NKT Cells in the Rat: Organ-Specific Distribution of NK T Cells Expressing Distinct Vα14 Chains1 2 (2024)

https://doi.org/10.4049/jimmunol.164.6.3140

A distinct function for mouse CD1 appears to be as a ligand directly recognized by a population of T cells that express the TCR and NK1.1 (NKR-P1C) Ags (1). Similar NKR-P1-positive T cells, hereafter referred to as NKT⁵ cells, have been identified in humans and rats (2, 3, 4). Mouse NKT cells and their human counterparts recognized evolutionarily conserved CD1d molecules in the absence of exogenously added Ags. Mouse and human CD1d molecules can also present glycolipids such as α-galactosylceramide to NKT cells (5, 6, 7). A striking feature of these cells is that they use an invariant TCR α-chain (Vα14-Jα281 in mice, Vα24-JαQ in humans) paired preferentially with particular Vβ-chains (Vβ8, 7, or 2 in mice and Vβ11 in humans) (8, 9, 10). NKT cells appear to play an important role in regulating immune responses through their rapid production of large amounts of IL-4 upon stimulation with anti-CD3 Ab in vivo or IFN-γ upon stimulation with anti-NKR-P1 Ab in vitro (11, 12).

The non-MHC-encoded CD1 family has recently emerged as a novel Ag-presenting system that is distinct from either MHC class I or class II molecules. Two classes of CD1 genes have been identified (13, 14). Whereas the classic CD1 genes were absent in rats and mice, CD1D gene has been conserved through mammalian evolution including mice, rats, rabbits, sheep, and humans (14, 15, 16, 17, 18, 19). CD1d molecules, recently referred to as Group 2 CD1 molecules including human CD1d, mouse CD1d, and rat CD1d, are expressed by a wide variety of organs, appearing strongly in the intestinal epithelium, hepatocytes, epidermal cells, and to a lesser extent in thymocytes and hematolymphoid cells (20, 21, 22, 23). On the contrary, Brossay et al. reported that murine CD1d was mainly expressed on hemopoietic cells but not by the epithelium of the digestive tracts (24). Therefore, from a phylogenetic standpoint, several investigators have hypothesized that CD1d molecules have a similar functional significance in various mammalian immune systems (1, 13, 18, 25).

To elucidate the immunological significance of NKT cells and their interaction with CD1d, especially in a variety of unique disease models established in rats, it is very important to obtain information about the rat invariant TCR α-chain. This is the first report of the nucleotide and predicted amino acid sequences of the invariant TCR α-chains in species other than mice and humans.⁶ Furthermore, two novel findings—a certain degree of diversity in the rat TRAV14 genes and their tissue-specific expression—were revealed with regard to the invariant TCRα. We also showed a culture system to establish the CD1d-reactive cells from heterogenous lymphocytes, including T cells, NKT cells, and NK cells. The significance of the interaction between the two types of invariant TCR α-chains and CD1d and the organ specificity of invariant TCR α-chains are discussed.

Ten- to 14-wk-old and aged (over 12-mo-old) F344/Crj (Fischer, RT1 haplotype; lv1) and F344/Crj nude rats were used. Aged Sprague Dawley (closed colony from Charles River Breeding Laboratories, Wilmington, MA) and Wistar/Smc (RT1^l) rats were also used. Cellular DNAs were prepared from 11 different rat strains; F344/Crj (lv1), LEW/Hkm (l), Wistar/Smc (l), NIGIII (q), LEJ/Hkm (j), ALB/Hok (b), BN/Hok (n), ACI/Hkm (av1), TO/Hkm (u), WKAH/Hkm (k), and W/N/Hkm (k), and BALB/c mouse. Single-cell suspensions of hematolymphoid organs were prepared as described previously (26). Hepatic sinusoidal lymphocytes (HL) were isolated by a high pressure perfusion method as described (27). After Ficoll-Conray (density, 1.096) gradient centrifugation, cells in the interface were collected, washed in PBS, and then purified by the modified nylon fiber column method described (28). Histological examination revealed that the meshwork of silver-staining fibers that highlighted endothelial cell linings parallel with liver cell plates was partially destroyed and almost all sinusoidal lymphocytes were perfused, whereas most periportal mononuclear cells remained in the liver.

