The Primary Site of CD4⁻8⁻B220⁺ αβ T Cells in lpr Mice: The Appendix in Normal Mice¹

Conventional T cells, generated by the mainstream of T cell differentiation in the thymus, belong to either the CD4⁺ or the CD8⁺ subset and, therefore, double-negative (DN)³ CD4⁻8⁻ cells are rarely seen in this population. However, a unique population of DN CD4⁻8⁻B220⁺ αβ T cells appears in MRL-lpr/lpr (lpr) mice, and these mice fall victim to autoimmune disease as a function of age (1, 2, 3, 4, 5). The lpr gene is now estimated to be an abnormal Fas gene into which an early transposon of a retrovirus is transfected (6, 7, 8). Due to this lpr gene, a lymphoproliferative disorder is induced, especially in CD4⁻8⁻B220⁺ αβ T cells. These T cells were found recently to be generated extrathymically in the liver and through an alternative intrathymic pathway (9, 10, 11, 12). However, the proportion of such DN T cells was small even in these organs of normal mice.

To date, there have been no reports of an abundance of DN CD4⁻8⁻B220⁺ αβ T cells in any immune organs of normal mice. Although a few cells were found to exist at some sites of normal mice (13), a large number of or a large proportion of such DN CD4⁻8⁻B220⁺ αβ T cells has not been detected at any sites in normal mice. We herein report that such αβ T cells were abundant at intraepithelial sites of the appendix in normal mice, irrespective of strains. This might be because the appendix developed from the intestine in early phylogeny. The appendix, therefore, still carries primordial T cells in mice. The present results reveal that abnormal αβ T cells seen in lpr mice might be a normal T cell subset that exists even in normal mice.

C57BL/6 (B6), BALB/c, C3H/HeN, and lpr mice at the age of 5 to 15 wk were used. All mice were fed under specific pathogen-free (SPF) conditions in the animal facility of Niigata University (Niigata, Japan). B6 mice fed under conventional conditions and germfree conditions were also used at the age of 15 wk. B6.Ly-5.1 mice were also used. All of these mice were purchased from Clea Japan (Tokyo, Japan).

Mice anesthetized with ether were killed after complete exsanguination through incised axillary arteries and veins, and the liver, spleen, appendix, small intestine, and colon were removed. The liver was pressed through 200-gauge stainless steel mesh and then suspended in Eagle’s MEM medium supplemented with 5 mM HEPES (Nissui Pharmaceutical Co., Tokyo, Japan) and 2% heat-inactivated newborn calf serum. After washing, the cells were fractionated by centrifugation in 15 ml of 35% Percoll solution (Pharmacia Fine Chemicals, Piscataway, NJ) for 15 min at 2000 rpm (14). The pellet was resuspended in erythrocyte-lysing solution (155 mM NH₄Cl, 10 mM KHCO₃, 1 mM EDTA-Na, and 170 mM Tris, pH 7.3). The spleen was pressed through 200-gauge stainless steel mesh. The pellet was treated with the erythrocyte-lysing solution. The appendix, small intestine, and colon were removed and flushed with PBS to eliminate luminal contents (15). The mesentery, Peyer’s patches, or appendix lymphoid follicles were then resected. To obtain intraepithelial lymphocytes (IEL), the intestine was cut longitudinally and then into 1- to 2-cm pieces. These fragments were incubated for 15 min in 15 ml Ca²⁺- and Mg²⁺-free PBS containing 5 mM EDTA, in a 37°C shaking-water bath. The supernatant was then collected. The cell suspensions were collected and centrifuged in a discontinuous 40%/80% Percoll gradient solution at 2800 rpm for 25 min. Cells from the 40%/80% interface were collected.

Lamina propria lymphocytes (LPL) were prepared after the digestion of intestine with collagenase type II at a concentration of 90 U/ml in the medium. Samples were incubated for 45 to 90 min in a 37°C shaking-water bath. Digested intestine was then pressed through 200-gauge stainless steel mesh and suspended in medium. Cells were fractionated by the Percoll solution.

