Comparative Contribution of CD1 on the Development of CD4⁺ and CD8⁺ T Cell Compartments¹

The CD1 molecules are cell surface glycoproteins that have been conserved throughout mammalian evolution (1, 2, 3, 4, 5, 6). The overall structure of CD1 resembles that of MHC class I molecules, with three extracellular domains (α1, α2, and α3), a transmembrane region, and a short cytoplasmic tail. The α3 domain is noncovalently associated with β₂-microglobulin (β₂m).³ Unlike classical class I molecules, CD1 is relatively nonpolymorphic and is expressed at lower levels (5, 7). Thus, CD1 molecules were classified as a member of MHC class Ib family. However, unlike most of MHC class Ib genes, CD1 genes map outside of MHC both in humans and mice (8, 9), and they are significantly divergent from other class I genes. The sequence homology between CD1 and other class I molecules is only 25–30% (5). The same degree of homology was also detected between CD1 and class II molecules, suggesting that CD1 may represent a third lineage of Ag-presenting molecules (10). Consistent with this idea, CD1 molecules have been shown to present lipid and glycolipid Ags to T cells (11, 12, 13, 14), while MHC class I and class II molecules present peptide Ags to T cells.

Mouse CD1 is encoded by two closely related genes, CD1d1 and CD1d2 (7). CD1d1 is widely expressed on cells of multiple hemopoietic lineages (15, 16, 17), including B and T cells, macrophages, and dendritic cells, while CD1d2 can be detected only on thymocytes (18, 19). Human CD1d can be detected in the apical and lateral regions of small and large intestinal epithelial cells (IEC) (20). This unique localization of hCD1d may allow recognition by intraepithelial lymphocytes (IEL). However, the expression of CD1 on mouse IEC is still controversial, as anti-CD1 mAbs differ in detection of CD1 expression on mouse IEC (15, 16, 17, 21). Unlike MHC class I molecules, the expression of CD1 in both human and mouse does not require functional TAP (22, 23).

Study of T cell development in mutant mice lacking MHC molecules revealed that MHC class I and class II molecules play a central role in the development of CD8⁺ and CD4⁺ T cells, respectively (24, 25, 26, 27, 28). Recent studies using CD1-deficient mice have shown that CD1 is essential for the development of a major subset of NK1⁺ T cells (29, 30, 31), which use an invariant TCR α-chain in conjunction with a restricted set of TCR β-chains (32, 33). These NK1⁺ T cells promptly produce large amounts of cytokines, in particular IL-4, upon primary stimulation by TCR engagement (34). However, the role of CD1 in the development of other T cell subsets was unclear. Due to the presence of other MHC class I and class II molecules in CD1^o mice, no significant changes in either CD4⁺ or CD8⁺ population were detected in CD1^o mice (29, 30, 31). Yet, several lines of evidence suggested that CD1 might be involved in the development of some CD4⁺ and CD8⁺ T cells. In MHC class II-deficient mice, a small population of CD4⁺ T cells can be detected in the periphery (26, 27, 28). Many of the hybridomas derived from the CD4⁺ T cells of class II-deficient mice have been shown to recognize CD1, implicating a role for CD1 in the development of some CD4⁺ T cells (35). Although the expression of the class I molecules is reduced significantly both in β₂m^o and in TAP^o mice, the residual number of CD8⁺ T cells in TAP^o mice is slightly higher than that in β₂m^o mice (36, 37, 38). One possible source of these residual CD8⁺ T cells may be selection by TAP-independent, nonclassical class I molecules, such as CD1 and TL molecules (22, 23, 39).

In this report we have generated CD1^oII^o and CD1^oTAP^o mice to directly examine the role of CD1 in the development of CD4⁺ and CD8⁺ T cells in the thymus and peripheral lymphoid organs. In addition, the relative contributions of CD1-restricted CD4⁺ T cells and MHC class II-restricted CD4⁺ T cells in several immune responses were analyzed.

CD1-deficient (CD1^o) mice were established by homologous recombination in our laboratory as previously described (29) and were backcrossed six generations onto B6. I-Aβ-deficient (II^o) mice, provided by Dr. Steven Reiner (University of Chicago), were backcrossed five generations onto B6. TAP1^o mice were on a mixed B6 × 129 background (The Jackson Laboratory, Bar Harbor, ME). CD1^oII^o mice were generated by crossing CD1^o mice with II^o mice in B6 background. CD1^oTAP^o mice with a mixed B6 × 129 background were established by crossing CD1^o mice with TAP^o mice. Pathogen-free B6 mice were purchased from The Jackson Laboratory.