Three of the primers were based on originally reported mouse Vα14 and Jα281 sequences (29). Forward primers N323 (CAGAACAACCATGAAAAAGCGCCTGA) and N313 (CTAAGCACAGCACGCTGCACAT) were from exon 1 and exon 2, respectively. The reverse primer N314 (CAGGTATGACAATCAGCTGAGTCC) was from the 3′ region of mouse TRAJ281. Rat TRAV14 primers were prepared according to the sequences obtained by this study.⁷ Forward primers M69 (GGCTGAGGAATCAGGCAGCA) and M70 (GCTTTGGGGCTAGGCTTCTG) were from the 5′ region of exon 1, and N388 (GTGGAGCAGAGTCCCCAG) and M15 (GTCCTTCAATGCAATTACAC) was from exon 2. Rat TRAV14 type-specific primers M72 (GACAAACAAGGAAGAGAAA) for type 1 and M73 (TGCATACAAAAAGGAGACG) for type 2 were derived from exon 2. A reverse primer, M71 (CACCACACAGATGTAGGTGG), was from the 3′ end of exon 2, where the sequences were conserved among TRAV14s. TRAJ281 reverse primers were M47 (ACTCAGCTGACTGTCACACCTG), M48 (GTTCCAATTCCAAAATACAGC), and M49 (AGCTTCCCTAGAGCTGAACCTC).

Rat TCR α-chain constant region gene (TRAC) primers were also made according to reported rat TRAC sequences (30); these were a forward primer, N369 (ACCCAGAACCTGCTGGGTACCAG), and reverse primers N325 (TTGCTCTTGGAATCCAGAGC), N324 (AAAGTCGGTGAACATGCAGAGGGT), and N370 (TCAACTGGACCACAGCCTTAGCGT). For screening the TRA and TRAV14 cDNA and genomic clones, two probes, pV14J (N313/N314) and pV14 (M69/M71), were prepared by amplification of HL cDNA from Sprague Dawley rats. A Vα14-Jα281 cDNA clone (FH-II-2) was also used as a probe.

Total RNA or mRNA of hematolymphoid cells were converted to cDNA and subjected to RT-PCR with Tth DNA polymerase (Toyobo, Tokyo, Japan). HL cDNA was used for amplification with the sense primer N323 derived from the 5′ region of mouse TRAV14S1 (8) and anti-sense primer N325 from the rat Cα (TRAC) (30). The amplified fragment with an expected length of 590 bp was purified from agarose gel and subcloned into pMOSBlue vector (Amersham Japan, Tokyo, Japan). cDNA clones positive for a Vα14 internal oligonucleotide probe from exon 2 of the mouse TRAV14S1 (N313) were isolated and sequenced. To determine the 5′ untranslated regions of TRAV14, rapid amplification of cDNA ends (RACE) method and cycle sequencing were performed. Expressed TRAV14 subfamily genes, TRAV14S1, TRAV14S2, and TRAV14S3, were originally defined by sequencing TRAV14-positive cDNA clones from thymus (30 clones), HL (13 clones), and spleen (33 clones) and by close examination of their sequences. Because errors of nested PCR were sometimes observed, each subfamily was defined as the same group if only one or two nucleotide alterations existed in the amplified segments within 369 bp; moreover consideration was given to differences at the variable positions among rat TRAV14s. These clones were derived from three different lymphocyte preparations from different individual F344 rats. Three subfamilies were confirmed in the latter experiments.

Genomic DNA of F344 rats was first amplified with high-fidelity KOD DNA polymerase (Toyobo) by using M69 from the 5′ untranslated region and M71 from the 3′ end of TRAV14. Then nested PCR with M70 and M71 was conducted. Amplified fragments with expected length were subcloned and individual clones were isolated and sequenced. To estimate gene numbers and study polymorphism, Southern blot analysis was done as previously described (13).

