Through analysis of athymic (nu/nu) mice carrying a transgenic gene encoding GFP instead of RAG-2 product, it has recently been reported that, in the absence of thymopoiesis, mesenteric lymph nodes and Peyer’s patches (PP) but not gut cryptopatches are pivotal birthplace of mature T cells such as the thymus-independent intestinal intraepithelial T cells (IEL). To explore and evaluate this important issue, we generated nu/nu mice lacking all lymph nodes (LN) and PP by administration of lymphotoxin-β receptor-Ig and TNF receptor 55-Ig fusion proteins into the timed pregnant nu/+ mice that had been mated with male nu/nu mice (nu/nu LNP− mice). We also generated nu/nu aly/aly (aly, alymphoplasia) double-mutant mice that inherently lacked all LN, PP, and isolated lymphoid follicles. Although γδ-IEL were slightly smaller in number than those in nu/nu mice, substantial colonization of γδ-IEL was found to take place in the intestinal epithelia of nu/nu LNP− and nu/nu aly/aly mice. Notably, the population size of a major CD8αα+ γδ-IEL subset was maintained, the use of TCR-γ-chain variable gene segments by these γδ-IEL was unaltered, and the development of cryptopatches remained intact in these nu/nu LNP− and nu/nu aly/aly mice. These findings indicate that all LN, including mesenteric LN, PP, and isolated lymphoid follicles, are not an absolute requirement for the development of γδ-IEL in athymic nu/nu mice.
Over the past 2 decades, it has been revealed that numerous intestinal intraepithelial T cells (IEL)3 have cellular and behavioral characteristics distinct from those of thymus-derived peripheral T cells (1, 2, 3, 4, 5, 6, 7, 8). In mice, IEL are enriched with TCR-γδ T cells (γδ-IEL) (9, 10), and virtually all γδ-IEL and many αβ-IEL, unlike thymus-derived CD8αβ T cells that use the ζ-chain as part of their CD3 complex, express the unique CD8αα homodimer (11, 12, 13, 14) and can use the FcRγ-chain in place of the ζ-chain (15, 16, 17). Along these findings, growing evidence has indicated thymus-independent (TI) development of such CD8αα-expressing IEL (TI-IEL) (5, 7, 11, 12, 18). Detection of RAG-1 and RAG-2 transcripts (12, 19, 20, 21, 22) and identification of a small number of T-lineage-committed TCR− lymphocytes in IEL from wild-type mice (2, 12, 19, 20, 23, 24, 25) supported the concept of localized development of IEL in the epithelial layer in situ. However, it should be pointed out that the original view of extrathymic generation of CD8αα+ αβ-IEL is now inconsistent with the results of recent studies in which the thymus-dependent generation of every αβ-IEL, including the CD8αα-expressing subset, is unequivocally demonstrated (26, 27).
Our search (28) for anatomical sites of IEL generation revealed multiple tiny clusters filled with ∼1000 c-Kit+IL-7R+Lin− (Lin, lineage markers) lymphohemopoietic cells in the lamina propria (LP) of the intestinal crypt (cryptopatches (CP)). Data obtained through a series of CP studies strongly indicated that CP were essential sites for the extrathymic development of precursor T cells destined to become TI-IEL (22, 28, 29, 30). Specifically, the presence of both TCR-γ and -β germline transcripts in the c-Kit+IL-7R+Lin− CP lymphocytes (30) has emphasized that various DNA recombination enzymes are able to approach these chromosomal segments to commence the region-specific recombinations (31, 32, 33). On the whole, these findings lend strong support to the idea that T lineage-committed precursors, which match the developmental stage of triple-negative c-KithighCD44+CD25low/− thymocytes before pre-Tα gene transcription (34, 35), but after expression of CD3ε-specific mRNA (35, 36), are present in gut CP (30). One impediment to this conclusion has been the detection of a marginal level of RAG-2 transcripts for CP lymphocytes (22). However, the analysis of athymic (nu/nu), bone marrow (BM) chimeric mice revealed that the development of donor BM-derived TI-IEL proceeded through several consecutive steps (30). BM-derived TCR− IEL first appeared within villous epithelia overlying the regenerated CP filled with BM-derived c-Kit+IL-7R+Lin− cells. These TCR− IEL subsequently emerged throughout the epithelia, and thereafter, conversion of TCR− to TCR+ IEL, the final step, took place very slowly. These results in conjunction with above-mentioned findings (2, 12, 19, 20, 21, 22, 23, 24, 25) have led us to conclude that TI-IEL complete their late maturational events, such as RAG-mediated TCR gene rearrangement, at a very slow rate in the epithelial layer in situ.