The surface phenotype of cells was analyzed using mAbs in conjunction with two-color or three-color immunofluorescence tests (14). The mAbs used in this study included FITC-, phycoerythrin (PE)-, or biotin-conjugated reagents of anti-CD3 (145-2C11), anti-B220 (RA3-6B2), anti-CD2 (RM2-5), anti-CD5 (53-7.3), anti-TCR-αβ (H57-597), anti-TCR-γδ (GL-3), anti-CD4 (RM4-5), anti-CD8 (53-6.7), anti-Gr-1 (RA3-8C5), anti-Mac-1 (M1/70), and anti-c-kit (SC1) mAbs (PharMingen Co., San Diego, CA). Biotin-conjugated mouse anti-Ly-5.1 mAb was kindly provided by Dr. T. Kina, Chest Disease Research Institute, Kyoto University (Kyoto, Japan). Biotin-conjugated reagents were developed with PE or Red 613-conjugated streptavidin (Becton Dickinson, Mountain View, CA). To prevent nonspecific binding of mAbs, CD32/16 (24G2) was added before staining with labeled mAbs. The fluorescence-positive cells were analyzed with FACScan using Lysis II software (Becton Dickinson).

Cytotoxicity was examined by a ⁵¹Cr release assay (16). Fresh mononuclear cells (MNC) in various organs and IEL in the appendix and small intestine were prepared from B6 mice. ⁵¹Cr-labeled YAC-1 cells (1 × 10⁴/well) or ⁵¹Cr-labeled syngeneic thymocytes (2 × 10⁴/well) were incubated with effector cells at the indicated E:T ratios at 37°C for 4 h. After the incubation, radioactivity released into supernatants was counted and percent specific lysis was calculated.

IEL isolated from the appendix and small intestine for morphologic study were fixed in 2.5% glutaraldehyde and 1% OsO₄, as described previously (14). Pellet samples were dehydrated through a graded series of ethanol and embedded in Epon 812. Ultrathin sections stained with uranyl acetate and lead citrate were observed at the same magnification with a Hitachi H-7000 transmission electron microscope. Anti-c-kit (ACK-2) mAb was a gift from Dr. Y. Kanamori (Keio University, Tokyo, Japan). The longitudinally opened appendix was embedded in O.C.T. compound at −80. The tissue segments were sectioned with a cryostat at 6 μm and applied to poly-l-lysin-coated slide glasses (14). The tissue sections that had been air dried and fixed in acetone for 10 min at room temperature were incubated with appropriately diluted rat anti-c-kit mAb for 30 min at 37°C or overnight at 4°C, and rinsed three times with PBS, followed by incubation with biotin-conjugated goat anti-rat IgG (Cedarlane Laboratories, Ontario, Canada). Subsequently, the sections were incubated with avidin-biotin peroxidase complex (Vectastain ABC kit; Vector Laboratories, Burlingame, CA).

Lineage marker-negative (Lin⁻: Ly-1⁻, B220⁻, Gr-1⁻, Mac-1⁻) cells were thought to be fluorescence-negative cells after staining with all markers described above. Lin⁻ c-kit⁺ cells were obtained from IEL of the appendix in BALB/c mice by FACStar II^Plus (Becton Dickinson). These cells (2 × 10⁵/mouse) were injected i.v. into 4-Gy irradiated C.B17-Icr/scid(scid) mice. Similarly, Lin⁻ c-kit⁺ cells were also obtained from IEL of the appendix in B6.Ly-5.1 mice, and were injected into 4-Gy irradiated B6.Ly-5.2 mice. Four weeks after injection, the surface phenotype of cells was analyzed by using mAbs in conjunction with two-color or three-color immunofluorescence tests.