Expression of TRAV14 was analyzed by RACE and RT-PCR. RACE was performed on thymus and HL. TRAC-positive clones were picked up with toothpicks, transferred to new agarose plates (400 colonies/plate), regrown, and blotted onto a Gene Screen Plus blotting membrane (NEN, Daiichikagaku, Tokyo). The blots were hybridized with an FH-II-2 Vα14 cDNA probe. The Vα14-positive clones were sequenced with a TRAC reverse primer (N324 or N328) or primers suitable for the vector. For RT-PCR, cDNAs derived from HL, spleen, thymus, and bone marrow cells were amplified first with N388 forward primer from exon 2 and N325 reverse primer from TRAC, and then with internal primers, M15 from exon 2 and N324 from TRAC. To detect TRAV14 subfamilies and types, sequencing and/or PCR analyses of cDNA clones were conducted with type-specific primers; M72 for type 1 and M73 for type 2.

Semiquantitative analysis of TRAV14 expression was done as previously described (31). Briefly, 5 μg of each total RNA was converted to cDNA by the reverse transcription (RT) step. These cDNAs were resuspended in 50 μl of Tris-EDTA buffer and 1 μl of these, corresponding to 100 ng of total RNA (referred to as 1×) was used for amplification. Because the efficiency of RT seemed to be different in each experiment and in each batch of total RNA, these procedures were done at the same time, in the same experimental conditions. Furthermore, serially diluted samples were amplified with β-actin primers to evaluate the quality and quantity of cDNA. Based on the intensity of β-actin bands, the amount of cDNA used for amplification could be adjusted. cDNA used for amplification corresponded to the total RNA as follows: lane 1, 100 ng (1×); lane 2, 50 ng (1/2, one second dilution); lane 3, 25 ng (1/4); lane 4, 12.5 ng (1/8); lane 5, 6.25 ng (1/16); lane 6, 3.125 ng (1/32); lane 7, 1.5625 ng (1/64); lane 8, 0.78125 ng (1/128); lane 9, 0.390625 ng (1/256); and lane 10, 0.1953125 ng (1/512). Because β-actin content is very high in the RNA preparation, serial dilution was started from a more diluted point: lane 1, 5 ng (1/20); lane 2, 2.5 ng (1/40); lane 3, 1.25 ng; lane 4, 0.625 ng; lane 5, 0.3125 ng; lane 6, 0.15625 ng; lane 7, 0.078125 ng; lane 8, 0.0390625 ng; lane 9, 0.0195312 ng; and lane 10, 0.097656 ng.

⁵¹Cr-release assay was done according to the procedure previously described (32). For the cold-target inhibition test, 50 μl of ⁵¹Cr-unlabeled cells (cold target cells) was added with 50 μl of ⁵¹Cr-labeled CD1d transfectants at ratios of 4:1, 2:1, and 1:1. Then, 100 μl of effector hepatic lymphocytes was added at an E:T ratio of 25:1. The percent specific lysis was calculated as follows: [(mean experimental cpm − mean spontaneous cpm)/(mean maximum cpm − mean spontaneous cpm)] × 100.

Monoclonal Abs against rat lymphoid cells for CD3 (1F4), CD5 (R1-3B3), CD4 (W3/25), and CD8 (R1-10B5) were made in our laboratory or obtained from Seikagaku Kogyo (Tokyo, Japan). The mAbs for NKRP-1A (10/78), TCR-Vβ8.2 (R78), TCR-Vβ8.5 (B73), TCR-Vβ10 (G101), FITC-conjugated 10/78, PE-conjugated R73, biotin-conjugated anti-rat CD44 (OX49), and streptavidin Cy-Chrome were purchased from PharMingen (San Diego, CA). The mAbs for TCR-Cβ (R73) and TCR-δ (V65) were kindly supplied by Dr. T. Hünig (University of Würzburg, Würzburg, Germany). The polyclonal Ab against rat CD1d was described previously (16). Cells were harvested after different periods of time and stained by the indirect immunofluorescence technique as described. Stained cells were analyzed and sorted by FACScan (Becton Dickinson, San Jose, CA) and FACScalibur (Becton Dickinson).