Recently, however, a new scenario for the extrathymic development of TI-IEL in nu/nu mice was described (37). By assessing RAG-2 expression in transgenic (Tg) nu/nu mice carrying a bacterial artificial chromosome encoding a GFP reporter instead of RAG-2, it was demonstrated that extrathymic T lymphopoiesis occurred mainly in mesenteric lymph nodes (MLN) and less in Peyer’s patches (PP), but not in CP (37). To evaluate these new and important findings, we generated nu/nu mice that lacked all LN and PP by administration of lymphotoxin-β receptor (LTβ-R)-Ig and TNF-R55-Ig fusion proteins into pregnant nu/+ mice (38) and double-mutant nu/nu aly/aly (aly, alymphoplasia) mice that lacked all LN, PP, as well as newly identified intestinal isolated lymphoid follicles (ILF) (39). We confirmed that these two kinds of mice harbored numerous γδ-IEL in the epithelial compartments of the small intestines and were found to retain gut CP. The significance of these findings is discussed from the viewpoint that all LN, PP, and ILF are dispensable anatomical sites for the generation of TI-IEL in the athymic nu/nu condition.
Materials and Methods
BALB/cA Jcl nu/nu (nu/nu), BALB/cA Jcl nu/+ (nu/+), aly/aly Jcl mutant and C.B-17/Icr Jcl scid/scid (scid/scid) mice were purchased from CLEA Japan (Tokyo, Japan). TCR-γδ (KN6)-Tg mice on the BALB/c background were described previously (40). Female KN6-Tg mice were crossed with scid/scid mice to generate KN6 scid/+ mice, then they were backcrossed with scid/scid mice to obtain KN6 scid/scid mice. The presence of KN6-Tg was determined by PCR analysis of tail DNA with a set of primers to the KN6 Tg (5′-CAGATCCTTCCAGTTCATCC-3′ and 5′-CAGTCACTTGGGTTCCTTGTCC-3′), and the homozygous scid/scid genotype was determined by the absence of TCR-αβ+ T cells in PBL. We generated transplacentally manipulated nu/nu and nu/+ mice that lack all LN and PP according to essentially the same method described previously (38). In brief, timed-pregnant nu/+ mice that had been mated with male nu/nu mice were i.v. injected with 200 μg of both LTβ-R-Ig and TNFR-55-Ig fusion proteins on gestational days 13 and 16. All mice used for experiments were between 8 and 18 wk of age, and absence of a thymus in various athymic mice was checked at necropsy. All animal procedures described in this study were performed in accordance with the guidelines for animal experiments of Keio University School of Medicine.
Production of nu/nu aly/aly mice and genotyping of aly mutation
Because nu/nu and aly/aly mothers are incapable of nursing the neonates, we used an in vitro fertilization technique (41) to produce (nu/nu×aly/aly)F1 hybrid mice, then these heterozygous nu/+ aly/+ mice were intercrossed to obtain nu/nu aly/+, nu/nu aly/aly, and nu/+ aly/+ nu/+ aly/aly littermates. To determine aly/aly, aly/+, and +/+ alleles, the TaqMan assay of tail DNA was performed using the ABI PRISM 7000 sequence detection system (PerkinElmer) as previously described (42). The primer sequences for aly were 5′-GCCTACTGACATCCCGAGCTA-3′ (forward primer) and 5′-GCAGGACTGGGCTGGAAGA-3′ (reverse primer). The oligonucleotide probe corresponding mutant aly allele was 5′-AGACCGTACTGTTGAAG-3′ (FAM labeled), and the oligonucleotide probe corresponding wild-type allele was 5′-AGACCGTACCGTTGAA-3′ (VIC labeled). Underlining in sequences indicates point mutation. The 3′ end of each probe carried the quencher that suppressed the fluorescence of the reporter dyes. Each DNA sample was amplified with the TaqMan Universal master mixture containing AmpliTaq Gold DNA polymerase according to the manufacturer’s instructions (Applied Biosystems). PCR conditions were 2 min at 50°C, 10 min at 95°C, 15 s at 95°C, and 1 min at 60°C for 40 cycles. During PCR, fluorescence developed when the oligonucleotide hybridized to perfectly matching DNA, and the exonuclease activity of Taq polymerase separated the quencher from the reporter dye. After PCR, the fluorescence yield for the two different dyes was measured and presented in a two-dimensional graph.