MNC were isolated from the liver, spleen, appendix, and small intestine of B6 mice at the age of 8 wk (Fig. 1,A). Two-color staining for CD3 and B220, that for CD3 and a mixture of CD4 and CD8, and that for CD3 and CD2 were conducted (14). In the liver and spleen, the populations of CD3⁺ T cells and B220⁺ B cells were found to be isolated from each other. In contrast, a large proportion of B220⁺ cells was found among CD3⁺ cells in IEL of the appendix. It was estimated that CD3⁻B220⁺ cells were surface Ig⁺ (sIg), while CD3⁺B220⁺ cells were sIg⁻ in IEL of the appendix (data not shown). Although a few CD3⁺B220⁺ cells were present in IEL of the small intestine, the majority were CD3⁺B220⁻ T cells. In addition to the uniqueness of CD3⁺B220⁺ cells, IEL of the appendix contained considerable proportions of DN CD4⁻8⁻ cells and CD2⁻ cells among CD3⁺ cells (Fig. 1 A, center and lower columns). In the liver, there were DN CD4⁻8⁻ cells, but not CD2⁻ cells. In the case of the intestine, there was an abundance of CD2⁻ cells, but few DN CD4⁻8⁻ cells. In other words, the phenotype of IEL in the appendix (i.e., CD3⁺CD4⁻8⁻B220⁺CD2⁻) is different from those in the liver, spleen, and intestine. Rather, it resembles the phenotype of abnormally expanding T cells in lpr mice.

To further compare such points, two-color staining for CD5 and B220, that for CD4 and CD8, and that for TCR-αβ and TCR-γδ were conducted (Fig. 1 B). It was estimated that only a few CD3⁻B220⁺sIg⁺ B cells existed in the appendix (<5%). In the cases of the liver and spleen, both CD4⁺ and CD8⁺ cells, mainly αβ T cells, existed. In contrast, the majority of IEL T cells in the appendix and intestine were CD8⁺ and contained a considerable proportion of γδ T cells.

We obtained IEL from the appendix, cecum, and colon and compared their phenotypes with each other (Fig. 2). IEL in the cecum and colon were also found to contain CD3⁺B220⁺ cells. However, the proportion of CD3⁺B220⁺ cells was the highest in the appendix. These experiments were repeated three times.

We then compared the levels of CD3⁺B220⁺ cells in the appendix among mice fed under conventional conditions, SPF conditions, and germfree conditions. The proportion of CD3⁺B220⁺ cells was found to somewhat decrease in the appendix of mice fed under conventional conditions, although the absolute numbers of CD3⁺B220⁺ cells were comparable. Under conventional conditions, the major expansion occurred in usual CD3⁺B220⁻ cells.

To directly confirm the existence of CD3⁺CD4⁻CD8⁻B220⁺CD2⁻ cells in the appendix, three-color staining for CD3, B220, and a mixture of CD4 and CD8, and that for CD3, B220, and CD2 were conducted. CD3⁺B220⁺ cells in the appendix were found to contain DN CD4⁻8⁻ cells and CD2⁻ cells.

Since many T cells in the liver (17, 18, 19) and intestine (15, 20, 21, 22, 23) are known to be generated extrathymically, athymic nude mice at the age of 8 wk were also examined (Fig. 3 A). In general, the proportion of CD3⁺ T cells was extremely low in all tested organs of nude mice. However, the phenotype of IEL T cells in the appendix of these mice was mainly CD3⁺B220⁺CD8⁺ (data not shown) CD2⁻. Except for the dominance of CD3⁻CD8⁺ cells, the other phenotype resembled those of IEL T cells in the appendix of normal mice.

Phenotypic characterization was extended to lpr mice (Fig. 3 B). lpr mice were used at the age of 5 wk before the onset of autoimmune disease and at the age of 15 wk after the onset of disease. Even before the onset of disease and splenomegaly, a large proportion of CD3⁺B220⁺CD4⁻8⁻ cells was seen in the appendix (the absolute number of CD3⁺B220⁺CD4⁻8⁻ cells in this organ was more than twice that of control mice at the same age). The only way in which they differed from those of B220⁺CD4⁻8⁻ cells in the other organs of lpr mice was CD2⁺. After the onset of autoimmune disease and splenomegaly (i.e., 15 wk), CD3⁺B220⁺CD4⁻8⁻ cells became dominant in all other tested organs. Some CD2⁺ cells were also present in organs other than the appendix.