Freshly isolated HL or splenic lymphocytes (1 × 10⁴), purified using the modified nylon fiber column method, were combined with the 1 × 10⁴ or 5 × 10⁴ irradiated CD1d transfectants or irradiated parent cells for 24, 48, 72, and 96 h in flat-bottom 24-well plates in the absence of the stimulants. After the specified time, cells were pulsed for 8 h with 1 μCi [³H]thymidine and harvested. Abs to CD1d (2.5 μg/ml) were added to HL at the initiation of the incubation with the CD1d transfectants.

Four rat TRAV14 genes, referred to as TRAV14S1, TRAV14S2, TRAV14S3, and TRAV14S4, highly similar to each other and to mouse TRAV14 and human TRAV24, were identified. Deduced amino acid sequences are aligned in Fig. 1,A. Complete 5′ untranslated regions of rat TRAV14 were determined for TRAV14S3 by RACE and nucleotide sequencing (data in DDBJ accession no. AB036696). A fourth TRAV14S4 subfamily was found by genomic PCR analysis of F344 strain, but no mRNA was detected by RT-PCR analysis. The predicted TRAV14S4 protein had an amino acid phenylalanine (TTT) at codon 40 instead of cysteine (TGT) in other TRAV14s (Fig. 1A). Lack of the cysteine residue may nullify the intrachain disulfhydryl bond affecting the V domain of TRAV14S4, which in turn destroys TCR cell-surface expression or may affect the interaction of TRAV14S4 TCR α-chain with other molecules important for the trafficking of the TRAV14 to cell-surface membrane. Therefore, TRAV14S4 mRNA might be excluded from the positively selected T cell populations.

Southern blot analysis of 11 rat strains indicated that all of them have several TRAV14 genes and that there is some polymorphism of restriction fragment length (Fig. 2,A). When F344 rat cellular DNA was digested with different restriction enzymes, several bands were seen in each digest (Fig. 2B). In contrast, BALB/c mouse has only a single 1.3-kb band. Thus, the existence of multiple (at least four) genes belonging to the TRAV14 family in the F344 rat genome was confirmed.

Eleven amino acid positions, 4, 36, 40, 64, 72, 73, 75, 77, 80, 89, and 90, numbering first methionine as No. 1, were diversified among rat TRAV14 proteins (Fig. 1,A). Six positions, 4H (histidine), 72T (threonine), 73N (asparagine), 75E (glutamic acid), 77K (lysine), and 80R (arginine), were common to TRAV14S1, TRAV14S2, and TRAV14S4. Amino acids composed of complementarity determining region (CDR) 1, from 47 to 53 (TVTPFNN) according to Chotia et al. (33), were conserved among four families. However, the sequences of the CDR2 region, from 70 to 77, were obviously different and could be grouped into two types. Type 1 included TRAV14S1, TRAV14S2, and TRAV14S4 (VLTNKEEK), and type 2 included TRAV14S3 (VLAYKKET). Four of eight amino acids were diversified. In contrast, only a 3-aa stretch within the CDR2 region was different between the two mouse subfamilies: DQK (position 73–75) in mouse TRAV14S1 was substituted to HEN in mouse TRAV14S2 (Ref. 34 and Fig. 1A).

Evolutionary trees constructed by both UPGMA and neighbor-joining methods further support the categorization into two types, because type 1 TRAV14s genes had a common root and type 2 TRAV14 (TRAV14S3) was more related to mouse TRAV14s and human TRAV24 (Fig. 1D).