The following mAbs, described previously (22, 28, 29, 30, 39), were used. For immunohistochemical and immunofluorescence stainings: anti-γδ (GL-3), anti-c-Kit (ACK-2), anti-B220 (RA3-6B2), and anti-IgA (C10-3) were used. For flow cytometric analysis, FITC-conjugated anti-αβ (H57-597), anti-γδ (GL-3), anti-Vγ1 (2.11; gift from Dr. S. Tonegawa, Center for Learning and Memory, MIT, Cambridge, MA), anti-Vγ4 (UC3-10A6), anti-Vγ7 (GL-1; gift from Dr. L. Lefrancois, Department of Medicine, Division of Immunology, University of Connecticut Health Center, Farmington, CT), anti-CD4 (GK 1.5), biotinylated anti-γδ (GL-3), anti-B220 (RA3-6B2), anti-CD8α (53-6.7), and anti-c-kit (ACK-2), and PE-conjugated anti-CD8β (53-5.8) and anti-CD4 (GK 1.5) were used.
Longitudinally opened small intestine, ∼10 mm in length, was pasted on a filter paper to form a horizontal section and then embedded in OCT compound (Tissue-Tek; Miles) at −80°C. The tissue segments were sectioned with a cryostat at 6 μm, and sections were preincubated with Block-Ace (Dainippon Pharmaceutical) to block nonspecific binding of mAbs. The sections were then incubated with hamster (anti-γδ) or rat (anti-c-Kit, anti-B220, or anti-IgA) mAb for 30 min at 37°C and rinsed three times with PBS, followed by incubation with biotin-conjugated goat anti-hamster IgG Ab (5 μg/ml; Cedarlane Laboratories) or with biotin-conjugated goat anti-rat IgG (5 μg/ml; Cedarlane Laboratories). Subsequently, the sections were washed three times with PBS, then incubated with avidin-biotin peroxidase complexes (Vectastain ABC kit; Vector Laboratories). Histochemical color development was achieved with Vectastain 3,3′-diaminobenzidine substrate kit (Vector Laboratories) according to the manufacturer’s instructions. Finally, the sections were counterstained with hematoxylin for microscopy. Endogenous peroxidase activity was blocked with 0.3% H2O2 and 0.1% NaN3 in distilled water for 10 min at room temperature. Tissue sections incubated with either nonimmune hamster serum or isotype-matched normal rat IgG showed only minimal background staining.
Tissue segments from thymus, spleen, inguinal LN, MLN, and PP from KN6 scid/scid mice were embedded in OCT compound at −80°C. The small intestine of KN6 scid/scid mice was longitudinally opened along the mesenteric wall, then intestine, ∼10 mm in length, that had been rolled to form a vertical section was embedded in OCT compound at −80°C. Cryostat tissue sections, 6-μm thick, were fixed in acetone for 10 min at room temperature, washed three times with PBS, then pretreated with Block-Ace. Subsequently, the sections were incubated with anti-c-Kit mAb (ACK-2) for 60 min at 4°C, followed by incubation with PE-conjugated goat F(ab′)2 anti-rat IgG (H+L) (Invitrogen Life Technologies). The sections were then incubated with anti-γδ mAb (GL-3) and counterstained with FITC-conjugated goat anti-hamster IgG (H+L) (Jackson ImmunoResearch Laboratories). Finally, the sections were examined under a fluorescence microscope (Axiovert 100; Carl Zeiss) equipped with an image analysis system (Signal Analytics).