The Abs used in this study include FITC-conjugated mAbs specific for CD4 (RM4-5), TCRβ (H57-597), CD69 (H1.2F3), Vβ5 (MR9-4), Vβ6 (RR4-7), Vβ7 (TR310), Vβ8 (MR5-2), Vβ9 (MR10-2), Vβ12 (MR11-1), Vβ14 (14-2), and Vα3 (RR3-16); PE-conjugated mAbs specific for CD8α (53-6.7), NK1.1 (PK136), CD4 (RM4-5), CD44 (IM7), Vβ2 (B20.6), Vβ3 (KJ25), Vβ4 (KT4), Vβ10 (B21.5), Vβ11 (RR3-15), Vβ13 (RR12-3), Vα2 (B20.1), Vα8 (B21.14), and Vα11 (RR8–1); biotin-conjugated mAb specific for CD62L(MEL-14); and Cy-Chrome-conjugated mAbs specific for TCRβ (H57-597), CD4 (RM4-5), and Cy-Chrome streptavidin (PharMingen, San Diego, CA). The lymphocytes from perfused liver were isolated according to the method described by Goossens et al. (40). The IELs were prepared and purified through discontinuous 40/70% Percoll gradient centrifugation as described by Tagliabue et al. (41). Single-cell suspensions from thymus, spleen, and lymph node were prepared using standard procedure. Cell suspensions were stained using combinations of fluorescent-conjugated Abs and were analyzed with a Becton Dickinson (Mountain View, CA) FACS caliber flow cytometry using CellQuest software.

CD8⁺ T cells and B cells were depleted from the splenocytes of B6, CD1^o, II^o, and CD1^oII^o mice by incubating cells with CD8α-FITC and B220-FITC, then cells were incubated with avidin-magnetic beads and applied to magnetic separation (PerSeptive Diagnostics, Cambridge, MA). The cells were then sorted for CD4⁺, CD4⁺NK1.1⁺, and CD4⁺ NK1.1⁻ by FACS, resulting in a >95% pure population. Sorted CD4⁺ T cells (2.5 × 10⁴ to 1 × 10⁵ cells/well) were stimulated in anti-CD3 (2C11)-coated 96-well plate in a final volume of 200 μl of RPMI 1640 medium (supplemented with 10% FCS, 2 mM l-glutamine, 20 μM 2-ME, and 100 U/ml penicillin/streptomycin). After 48 h, the culture supernatants were harvested, and the levels of IL-4 and IFN-γ were quantitated by ELISA (PharMingen).

B6, CD1^o, II^o, and CD1^oII^o mice were immunized i.p. with 25 μg of TNP-conjugated Ficoll (Biosearch Technologies, Novato, CA) or 50 μg of TNP-LPS (provided by Dr. Guido Franzoso, University of Chicago) in 0.1% alum. Animals were bled before immunization and 10 and 14 days postimmunization. Anti-TNP-specific Abs in the sera were determined by isotype-specific ELISA. Briefly, flat-bottom microtiter plates were coated overnight at 4°C with 50 μg/ml of TNP-BSA (Biosearch Technologies) in PBS (pH 7.4). After washing three times with PBS-Tween 20 (0.5%), serial dilutions of murine sera in 10% FCS-HBSS were added to the plates and incubated overnight at 4°C. Plates were washed three times with PBS-Tween 20 before adding biotinylated goat anti-mouse isotype-specific Abs (1/250 in 10% FCS-HBSS; Southern Biotechnology Associates, Birmingham, AL). After 1-h incubation at room temperature, plates were washed three times with PBS-Tween. Alkaline phosphatase-conjugated streptavidin (1/1000; Jackson ImmunoResearch Laboratories, West Grove, PA) was then added to the plates and incubated for 25 min at room temperature. After five washes with PBS-Tween, the assays were developed with alkaline phosphatase substrate (Sigma).

Mean values were compared using Student’s t test for independent variables. Statistical significance was considered to be p < 0.05.

