Essential and Partially Overlapping Role of CD3γ and CD3δ for Development of αβ and γδ T Lymphocytes

The targeting construct was generated by standard methods. In brief, a genomic DNA clone containing a 15.5-kb fragment of CD3γδ genes was isolated from a 129/sv mouse genomic DNA library (provided by Dr. Manley Huang, GenPharm Int., Mountain View, CA) and subcloned into pBluescriptSK+ (Stratagene, La Jolla, CA). A 2.8-kb SalI-XhoI DNA fragment containing the PGK-TK^r gene was isolated from pPGK-TK (provided by Dr. Manley Huang), and ligated to the XhoI site of pPGK-hygromycine^r (hyg^r) (a gift of Dr. Richard Mortensen). A 1.9-kb XbaI-XbaI intronic fragment between exon 1 and 2 of CD3δ was obtained by XbaI digestion of the 15.5-kb CD3γδ genomic DNA fragment. And a 3-kb intronic fragment between exon 1 and 2 of CD3γ was obtained by first subcloning a 5-kb EcoRI-XbaI fragment into SK+ followed by a HindIII cut, so that a HindIII site from the polylinker region of the plasmid was transferred to one end of the 3-kb fragment. The 1.9-kb XbaI-XbaI fragment and the 3-kb HindIII-HindIII fragment were inserted into the 5′ and 3′ sites of the PGK-Hyg^r gene. In the resulting construct, a 3.1-kb DNA fragment containing the 1.4-kb intergenic DNA fragment between the CD3γ and CD3δ genes and exons 1 of both genes were replaced by the 2.8-kb PGK-Hyg^r cassette. 10 μg of purified targeting molecules were electroporated into 10⁷ J-1 ES cells. ES cells were positively selected by hygromycin-B at 200 μg/ml and negatively selected by FIAU at 0.2 μM. 355 clones were selected and examined by Southern blots for homologous recombination using a 0.8-kb (StuI-XbaI) 5′ probe located outside of the construct. Eight clones were identified as targeted clones, which were confirmed by another Southern analysis with a hyg^r probe. Four of the targeted clones were injected into the blastocysts of either C57BL/6 or BALB/C origin, and 90– 100% fur color chimerism was observed in 45 founder mice. Test breeding of the chimeras indicated that all of the males (n = 28 from 3 embryonic stem [ES] clones) transmitted the ES cell genome. Four males were mated to C57BL/6 females to generate heterozygous mice, and homozygous CD3γδ^−/− lines were obtained by sibling breeding. Identical results were obtained from homozygous CD3γδ^−/− lines of different ES clones.

Single cell suspensions of thymocytes, LN cells, spleen cells, PBL, and small intestine intraepithelial lymphocytes (iIEL) were prepared as described (26, 27). Three-color staining of the cells was performed as previously reported elsewhere (28).

Northern blot analysis was performed as described (29).

To generate mice deficient in both CD3γ and CD3δ gene expression, a 3.1-kb DNA fragment containing the promoters (25) and exons 1 of the CD3γ and CD3δ genes was replaced by a PGK-Hyg^r cassette (Fig. 1,A). The PGK-hyg^r cassette was chosen here over the PGK-neo^r cassette to prevent a possible suppressive effect of the PGK-neo^r on neighboring gene expression (30, 31). Homozygous mice carrying this mutation in the CD3γ and δ genes were generated (Fig. 1,B). Northern blot analysis demonstrated that the expression of both CD3γ and CD3δ mRNA was absent in the CD3γδ^−/− thymocytes (Fig. 2). Moreover, no aberrant expression of the truncated CD3γ or δ mRNAs were ever detected in Northern blotting of thymocytes from more than 20 CD3γδ^−/− mice. However, the expression of the neighboring CD3ε gene and the nonlinked CD3ζ was normal (Fig. 2), and pTα expression was detected (data not shown).

Total cellularity of the thymi of CD3γδ^−/− mice was 2–5% of that in wild-type or heterozygous littermates (Fig. 3,A). Flow cytometric analysis of the thymocytes showed that these cells are DN, with the majority of them being CD44⁻CD25⁺c-Kit⁻Sca-1⁺, identical to the thymocytes found in RAG^−/− mice (Fig. 3,B). Northern blot analyses of the thymocytes of CD3γδ^−/− mice did not detect the mRNA for rearranged TCR-α and TCR-β genes, whereas only the 1.0-kb germline C_β mRNA was detectable (Fig. 2). Consistent with these analyses, no mature αβ⁺ T cells were detected in the LN, the spleen, or the gut of the CD3γδ^−/− mice (Figs. 3,C and 4,C, Table 1). B cell development appeared unaffected (Table 1). Taken together, αβ T cell development in CD3γδ^−/− mice is blocked at the same DN CD44⁻CD25⁺ check point as RAG^−/− mice (32, 33).

Next, γδ T cell development in CD3γδ^−/− mice was examined. As shown in Fig. 4, A and B, γδ T cells were absent in the thymus and periphery of CD3γδ^−/− mice. Since γδ T cells normally account for only a very small fraction of thymocytes and peripheral T cells, we assessed γδ T cell development in the small intestine, where γδ T cells represent a major population of the iIEL in wild-type mice. In CD3γδ^−/− mice, γδ T cells were again nondetectable in the intestine (Fig. 4,C). However, normal number of CD8αα⁺B220⁺CD32⁺NK1.1⁻ cells, representing T cell progenitors in the gut (27) could be detected in the gut of CD3γδ^−/− mice (Fig. 4, C–E, Table 1, and data not shown). Therefore, these analyses indicate that deficiency in CD3γ and δ completely blocked γδ T cell development beyond the CD8αα⁺ stage.