IEL were isolated according to methods described previously (22). Lymphoid cells were incubated first with biotinylated mAb, then with streptavidin-PE (BD Biosciences) and FITC-conjugated second mAb. Stained cells were suspended in staining medium (Hanks’ solution without phenol red, 0.02% NaN3, and 2% heat-inactivated FBS) containing 0.5 μg/ml propidium iodide and analyzed using FACScan with CellQuest software (BD Biosciences). Dead cells were excluded by propidium iodide gating. Three-color analysis of IEL was also performed. IEL were incubated first with anti-CD8α mAb (biotinylated), then with streptavidin-Tri-Color (Caltag Laboratories). After washing, IEL were counterstained with two combinations of two PE-conjugated mAbs (anti-CD8β and anti-CD4 mAbs) and one FITC-conjugated mAb (anti-γδ), respectively. Lymphoid cells were incubated with anti-FcγR II/III mAb (2.4G2) before staining to block nonspecific binding of labeled mAbs to FcR.
Development of γδ-IEL in nu/nu mice is independent of all LN and PP
To explore whether MLN are essential anatomical sites for the generation of TI-IEL in athymic nu/nu mice (37), we generated nu/nu mice that lacked all LN and PP. Pregnant nu/+ female mice that had been mated with nu/nu male mice were injected with LTβ-R-Ig and TNF-R55-Ig fusion proteins according to the protocol described by Rennert et al. (38), and the presence or the absence of LN and PP was determined in the progeny at 8 wk of age under a stereomicroscope. Although PP were absent from all treated mice, markedly attenuated remnants of MLN were present in about one-fifth of them. However, in every MLN-deficient nu/nu and nu/+ offspring, the development of mandibular, axillary, inguinal, and popliteal (data not shown) LN (i.e., peripheral LN) and cervical (data not shown), iliac, and sacral LN (i.e., mucosal LN) was also ablated (LNP− mice; Fig. 1).
Flow cytometric analysis of IEL isolated from nu/nu, nu/nu LNP−, nu/+, and nu/+ LNP− mice was performed using anti-TCR-αβ and anti-TCR-γδ mAbs. Consistent with well-established findings (3, 4, 8, 43), the proportion of αβ-IEL to γδ-IEL was sharply reduced, and the composition of IEL not expressing either type of TCR was expanded in athymic nu/nu conditions regardless of the presence or the absence of all LN and PP (Fig. 2,A). In contrast, no significant differences in absolute numbers of IEL were observed between LNP− and control LNP+ mice (data not shown). Notably, although the population size was slightly smaller by a factor of ∼1.5 compared with that in IEL from control nu/nu mice, a large number of γδ-IEL was detected in IEL from nu/nu LNP− mice (Fig. 2,A). Compartmentalization of γδ-IEL within the epithelial layer of small intestine in nu/nu LNP− mice was also verified by immunohistochemistry (Fig. 2 B). These results indicate that the development of γδ-IEL per se is independent of thymus, all LN, and PP.
With these findings in mind, we examined whether CP and ILF were present in these in utero manipulated LNP− mice, because, in contrast to PP that are already microscopically well developed just before birth (44), organogenesis of CP (28) and ILF (39) commences in early postnatal life. In fact, it was corroborated that the development of CP filled with closely packed c-Kit+ lymphocytes (Fig. 2,C) and ILF containing B220+ B cell aggregation (Fig. 2 D) remained intact in these nu/nu LNP− and nu/+ LNP− mice.
Development of γδ-IEL in nu/nu aly/aly double-mutant mice
To ascertain the universality of the above findings, we explored the development of IEL in a mutant mouse that inherently lacked thymus, all LN, and PP, because fusion protein-treated LNP− mice might possess a minute and stereomicroscopically invisible MLN, even though this possibility appeared to be remote (38). With this purpose in mind, we generated nu/nu aly/aly double-mutant mice. In accordance with the earliest description (45), nu/nu aly/aly mice were devoid of all LN and PP (data not shown) as well as thymus. Importantly, substantial colonization of γδ-IEL in the small intestine of nu/nu aly/aly mice was verified by flow cytometric (Fig. 3,A) and immunohistochemical (Fig. 3,B) analyses. Furthermore, as inferred from our previous observations (28, 39), histogenesis of CP was detected (Fig. 3,C), whereas development of ILF was completely blocked (Fig. 3,D), in these double-mutant animals. Taking all of these results together (Figs. 2 and 3), neither thymus, all LN including MLN, PP, nor ILF is an absolute requirement for the development of γδ-IEL.