The J regions of TRAV14-positive cDNA clones from HL were sequenced. Almost all of the TRAV14s (proportion described in next sections) were rearranged with a new rat J α-chain, referred to here as rat TRAJ281, which is highly similar to mouse TCR Jα281 (TRAJ15) (Fig. 1,B). The percent similarity of rat TRAJ281 to its mouse and human counterparts was 76.2 and 80.9%, respectively. The V14-J281 joint consisted of a single small amino acid, alanine (encoded by GCc or GCg) or glycine (GGc, GGt, GGg) (Fig. 1C). Whereas the exact origin of these nucleotides could not be determined as no information about rat germline J region is available, the CDR3 region of rat TRAV14-J281 is hom*ogeneous, as is the case with mouse and human invariant TCR α-chains.

By RACE performed on thymus and HL, 10 of 384 TCR Cα-positive clones from thymus (2.6%) and 19 of 662 clones from HL (2.87%) contained TRAV14 sequences. Of these, none of the thymus (0%) and five of the HL (0.76%) had hom*ogeneous junctional sequences (invariant TRAV14-J281). Thus, TRAV14 expression is very low in rat T cells. To more accurately determine the level of TRAV14 expression, semiquantitative RT-PCR analysis was performed. It was highest in HL (amount scored 1), followed by spleen (1/32) and then thymus (1/64), as estimated from the visible bands appearing in the series of diluted samples (Fig. 3A).

The proportion of invariant TCRα among expressed TRAV14 genes was investigated for HL, spleen, thymus, and bone marrow by RT-PCR amplification using TRAV14 and TRAC primers. The results are summarized in Fig. 3B. In the thymus, only 12.9% of Vα14⁺ clones had invariant TCRα. The proportion of invariant TCRα among Vα14⁺ clones was increased in the periphery, 31.1% in spleen, and extremely high, up to 94.6%, in HL.

We then asked whether the two types of rat TRAV14 were expressed in a tissue-specific manner. Almost the same proportion of type 1 and type 2 transcripts were found in the thymus, liver, and bone marrow (Fig. 4,A). In particular, type 2 TRAV14S3 gene was expressed predominantly in spleen. Sixty-seven of 82 (81.7%) TRAV14 clones from the spleen were in the type 2 category. The proportion of type 1 and type 2 invariant TCRα in each organ is summarized in Fig. 4,B. In the liver, type 1 invariant TRAV14-J281 (type 1 invariant TCRα) clones were more frequently observed than type 2 (type 2 invariant TCRα) clones; 69 type 1 and 44 type 2 clones. Almost the same number of type 1 and type 2 invariant TCRα clones were observed in bone marrow (26 type 1 and 24 type 2). Of particular interest, type 2 invariant TCRα was most frequently observed in spleen (81.5%, 22 of 27 invariant TCRα, Fig. 4B). From these results, we concluded that the rat TRAV14s and their invariant derivatives showed a certain degree of tissue-specific expression.

Because little is known about lymphocyte subpopulations in the normal rat liver, we analyzed the surface phenotypes of lymphocytes isolated from the liver sinusoids, and spleen, staining them with a set of mAbs. In the studies described below we used lymphocytes, purified using a modified nylon fiber column method, which contained T cells, NK cells, and NKT cells, but not B cells or macrophages. Nearly half of the total of hepatic and splenic lymphocytes were αβT cells, 3–7% were NKT cells, and 10% were NK cells (Table I). Furthermore, the fluorescence intensity of NKR-P1A had two-peak patterns for rat T cells and NKT cells in the liver and the spleen (Fig. 5,A), like that of TCR for mouse NKT cells in the liver and the thymus (35). NKT cells existed both in the NKR^high and NKR^dull subsets. Interestingly, hepatic NKT cells existed mainly in NKR^dull subsets. The ratio of NKR^high-NKT cells to NKR^dull-NKT cells in the hepatic sinusoids and the spleen was 1:10 and 1:3, respectively. Approximately 80% of hepatic T cells expressed CD4, 17% of them expressed CD8, and 1.5% expressed neither CD4 nor CD8. Half of splenic T cells expressed CD4, and the majority of the other splenic T cells expressed CD8 (Table II). Unexpectedly, most rat hepatic and splenic NKT cells expressed CD8. Thus, most hepatic and splenic NKT cells expressed CD8 and CD44^high.