In a series of recent studies, we have emphasized that extrathymic T cells in the liver and intestine contain forbidden clones estimated in the minor lymphocyte-stimulating antigen (Mls) system and that they exert self-reactivity against syngeneic thymocytes (as well as NK targets) (14, 16, 18, 24). In this experiment, such a possibility was examined in IEL of the appendix in parallel with MNC in the spleen, liver, and small intestine (Fig. 4). C3H/HeN mice (Mls-1^b2^a) at the age of 8 wk were used. The Vβ3⁺ and Vβ11⁺ cells are forbidden clones, whereas Vβ2, Vβ7, and Vβ8 (8.1+8.2) are nonforbidden clones in these mice. Three-color staining for CD3, IL-2Rβ, and each Vβ was conducted. By gated analysis, the proportions of Vβ⁺ cells in each fraction were determined. As shown previously (14), forbidden clones were abundant in CD3^int cells (i.e., extrathymic T cells) in the liver. This was true in both Vβ3⁺ and Vβ11⁺ cells. On the other hand, CD3^high cells (i.e., thymus-derived T cells) contained only a few Vβ3⁺ and Vβ11⁺ cells in the spleen and liver. In the case of IEL in both the small intestine and appendix, such forbidden clones were abundant. The levels of nonforbidden clones, Vβ2⁺, Vβ7⁺, and Vβ8⁺, were almost the same in all tested organs.

It was then examined whether IEL in the appendix had killer activity against NK-sensitive YAC-1 targets or syngeneic thymocytes (Fig. 5 A). Liver MNC exerted cytotoxicity against both targets. In the case of the appendix, the cytotoxicity against syngeneic thymocytes was extremely high in comparison with those of other organs. Appendix IEL were also cytotoxic against YAC-1 targets. Since appendix IEL lacked any NK1.1⁺ cells (i.e., neither CD3⁻NK1.1⁺ NK cells nor CD3⁺NK1.1⁺ T cells), all of these cytotoxicities were mediated by CD3⁺ T cells in the appendix. Interestingly, IEL in the intestine as well as splenocytes had very low cytotoxic activity against YAC-1 targets, while IEL in the intestine had some cytotoxic activity against syngeneic thymocytes.

The morphology of IEL in the appendix and IEL in the intestine was then compared using electron microscopy (Fig. 6). It was found that all IEL in the appendix were large granular lymphocytes (Fig. 6,A). CD3⁺B220⁺ cells had a larger number of cytoplasmic granules than did CD3⁺B220⁻ cells (data not shown). IEL in the intestine had fewer granules and were smaller in size than those in the appendix (Fig. 6,B). In the final portion of this experiment, immunohistochemical staining for c-kit⁺ cells in the appendix was performed. The impetus for this study was recent evidence that c-kit⁺ cells for their own extrathymic T cells exist in the liver and intestine (25, 26). We found that many c-kit⁺ cells were present in the lamina propria (Fig. 6,C) and at the cryptopatches (Fig. 6 D), similar to the case of the intestine. In the cryptopatches, c-kit⁺ cells tended to form clusters.

Characterization of c-kit⁺ cells yielded by the appendix was further performed (Fig. 7). MNC were isolated from the liver, spleen, small intestine, and appendix. In the intestine and appendix, MNC were isolated from both intraepithelial sites (i.e., IEL) and the lamina propria (i.e., LPL). As shown in Figure 7 A, it was found that the appendix was a site with an abundance of Lin⁻ c-kit⁺ cells; this was true for both IEL and LPL. However, the majority of LPL in the appendix were B220⁺ B cells, and there were no CD3⁺B220⁺ T cells.