In this context, it is important to determine T and B cells and IgA+ B cells that sojourn, respectively, in the spleen and LP of nu/nu LNP− and nu/nu aly/aly mice, because spleen is most likely the sole organized peripheral lymphoid tissue remaining in these animals, and nu/nu aly/aly mice lack intestinal IgA+ B cell-producing plants such as PP and ILF. In contrast to abundant B220+ B cells, mature T cells were virtually absent in the spleens of nu/nu LNP− (Fig. 4,A) and nu/nu aly/aly (Fig. 4,C) mice. In the small intestines, however, nu/nu LNP− mice possessed IgA+ B cells (Fig. 4,B), γδ-IEL, CP, and well-developed ILF (Fig. 2), whereas nu/nu aly/aly mice possessed γδ-IEL (Fig. 3, A and B) and CP (Fig. 3,C), but lacked IgA+ B cells (Fig. 4,C) and ILF (Fig. 4 D). These results indicate that all LN, including MLN, PP, and ILF, are not an absolute requirement for the generation of γδ-IEL in nu/nu mice and that the aly mutation interferes the formation of PP and ILF, resulting in the impaired development of IgA+ B cells in villous LP.
Phenotypic and Vγ gene usage analyses of γδ-IEL in nu/nu LNP− and nu/nu aly/aly mice
The data reported to date indicate that γδ-IEL are potentially capable of developing in nu/nu mice lacking all LN, PP, and ILF. In this regard, however, it is reasonable to consider the possibility that γδ-IEL generated under such harsh conditions might differ from those generated in nu/nu mice possessing all LN, PP, and ILF. To address this issue, we conducted flow cytometric analysis of γδ-IEL isolated from nu/nu, nu/nu LNP−, and nu/nu aly/aly mice. Although absolute numbers of γδ-IEL were lower by a factor of 2 compared with those in nu/nu mice (Fig. 5,A), the composition of the major γδ-IEL subset expressing CD8αα homodimer (Fig. 5,A) and the Vγ1, Vγ4, and Vγ7 gene segment used by such γδ-IEL (Fig. 5 B) remained the same in both nu/nu LNP− and nu/nu aly/aly mice. Collectively, these results indicate that the absence of all LN, PP, and ILF exerts only a small effect on the phenotypic configuration of γδ-IEL in nu/nu mice.
Development of c-Kit+ γδ-IEL in scid/scid mice expressing Tg TCR-γδ
We have shown that IL-7 produced by intestinal epithelial cells (IEC) is important for intraintestinal development of γδ-IEL and is crucial for organization of intestinal mucosal lymphoid tissues, such as PP and CP (18). It has also been shown that c-Kit and stem cell factor (SCF) are expressed by γδ-IEL and IEC, respectively, and signaling through c-Kit/SCF is indispensable for normal development of γδ-IEL (46, 47). Thus, these previous (46, 47) and the present findings in conjunction with results obtained with BM chimeric mice (30), reinforce the idea that the development of γδ-IEL takes place in the intestinal mucosa in situ. In an attempt to confirm and visualize directly the cellular events that proceed toward gut-oriented γδ-IEL generation, i.e., c-Kit+ CP cells→c-Kit+ TCR-γδ T cells→c-Kit− γδ-IEL, we generated scid/scid mice expressing KN6-Tg TCR-γδ (48). Double-immunofluorescence analysis of small intestinal tissues containing CP highlighted these presumptive cellular events. Thus, a representative picture of jejunal tissue sections from KN6 scid/scid mice (Fig. 6,A) shows that a cluster of c-Kit+ cells in a CP (red) does not express TCR-γδ, and that large numbers of γδ-IELs (green) are present in the epithelial layer, especially in the epithelium adjacent to CP. Notably, quite a large number of lymphocytes expressing both c-Kit and TCR-γδ molecules (yellow or orange) is also present in the epithelial and LP compartments of villi (Fig. 6,A). Neither clustering c-Kit+ cells nor lymphocytes stained yellow/orange were detected in other lymphoid tissues, such as thymus, LN, PP, and spleen (data not shown). Flow cytometric analysis of cell surface c-Kit molecules on the gated γδ T cells confirmed that a large fraction of γδ-IEL was c-Kit positive, namely, double-positive TCR-γδ+c-Kit+ cells (Fig. 6,B, upper panel). In contrast, abundant Tg γδ T cells residing in MLN, spleen, and thymus did not include such c-Kit-expressing, double-positive cells (Fig. 6,B, lower three panels). Overall, these results support the above-described basic premise that the development of γδ-IEL takes place in the intestinal mucosa in situ. However, because KN6-Tg TCR-γδ-expressing cells have not been detected in the CP of KN6 scid/scid mice (Fig. 6 A), whether CP are essential and indispensable anatomical sites in generating γδ-IEL remained highly contentious.