We report here that in the CD3γδ^−/− double mutant mice, intrathymic development is completely arrested at the DN CD44⁻CD25⁺ prothymocyte stage, a central check point at which pre-TCR begins to mediate further thymocyte differentiation into the DP stage. This observation indicates that the function of pre-TCR is completely abrogated in CD3γδ^−/− mice. In contrast, in recently reported CD3δ^−/− mice, thymic development is undisturbed up to the DP stage (23), whereas the transition from DN to DP stages was severely but not completely blocked in CD3γ^−/− mice (24). The phenotypes of CD3δ^−/− and CD3γ^−/− mice are consistent with the biochemical evidence that CD3γ is preferentially required over CD3δ in prothymocytes for the assembly of the pre-TCR–CD3 complex (22). However, the present data revealed that CD3δ also participated in vivo in the assembly and function of the pre-TCR–CD3 complex. Moreover, small numbers of TCR-αβ⁺ T cells were detected in the periphery of CD3δ^−/− and CD3γ^−/− mice, but were absent in CD3γδ^−/− mice. These observations are consistent with the biological evidence that in mature T cells, the TCR– CD3 complex lacking either CD3γ or δ could sometimes be detected on the cell surface at reduced levels. However, no surface expression of the TCR–CD3 complex could be detected in cells lacking both CD3γ and δ (15, 16). Taken together, CD3γ and CD3δ collectively play an essential, yet partially overlapping, role in the assembly and function of the pre-TCR. It is most likely that in the absence of CD3γ and CD3δ, pre-TCR cannot be expressed on the surface of prothymocytes.

In addition to the structural requirement, CD3γ and CD3δ may regulate pre-TCR function through the signaling capacity of the immunoreceptor tyrosine-based activation motifs (ITAMs) presented in their cytoplasmic domains (34). It is known that not every ITAM plays a distinct role in pre-TCR function. For instance, pre-TCR function is competent in mutant mice deficient in the CD3ζ cytoplasmic domain (35). Moreover, the defect in pre-TCR function in CD3γ^−/− (24), CD3ζ^−/− (36), or RAG^−/− (19, 27, 37) mice can be overcome by anti-CD3ε-mediated cross-linking. However, the same anti-CD3ε treatment in vivo in CD3γδ^−/− mice failed to relieve the block at the DN check point (data not shown). Since the anti-CD3ε antibody used in all of these studies, namely 2C11 (or 500A2), binds CD3ε efficiently when either CD3γ or CD3δ is presented but poorly when both CD3γ and CD3δ are missing (38; data not shown), the lack of thymocyte differentiation upon 2C11 treatment of CD3γδ^−/− mice might be explained by the following nonexclusive possibilities: (a) pre-TCR could not be expressed on the surface of CD3γδ^−/− prothymocytes; (b) the inefficient binding of 2C11 to CD3ε on the surface of CD3γδ^−/− prothymocytes results in a weak signal that is below the threshold level for further thymic development; and (c) the cytoplasmic domains of CD3γ and CD3δ collectively play an essential role in pre-TCR function. The last possibility, nevertheless, is less likely because it has been shown that under artificial circumstances, either CD3ε or CD3ζ cytoplasmic domain alone can independently generate signals for thymocyte development to the DP stage (39). Thus, the ultimate assessment of the physiological role of the cytoplasmic domains of CD3γ and CD3δ awaits the generation of mutant mice in which the cytoplasmic domains of CD3γ and CD3δ are deleted.

An important observation of this study was that γδ T cell development was completely blocked in the CD3γδ^−/− mice. In comparison, γδ T cell development was partially blocked in the CD3γ^−/− mice and was undisturbed in CD3δ^−/− mice (23, 24). Thus, this study demonstrated that CD3δ also plays a role in regulating the development of the γδ T cell lineage, and CD3γ and CD3δ collectively are essential for γδ T cell development. Like their regulation of αβ T cell development, CD3γ and CD3δ may regulate γδ T cell development by their structural contribution and/or signaling capacity. Nevertheless, the function of CD3γ or CD3δ for γδ T cells may not be a duplication of their respective roles for αβ T cells. For instance, although surface expression of TCR-αβ is severely reduced (8–10-fold) in CD3δ^−/− mice, their TCR-γδ expression is only mildly (less than twofold) reduced (23). On the other hand, severe reduction of both TCR-αβ and TCR-γδ expression in CD3γ^−/− mice indicated a pivotal role of CD3γ in the assembly of TCR-αβ–CD3 and TCR-γδ–CD3 complexes (24). Taken together, it is likely that the complete block in γδ T cell development in CD3γδ^−/− mice was a result of the incomplete TCR-γδ–CD3 complex not being expressed on cell surface in the absence of CD3γ and CD3δ. It remains to be investigated whether the cytoplasmic domains of CD3γ and CD3δ also have distinct functions in the development of γδ T cells.

In conclusion, in the CD3γδ^−/− mice, early thymic development mediated by pre-TCR was completely blocked, and TCR-αβ⁺ and TCR-γδ⁺ T cells were absent. These observations are different from those made on either CD3δ^−/− or CD3γ^−/− mice, in which pre-TCR function was either undisturbed or incompletely blocked, as TCR-αβ⁺ and TCR-γδ⁺ T cells were detected in the periphery. Taken together, these studies demonstrated that CD3γ and CD3δ play an essential, yet partially overlapping, role in the development of both αβ and γδ T cell lineages.

This work was supported by National Institutes of Health grants CA 74233 (to B. Wang) and AI35714 and R37-17651 (to C. Terhorst).