It was finally investigated whether c-kit⁺ cells isolated from the appendix were able to reconstitute T cells in various organs of scid mice (Fig. 7 B). Originally, scid mice did not carry any TCR-αβ⁺ and TCR-γδ⁺ cells and carried only a few CD3, CD4, and CD8 cells (<2%) in all tested organs. On the other hand, 1 mo after an injection of Lin⁻ c-kit⁺ cells of the appendix (2 × 10⁵ cells) into 4-Gy irradiated scid mice, many CD4⁺ and CD8⁺ cells had been newly generated in all tested organs. One point to be noted was a failure of the reconstitution of double-positive (DP) CD4⁺8⁺ cells in the thymus. This experiment was repeated five times. Although the majority of the CD3⁺ cells that appeared in the intestine and appendix were B220⁻, there was a significant proportion of CD3⁺B220⁺ T cells in IEL of both organs. Such CD3⁺B220⁺ cells were very few in all other organs.

A transfer experiment of Lin⁻ c-kit⁺ cells obtained from IEL of the appendix was also conducted by using B6.Ly-5.1 congenic mice (Fig. 8). Namely, 2 × 10⁵ Lin⁻ c-kit⁺ cells of the appendix of B6.Ly-5.1 mice were i.v. injected into 4-Gy irradiated B6.Ly-5.2 mice. One month after such transfer, Ly-5.1⁺ cells appeared in various organs of recipient B6 mice. Gated analysis revealed that the majority of donor cells belonged to T cells and they were mainly CD3⁺B220⁺, especially in the appendix. These results confirmed the data of cell transfer experiment in scid mice.

In the present study, we demonstrated for the first time that the appendix, especially in the intraepithelial region, is a site with an abundance of CD4⁻8⁻B220⁺ αβ T cells in normal mice. Although such DN B220⁺ αβ T cells have been known to be the major T cell population expanding in lpr mice (1, 2, 3, 4, 5, 6, 7), they have never been detected in great numbers in normal mice. It is speculated that the property of CD2⁻ among CD4⁻8⁻B220⁺ αβ T cells in the appendix (as well as lpr T cells) may be due to the resemblance to intestinal T cells. Thus, all IEL T cells in the intestine were CD2⁻. In a previous study, we reported the fact that DN T cells in lpr mice are mainly generated through the extrathymic pathway in the liver or at the other extrathymic sites (9, 10, 11, 12). This speculation seems to be reasonable, because the appendix in athymic nude mice also comprises B220⁺ T cells.

In the case of lpr mice, the number and proportion of CD4⁻8⁻B220⁺ αβ T cells in the appendix were extremely high even before the onset of autoimmune disease. However, after the onset of disease, such DN B220⁺ αβ T cells became prominent in all tested organs of lpr mice. In a preliminary experiment, we conducted appendectomy in neonatal (3 days) and young (4 wk) lpr mice. In both cases, it was not possible to ameliorate the autoimmune disease, although some lpr mice showed a retarded onset (0–4 wk) of disease. One speculation is that the other sites also contain c-kit⁺ cells and offer microenvironments to support the differentiation of DN B220⁺ T cells in lpr mice. Depending on the microenvironments, DN B220⁺ T cells seem to change their phenotype, e.g., DN→CD8⁺, CD2⁻→CD2⁺, B220⁺→B220⁻, etc. This indication is based on findings of experiments in which c-kit⁺ cells were injected into scid mice (this study) and into MRL-+/+ mice (unpublished observation).

The majority of T cells expanding in lpr mice were CD3⁺B220⁺DN CD4⁻CD8⁻TCR^int cells, especially in the spleen and lymph nodes. However, there are some other types of cells, for example, CD3⁺B220⁺CD4⁺TCR^int cells. These T cells were dominant in the liver and thymus (thymic medulla) of lpr mice. We think that abnormal T cells expand at several sites, including the appendix, liver, and thymic medulla, of lpr mice, depending on the type of cells.