Because most of the numerous lymphocytes residing in the murine IEC compartment unexpectedly turned out to be T cells (IEL), additional revelation toward the end of the last century indicated that they display phenotypic and functional characteristics distinct from those of other T cell populations and that their development does not necessarily depend on the thymus (TI-IEL) (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18). Likewise, several lines of information have illuminated the distinctive T cell facets of human fetal intestine (49, 50), and on the basis of RAG expression, there may also be TI-IEL that develop in human (51, 52) and rat (53) intestines. Although evidence for the vestigial lymphocyte-producing function of gut mucosa is substantial, as mentioned above (2, 5, 7, 11, 12, 18, 19, 20, 21, 22, 23, 24, 25, 49, 50, 51, 52, 53), recent studies have made it clear that CD8αα does not serve as a marker for TI development of αβ-IEL (26, 27). In contrast, because positive and/or negative selection of TCR-γδ T cells in the thymus is not as evident (26, 27), thymic dependency for functional TCR-γδ T cells is less obvious.
In an assessment of the expression of Tg-encoded GFP in place of RAG-2 protein, no evidence was obtained for a lymphopoietic process involving CP cells migrating into gut epithelium to undergo TCR gene rearrangement and maturation into αβ- and γδ-IEL even in the athymic nu/nu condition (37). Instead, MLN and, less efficiently, PP have been identified as the major extrathymic T cell-producing plants in athymic nu/nu mice, and the newly generated T cells migrate from MLN into thoracic duct lymph to reach the gut epithelia, indicating that MLN and PP are the pivotal sites in generating TI-IEL, mostly γδ-IEL (37). In this context, the development of γδ-IEL should be hampered in nu/nu mice that simultaneously lack MLN and PP. Our present findings argue against this scenario by showing that not only transplacentally manipulated nu/nu LNP− mice lacking all LN and PP, but also genetically defined nu/nu aly/aly mice lacking all LN, PP, and ILF harbor a substantial population of γδ-IEL (Fig. 2, A and B, and Fig. 3, A and B), although these two different mouse models may share the same confounding factors. In contrast to what was observed in nu/nu mice, the extrathymic pathway of IEL generation was shown to be totally repressed in the euthymic condition using the same GFP RAG-2 Tg mouse model (37). The authors proposed that all IEL, including CD8αα+ IEL, in normal mice were the likely progeny of double-negative (DN) TCR-αβ+ and -γδ+ thymocytes and noted that extremely complex and unusual T cell characteristics of murine IEL with respect to their expression of various accessory, costimulation, activation, and adhesion markers (4, 5) might be brought about by the distinctive microenvironment of gut epithelium (37). In this context, for instance, many DN thymocytes somehow down-regulate cell surface expression of Thy-1 molecules, because a substantial fraction of IEL is Thy-1 negative, whereas most of them must up-regulate c-Kit molecules on the way to becoming IEL (46, 47) (Fig. 6). Even if all of those transfigurations (>20) are attributable to the inherent properties of gut epithelium, the biological significance as well as the molecular level of the mechanisms underlying such enigmatic cellular events remain highly contentious. It should also be pointed out that our recent findings (18) are inconsistent with this idea (37). γδ T cells are absent in IL-7−/− mice due to the selective blockade of TCR-γ gene rearrangements (54). Using the intestinal fatty acid-binding protein (iFABP) promoter, we reinstated the expression of IL-7 to mature IEC of IL-7−/− mice (iFABP-IL7) (18). Although γδ-IEL were restored in iFABP-IL7 mice as well as CP and PP, γδ T cells remained absent from all tissues, including thymus, spleen, and skin. These results clearly indicate that γδ-IEL generated in iFABP-IL7 mice are not of DN TCR-γδ+ thymocyte origin and that the recombination of TCR-γ genes in TCR-precursor T cells takes place in situ with the assistance of IL-7 produced locally by IEC.