Next we studied whether there was restricted heterogeneity of TCRVβ usage in the hepatic T cells and NKT cells using FITC-conjugated mAbs for rat Vβ8.2, Vβ8.5, Vβ10, and Vβ16 and PE-conjugated anti-αβTCR mAb or PE-conjugated anti-NKR mAb. Each of Vβ8.2, Vβ10, and Vβ16 were expressed in hepatic T cells and NKT cells at ∼5 and 10%, respectively (Fig. 5B). Vβ usage of splenic T cells and NKT cells was similar to hepatic T cells and NKT cells (data not shown).

Thus, TCR Vβ8, the major partner of an invariant TCRα in mice, was not expressed dominantly in either NKT cells or T cells in the normal rat liver and spleen.

To elucidate the function of rat CD1, we first examined the cytotoxic activity of the lymphocytes to CD1d transfectants in the thymus, lymph nodes, spleen, and hepatic sinusoids by CTL assay in vitro. As shown in Fig. 6,A, hepatic lymphocytes killed CD1d transfectants, but did not kill the parent cell line. Splenic lymphocytes had relatively low cytotoxicity. Thymocytes and lymph node cells did not show any cytotoxity. Cold (i.e., cells that had not been labeled with ⁵¹Cr) CD1d transfectants and CD1⁺ thymocytes were able to compete for lysis of hot (labeled with ⁵¹Cr) CD1d transfectants (Fig. 6B). These results indicated that hepatic lymphocytes directly recognized CD1d expressed on killing targets.

Thus, among adult rat lymphocytes (the mixture of T cells, NK cells, and NKT cells) in the thymus, spleen, liver sinusoids, and lymph nodes, only the cells from the liver sinusoids cytolyzed CD1d transfectants significantly in a CD1d-restricted manner.

Incubation of HL with CD1d transfectants for 72 h in the absence of exogenously added cytokines resulted in proliferation as assessed by incorporation of [³H]thymidine (Fig. 7,A). The reaction was blocked by addition of anti-CD1 Ab at the initiation of the incubation. In contrast, cultures of HL-containing medium alone showed progressive cell death over the duration of the culture. In cultures with the parent cell line, most surviving cells were NK cells by 24 h, and apoptosis was nearly complete by 72 h. In striking contrast to freshly isolated HL, the majority of surviving cells in cultures with CD1d transfectants expressed Vβ8.2 at 72 h (Fig. 7B).

Invariant rat Vα14-Jα281 expression was detected by RT-PCR in HL in cultures with CD1d transfectants at 96 h (Fig. 7C). However, TCR Cα, Vα14-Cα, and Vα14-Jα281 were not detected in HL in cultures with the parent cells at 48, 72, or 96 h. TCRγ expression was not detected in any of the cases.

To elucidate the possible roles of different types of TRAV14s in different organs, the type of TRAV14 expressed in the proliferating hepatic and splenic NKT cells by the stimulation of CD1d without exogenous Ag was studied using the quantitative PCR method and direct sequencing. The ratios of type 1 to type 2 TRAV14 expressed on proliferating NKT cells in this system were similar to those in freshly isolated Vα14⁺ NKT cells both in the spleen and liver sinusoids (Fig. 7D).