Prior studies have shown that in II^o mice, 5–15% of the wild-type numbers of CD4⁺ cells can be found in the spleen and lymph node (26, 27, 28). To address the question of whether CD1 is required for the development of these residual CD4⁺ T cells, we compared the CD4⁺ T cell compartment in II^o and CD1^oII^o mice. FACS analysis showed that the number of CD4⁺ T cells in liver and spleen of CD1^oII^o mice were reduced significantly compared with the corresponding population in II^o mice (Fig. 1). This reduction is most prominent in liver, where NK1.1⁺ T cells are normally prevalent (42, 43), and is moderate in spleen. CD4⁺ T cells constitute, on the average, 42.98 ± 8.8% liver lymphocytes isolated from II^o mice, while in CD1^oII^o littermates, 15.88 ± 2.36% of liver lymphocytes are CD4⁺ T cells. To examine whether the reduction of CD4⁺ T cells in CD1^oII^o mice is merely a result of the decrease in CD1-restricted NK1.1⁺ T cells, we compared the distribution of CD4⁺ NK1.1⁺ T cells and CD4⁺ NK1.1⁻ T cells in CD1^oII^o and II^o littermates. The representative CD4/NK1.1 plots (gated on TCR αβ⁺ cells) showed that both the CD4⁺ NK1.1⁺ and the CD4⁺ NK1.1⁻ populations were reduced in CD1^oII^o mice (Fig. 2). The percentage of CD4⁺ NK1.1⁺ T cells was reduced by 70–75%, and the percentage of CD4⁺ NK1.1⁻ T cells was reduced by 40–50%, respectively. These data suggested that CD1 could select not only CD4⁺ NK T cells, but also some NK1.1⁻ CD4⁺ T cells.

To determine whether CD1-dependent CD4⁺ T cells have a restricted repertoire, we compared the TCR usage in CD4⁺ populations from CD1^oII^o and II^o mice using a panel of mAbs specific for various Vβs and Vαs (Fig. 3,A). The pattern of Vβ segment usage by CD4⁺ T cells of CD1^oII^o mice was slightly different from that of II^o mice. In particular, the percentages of Vβ8- and Vβ7-expressing cells in CD1^oII^o mice are significantly lower than those in II^o mice. The reduction of Vβ8⁺ T cells can be detected in both CD4⁺NK1.1⁺ and CD4⁺ NK1.1⁻ population (Fig. 3 B), suggesting that CD1-restricted CD4⁺ T cells (both NK1.1⁺ and NK1.1⁻ T cells) preferentially use TCR with rearranged Vβ8 segments.

To compare the functional potential of CD4⁺ T cells from wild-type, CD1^o, II^o, and CD1^oII^o mice, purified CD4⁺ T cells from the four strains of mice were stimulated with plate-bound anti-CD3 in vitro. Two days later, the levels of IL-2, IL-4, and IFN-γ were measured by ELISA. Fig. 4,A shows that purified CD4⁺ T cells from II^o mice produce larger amounts of IL-4 than wild-type mice in response to anti-CD3. This is presumably due to the enrichment of CD4⁺ NK T cells in the remaining CD4⁺ population in II^o animals. Surprisingly, the residual CD4⁺ T cells from CD1^oII^o mice can be readily stimulated with anti-CD3 and produce large amounts of IFN-γ and substantial amounts of IL-4. The level of IFN-γ production in the CD1^oII^o mice is higher than that in the wild-type control animals, but lower than that in the II^o mice. In contrast, the amount of IL-4 produced by CD4⁺ T cells in CD1^oII^o is comparable to levels in wild-type mice, but much lower than that in II^o mice. CD4⁺ T cells from CD1^o mice do not produce significant amounts of IFN-γ and IL-4 in the same culture conditions, but produce significant amounts of IL-2 (Fig. 4,A). These data suggest that class II-restricted, CD1-restricted, and non-class II, non-CD1-restricted CD4⁺ T cells secrete different ratios of cytokines upon activation. Furthermore, when CD4⁺ cells are sorted into NK1.1⁺ and NK1.1⁻ populations, we found that NK1.1⁻ cells are largely responsible for IFN-γ secretion in both the CD1-restricted and CD1-independent populations (Fig. 4,B). However, CD1-restricted CD4⁺NK1.1⁻ T cells do secrete significant amounts of IL-4, in contrast to CD1-independent CD4⁺NK1.1⁻ T cells. The rapid secretion of cytokines by the residual CD4⁺ cells in CD1^oII^o mice correlates with our additional finding that this population appears to be enriched for cells that had a phenotype characteristic of activated T cells, such as CD44^highCD69^high (Fig. 4 C).