Our present findings provide compelling evidence for the development of γδ-IEL within the intestine of nu/nu mice that lack the thymus, all LN, PP, and ILF. It should be pointed out, however, that the population size and absolute numbers of γδ-IEL from nu/nu LNP− and nu/nu aly/aly mice are smaller than those from the corresponding control nu/nu mice (Figs. 2,A, 3,A, and 5 A). These features would indicate that LN and PP, in fact, contribute to γδ-IEL numbers in the nu/nu condition. In contrast to the results obtained in GFP RAG-2 Tg mouse model (37), our previous RT-PCR analysis of lymphocytes from normal euthymic mice showed that under conditions in which mRNA from 50 thymocytes displayed a strong signal for RAG-2 transcripts and mRNA from 6250 RAG-2−/− thymocytes failed to display any detectable signals, very low levels of RAG-2 transcripts were constantly detected in an amount of mRNA equivalent to 6250 IEL and CP cells (22). These findings suggest that a small minority of IEL and possibly CP cells also is undergoing TCR gene rearrangement, that they are able to do so with a minimum amount of RAG-2 transcripts, or both. Actually, we still do not know how many RAG-1 and -2 molecules per nucleus are required to drive the region-specific V(D)J recombinations of TCR genes. It is possible that the amount of mRNA encoding RAG-1 and -2 molecules required by thymocytes for the successful recombination of TCR genes may not be that large. It should also be pointed out that using cell fate mapping, almost all αβ-IEL have recently been shown to be the progeny of immature CD4+CD8+ thymocytes (55). Furthermore, by using elegant and sophisticated approaches (55), it has been revealed that the retinoic acid-related orphan receptors (RORγt) detected in fetal lymphoid tissue-inducer cells are also expressed in cells within gut CP, and that γδ-IEL are not the progeny of such RORγt-positive CP lymphocytes. Specifically, however, it has remained an open question whether a small, but significant, fraction of lymphocytes in CP does not retain RORγt molecules or whether almost all CP lymphocytes express RORγt (55). In any event, to substantiate the intraintestinal development of γδ-IEL in these nu/nu LNP− and nu/nu aly/aly mice, clear identification of T cells undergoing TCR gene rearrangement in the gut mucosa appears to be of critical importance.
We are grateful to Dr. L. Lefrancois for his critical reading of the manuscript. We thank N. Hosaka and M. Mori for their excellent technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by Grant-in-Aid for Creative Scientific Research (13GS0015); the Japan Society for the Promotion of Science; a Grant-in-Aid for Scientific Research on Priority Areas A; a Grant-in-Aid for the 21st Century of Excellence Program entitled Understanding and Control of Life’s Function via Systems Biology (Keio University); the Special Coordination Fund for Promoting Science and Technology, Ministry of Education, Culture, Sports, Science, and Technology; and Health Science Research Grants from the Ministry of Health, Labor, and Welfare.
Abbreviations used in this paper: IEL, intestinal intraepithelial T cell; aly, alymphoplasia; BM, bone marrow; CP, cryptopatch; DN, double negative; iFABP, intestinal fatty acid-binding protein; IEC, intestinal epithelial cell; ILF, isolated lymphoid follicle; Lin, lineage marker; LNP−, lymph node- and Peyer’s patch deficient; LP, lamina propria; LT, lymphotoxin; LTβ-R, lymphotoxin-β receptor; MLN, mesenteric lymph node; PP, Peyer’s patch; RORγt, retinoic acid-related orphan receptor; SCF, stem cell factor; Tg, transgenic; TI, thymus independent.