The proportion of CD3⁺B220⁺ cells as well as the absolute number of CD3⁺B220⁺ cells were extremely low in nude mice. It is speculated that they were thymus independent, but that thymic humoral factors might be important for their maximal expansion. This speculation may be reasonable because it was reported that the expansion of even extrathymic T cells in the liver and intestine was under the control of thymic influence (27, 28). Indeed, the numbers of T cells in nude mice were very low in all immune organs, even at extrathymic sites (e.g., the liver and intestine).

We have emphasized that forbidden T cell clones estimated in the Mls system are always confined to CD3^int cells generated in the liver or through an alternative intrathymic pathway (14, 17, 18, 19). This is true even in intestinal T cells (29, 30). There is no leakage of forbidden clones into CD3^high cells (i.e., thymus-derived T cells) in the periphery. In this study, we demonstrated that this was also the case with IEL T cells in the appendix (see Fig. 4). Reflecting this situation, hepatic CD3^int cells and appendix T cells mediated NK-like cytotoxicity against YAC-1 cells and autologous cytotoxicity against syngeneic thymocytes. It is interesting that intestinal T cells had a lower grade of cytotoxicity in comparison with that of the appendix. In addition to these primordial properties of self-reactivity, the morphology of appendix T cells as large granular lymphocytes is noteworthy. Taken together with their phenotype of DN B220⁺, all of these properties would appear to indicate that appendix T cells are prototype T cells (i.e., possibly the most primordial T cells in phylogeny).

In recent studies, c-kit⁺ cells were demonstrated even in the liver and intestine (25, 26). Such stem cells localize in the parenchymal space of the liver and in the cryptopatches of the small intestine. In the appendix, c-kit⁺ cells were also found to localize at both the intraepithelial sites and in the lamina propria (i.e., cryptopatches). Since the liver and appendix originated from the intestine in phylogeny (25), it is concluded that all such organs keep their own c-kit⁺ stem cells in situ. The major site of c-kit⁺ stem cells is the bone marrow in land-based organisms on land. However, the lower vertebrates before landing do not have bone marrow (31). This raises the possibility that c-kit⁺ stem cells in the liver, intestine, and appendix have a longer history than those in the bone marrow in phylogenic development. In a recent study, we demonstrated that c-kit⁺ cells isolated from the liver were pluripotent and produced erythroid, myeloid, and lymphoid cells (including DP CD4⁺8⁺ cells in the thymus) (25). In sharp contrast, c-kit⁺ cells isolated from the appendix and intestine produce only some lymphocytes and myeloid cells, but not any erythroid cells at all (our unpublished observation). One of the limited functions of c-kit⁺ cells of the appendix and intestine is also explained by the present result of the failure in the reconstitution of DP CD4⁺8⁺ cells in the thymus.

Finally, primordial T cells were found in IEL of the appendix. All sites in which primordial T cells exist express monomorphic MHC class I Ags (e.g., CD1 in the thymic medulla, HLA-F in the liver, HLA-E in the intestine, and HLA-G in the uterus) (32). It is therefore speculated that these primordial T cells recognize some self Ags in the context of monomorphic MHC Ags. Since primordial T cells preferentially mediate self-reactivity against abnormal (or rapidly proliferating) self cells, these T cells might play a role in innate immunity, as proposed by C. A. Janeway, Jr. (33). lpr mice produce such abnormal self cells, and such cells are therefore recognized by primordial T cells. Due to the abnormality of Fas Ag (6, 7, 8), such Fas⁻ targets are not killed appropriately by Fas ligand⁺ primordial T cells. Therefore, despite intensive expansion of DN CD4⁻8⁻B220⁺ αβ T cells in lpr mice, the autoimmune disorder is limited (1, 2, 3, 4, 5). In any case, the appendix is an interesting immune organ. The lymphocytes isolated from the lymphoid follicles in the appendix of normal mice were the same as those in the lymph nodes (data not shown). Therefore, the isolation of IEL from the appendix in this study was important for the detection of DN CD4⁻8⁻B220⁺ αβ T cells in normal mice.

The authors thank Mrs. Yuko Kaneko for preparation of the manuscript.