA was also not expressed in EL-4, RMA-S, BW5147 lymphoma cells, or in CD4 and CD8 T-cell hybridomas (BWLy2–3, IT H6/A11). Similarly, J558 and P3-X63Ag8 myeloma cells and B cell hybridomas did not express mMAFA. However, the CTL clone HY-Ad9 (Fig. 4,A, lane 3), which exhibits NK-like activity (11), expressed mMAFA levels comparable to CD8-effector T cells. This result led us to examine mMAFA expression in lymphokine-activated NK1.1⁺ killer cells. As shown in Fig. 4 B, lane 3, lymphokine-activated NK1.1⁺ cells (95% purity) generated by culturing T and B cell-depleted SC in the presence of IL-2 expressed high levels of mMAFA mRNA. The expression pattern of mMAFA is reminiscent of KIR and CD94/NKG2 molecules, which function as inhibitory receptors for both human NK cells (12, 13) and T cells (14, 15, 16). Similar to our data with mMAFA, KIRs are expressed on NK cells and on effector/memory phenotype CD8 T cells (17), and attempts to induce KIR expression on T cells by in vitro stimulation were unsuccessful (17, 18).

The MAFA gene was originally identified in the rat mucosal mast cell line RBL-2H3, and, in the rat, MAFA gene expression appears to be mast cell specific because signals were obtained from lung tissue but not from other organs including spleen and lymph nodes when the sensitive RT-PCR technique was used (3). Surprisingly, we did not observe mMAFA gene expression in IL-3/IL-4 induced bone marrow-derived mast cell cultures, which are thought to represent an in vitro analogue of mucosal mast cells (6) (Fig. 4,C, lane 2). In addition, stimulation of these cells with ionomycine did not lead to mMAFA gene expression (Fig. 4 C, lane 3). Our results indicate that MAFA gene expression in the mouse is not mast cell specific and may not even be expressed on mucosal mast cells in this species.

In analogy to the inhibitory effect on mast cell degranulation by rat MAFA, we postulate that the mouse homologue has a similar function and serves as an inhibitory molecule on CD8-effector T cells and on NK cells. Due to the lack of the appropriate serological reagents, we could not yet test this hypothesis directly. However, it is important to stress that the inhibitory function of the rat MAFA molecule has been clearly established (1). In contrast to rat MAFA, which is constitutively expressed in the rat mast cell line RBL-2H3, mMAFA expression in CD8 T cells was induced by Ag stimulation. Compared with other T cell-activation markers, mMAFA appears rather late in the course of a CTL response. This late kinetic of mMAFA expression fits in well with the postulated inhibitory role of this molecule for CD8 effector T cells.

We thank Dr. A. Oxenius for the generous gift of CD4 LCMV TCR transgenic mice and Dr. E. Schmitt for the mast cell cultures. We also thank S. Batsford for comments on the manuscript and S. Denkler and T. Imhof for animal husbandry.