Several studies demonstrated that the response to thymic-independent Ags (TI Ags) could be regulated by T cells despite their inability to stimulate MHC class II-dependent T cell help (44, 45, 46). We therefore compared wild-type, CD1^o, II^o, and CD1^oII^o mice to evaluate the role of CD1-restricted T cells in modulating the Ab production against type I TI Ag (TNP-LPS) and type II TI Ag (TNP-Ficoll). TNP-specific Ab responses of all isotypes could be elicited in CD1^o, II^o, CD1^oII^o, and wild-type mice following immunization with TNP-LPS and TNP-Ficoll (Fig. 5). There were no significant differences in the production of anti-TNP specific IgM, IgG1, IgG2a, and IgG2b in all four types of animals upon immunization with TNP-LPS (Fig. 5,A). Immunization with TNP-Ficoll elicited higher levels of TNP-specific IgM and IgG1 Abs in both II^o and CD1^oII^o mice compared with those in control and CD1^o mice (Fig. 5 B). This finding is consistent with a previous report that immunization with TNP-Ficoll induced higher levels of TNP-specific Abs in class II-deficient animals than in control animals (47). However, we detected no statistically significant difference in the production of anti-TNP-specific IgM, IgG1, IgG2a, and IgG2b between II^o and CD1^oII^o animals. Thus, CD1-restricted T cells play little role in providing cytokines for the Ab response against these two T-independent Ags, contrasting with the essential role of CD1-restricted NK T cells in the IgG response to GPI-anchored Ag (48).

To determine whether CD1 deficiency had any effect on the development of CD8⁺ T cells and γδ T cells, we prepared lymphocytes from TAP^o and CD1^oTAP^o mice; stained them with reagents specific for CD4, CD8, TCRαβ, and TCR γδ; and analyzed them by flow cytometry. TAP^o and CD1^oTAP^o mice have similarly reduced numbers of CD8⁺ cells in thymus, spleen, and lymph nodes (Table I). Compared with TAP^o mice, the percentage of CD8⁺ T cells was increased in the liver of CD1^oTAP^o mice, presumably due to the reduction of significant numbers of CD1-restricted CD4⁺ T cells in the liver. No significant difference in the number of γδ T cells was detected between TAP^o and CD1^oTAP^o mice.

Substantial numbers of TCRαβ⁺ CD8⁺ IEL are present in TAP^o mice despite their absence in β₂m-deficient mice (37, 38). Most of the TCRαβ⁺CD8⁺ IEL in TAP^o mice express the CD8αα homodimer. This suggests that TAP-independent but β₂m-dependent class Ib molecules may be responsible for the development of this subset of IEL. To explore the possible involvement of CD1 in the development of these TAP-independent TCRαβ⁺CD8⁺ IEL, we examined the phenotype of IEL isolated from wild-type, CD1^o, TAP^o, and CD1^oTAP^o mice by flow cytometric analysis. Surface staining for TCRαβ, TCRγδ, CD4, CD8α, and CD8β showed no significant difference in the percentage of TCRαβ⁺ and TCRγδ⁺ lymphocytes between CD1^o and control littermates (Fig. 6). In the TCRαβ⁺ population, the frequency of CD8αα- and CD8αβ-bearing cells did not decrease substantially in CD1^o animals (41.5 ± 5 vs 34.1 ± 7.6% for CD8αα⁺ cells; 37.3 ± 2.2 vs 31.3 ± 2.4% for CD8αβ⁺ cells in CD1^+/+ and CD1^−/− mice, respectively). In agreement with previous reports, the absolute number of TCRαβ⁺ IEL decreased in mice lacking TAP (TAP^o and CD1^oTAP^o mice), and a compensatory increase in the number of γδ⁺ IEL was detected (Fig. 6). Within the TCR αβ⁺ IEL subset, the percentage of CD8αβ IEL was reduced significantly in both TAP^o and CD1^oTAP^o mice, and the degree of reduction is comparable between TAP^o and CD1^oTAP^o mice. In contrast to the percentage of CD8αβTCR αβ⁺ IEL, the percentage of CD8ααTCR αβ⁺ IEL did not change significantly among the four types of animals. Thus, our data suggest that CD1 does not play a major role in the development of either CD8αβ IELs or CD8αα IELs.

In summary, analysis of the composition and functional properties of T cells in CD1^o mice in II^o and TAP^o backgrounds has permitted us to examine the contribution of CD1 to the development of CD4 and CD8 subsets of T cells. Although CD1 was classified as an MHC class Ib molecule due to its association with β₂m, our data suggest that CD1 contributes significantly in selecting CD4⁺ T cells but minimally in development of the CD8⁺ subset. The limited impact of CD1 in the development of CD8⁺ T cells was probably not due to the inability of CD1 to interact with CD8, because Teitell et al. (49) have demonstrated that mouse CD1 can bind to CD8 in redirected CTL assays. The ability of CD1 to interact with CD8 was further supported by the observation that constitutive expression of CD8 in transgenic mice resulted in a major depletion of CD1-restricted NK T cells that normally express either the CD4 coreceptor or no coreceptor at all (50). It has been suggested that NK T cells, which preferentially use an invariant α-chain (Vα14Jα281), might have high affinity for CD1. Expression of CD8 in CD1-restricted NK T cells would lead to their negative selection by increasing the avidity between CD1 and TCR complexes. However, this hypothesis would not preclude CD1 from positively selecting CD8⁺ T cells that express TCRs with lower affinity to CD1. It is worth noting that CD1-restricted CD8⁺ T cells have been isolated from mice immunized with plasmid DNA containing chicken OVA and from mice immunized with a CD1 transfectant coated with CD1 binding peptide (51, 52). Perhaps due to limited self-ligands, CD1-specific CD8⁺ T cells appear to be a minimal component of CD8⁺ T cell subset.

Compared with II^o mice, CD1^oII^o mice have reduced numbers of both NK1.1⁺ and NK1.1⁻ CD4⁺ T cells, suggesting that CD1 selects both types of CD4⁺ T cells. TCR analysis showed that both NK1.1⁺ and NK1.1⁻ CD1-restricted CD4⁺ T cells preferentially express TCR with Vβ8 and Vβ7. Our data are consistent with a recent report that CD4⁺ hybridomas derived from both NK1.1⁺CD4⁺ and NK1.1⁻ CD4⁺ T cells of class II^o mice recognize CD1, although recognition may require ligands derived from different cellular compartments (53). It has been shown that most of the NK1.1⁺CD4⁺ T cells express an invariant Vα-chain with Vα14Jα281 rearrangement, while most of the NK1.1⁻ CD4⁺ T cells express different Vα-chains (53). Therefore, NK1.1⁺CD4⁺ and NK1.1⁻ CD4⁺ T cells might be selected by different sets of CD1-bound ligands, while structural features of CD1 might be responsible for preferential selection of T cells expressing Vβ8 and, to a lesser extent, for T cells expressing Vβ7. Preferential selection of Vβ5⁺ T cells by some MHC class I molecules has also been demonstrated (54).

A detectable pool of CD4⁺ T cells still exists in CD1^oII^o mice. This T cell population, presumably selected by other class I molecules, shows diverse TCR usage, suggesting recognition of heterogeneous ligands. Compared with CD4⁺ T cells in II^o mice, the residual CD4⁺ T cells in CD1^oII^o mice secrete comparable amounts of IFN-γ but significantly lower amounts of IL-4 and IL-2 in response to anti-CD3 stimulation. These data indicate that CD1-restricted CD4⁺ T cells are the major source of IL-4 production, while non-CD1-restricted CD4⁺ T cells play a substantial role in IFN-γ production in the anti-CD3 induction model. It is important to note that the amount of IFN-γ secreted by the CD4⁺ cells from CD1^oII^o mice is significantly higher than the amount produced by conventional class II-restricted CD4⁺ T cells. Thus, this subset of the CD4⁺ T cell population might play an important role in modulating the immune response in vivo.

In contrast to classical class I molecules, the surface expression of CD1 and TL does not depend on TAP (22, 23, 39). The expression of TL on IEC has been clearly demonstrated (55), whereas the expression of CD1 on IEC is controversial. Our FACS analysis showed that the IEL in CD1^oTAP^o mice are phenotypically similar to the IEL in TAP^o mice, suggesting that CD1 may not be responsible for intestinal selection of TAP-independent CD8⁺ IEL. Intrathymic selection of TCRαβ⁺CD8⁺ IEL by TL Ag seems unlikely, because thymocytes from all mice used in this study were TL⁻ (data not shown). However, it is possible that TCRαβ⁺ CD8αα T cells may be selected by TL extrathymically. Alternatively, low levels of MHC class I molecules expressed in the absence of TAP might be responsible for the selection of CD8⁺ T cells in TAP^o mice.

Recent studies indicated that NK T cells are heterogeneous (29, 56, 57). Although a major population of NK T cells are CD1 restricted, some NK T cells are restricted by other MHC molecules. Our study of CDI^oII^o mice clearly demonstrates the existence of a functionally distinct NK T cell subset. Further analyses of CDI^oII^o mice might provide information on the immunological functions and ligand specificities of these remnant CD4⁺NK T cells.

We thank Nancy Chiu and Cherita White-Morris for their critical reading of the manuscript, and Lillian Zhao and Candice Mulder for technical assistance.