We, for the first time, identified novel rat TRAV family genes encoding invariant TCR α-chains highly similar to mouse TRAV14-J281 and human TRAV24-J18. Remarkable hom*ogeneity of their CDR3 regions implies that a part of the ligand molecules (carbohydrate or protruding portion) bound to the CD1d facing them is mainly constant in rats as in mice and humans. Multiplicity of rat TRAV14s may correspond to a previous notion by Southern blot analysis, in which a duplication event encompassing a large portion of the TRAV locus occurred in the rat (36). This is in extreme contrast to mouse TRAV14, as most laboratory mouse strains harbor a single-copy gene; mouse TRAV14S1 in 17 strains or mouse TRAV14S2 in 4 strains. Only two strains, DBA/1 and DBA/2, had both genes (29). This multiplicity was not a simple numerical expansion, rather the rat TRAV14 genes are more diverse and polymorphic than previously inferred from the studies of mice and humans, in which the invariability of the V domain has been emphasized (8, 9, 10, 29). The rat TRAV14s were divided into two types based on the diversity in the CDR2 region. Both types of invariant TRAV14-positive NKT cells, showing tissue-specific distribution, existed in rats and were simultaneously proliferated by and cytolysed to CD1d-expressing transfectants. Such CD1d-reactive NKT cells were enriched in the adult rat liver sinusoids. Whereas most features were common to mouse NKT cells, several characteristics were unique in the rat. For example, skewed TCR Vβ usage was not very obvious for freshly isolated rat NKT cells (Fig. 5,B; and A. Matsuura and M. Kinebachi, unpublished observation of Vβ expression analysis by RT-PCR). However, the majority of CD1d-responding cells after coculture were Vβ8.2-positive invariant TRAV14-J281-positive NKT cells (Fig. 7, B and C). Such a bias of Vβ repertoire of rat TRAV14⁺ cells may be due to stimulation through CD1d molecules under the different background of species. Furthermore, most rat NKT cells were CD8 positive. To further investigate rat TRAV14⁺ cells, specific Abs against recombinant TRAV14 protein is in preparation.

Proliferation and survival of hepatic NKT cells upon CD1d stimulation suggested that CD1d may play a role in hemopoiesis in the adult rat liver via the expansion or maintenance of NKT cells, in addition to the necessity of CD1d for the differentiation of mouse NKT cells in their early development and in the adult thymus (37, 38, 39, 40). As these autoreactive rat NKT cells have cytotoxic potential, as in mice (41), they may also contribute to tissue homeostasis, such as ordinary turnover of hepatocytes that strongly express CD1d molecules (22). Rat CD1d in the liver may also act as Ag-presenting molecules in emergent needs for protection against microorganisms from the intestine and blood. Besides hepatocytes, the kind of cells that express CD1d in the normal and inflamed rat liver needs to be determined.

Based on the crystal structures reported (42, 43), an insight into the molecular interaction between TCR, CD1d, and ligands was gained from the observation that “two types” of rat invariant TRAV14-positive NKT cells showed tissue-specific distribution. Because the variability was found in the CDR2 region, one can speculate that interaction surfaces created by the CD1 α2 helix and a part of the ligands might be different from organ to organ. Park et al. reported that autorecognition of mouse CD1 molecules by T cell hybridomas expressing the invariant TCR α-chains was highly dependent on the cell types in which mouse CD1 was expressed (44). Because most mouse strains have only one type of invariant TCRα, the fine specificity difference may be explained by the variety of self ligands and by diversity of the paired TCR β-chain. In such a situation, perhaps also in humans, the opposite side of the surfaces formed by CD1α1 helix and ligands is important for perception. Upon coculture with CD1d-transfected cells, the proportion of type 1 and type 2 invariant NKT cells from the liver and that of spleen were unaltered, indicating that the destination of types has already been determined in the origin of tissues or that self ligands/CD1d of transfectants have had insufficient force to drive them into distinctly differentiated pathways. Whether tissue-specific Ags loaded onto a hydrophobic ligand-binding groove of CD1d or tissue-specific modification of CD1d itself is responsible for the restriction of the types of TRAV14 needs to be clarified.

Recently, cellular GPI has been identified to be a major ligand of mouse CD1d1 by mass spectrometry and metabolic radiolabeling analysis (45). By assumption, >90% of CD1d1 was occupied by GPI. However, it has been shown that glycolipids, such as phosphatidylinositol, a synthetic PIM₂ (a phosphatidylinositol with two additional α-D mannose groups), and lipoarabinomannan purified from M. tuberculosis could not stimulate unprimed spleen cells from TAP^−/− mice which had TL- and CD1-dependent cells but did not have TAP-dependent, MHC class I-dependent cells. In contrast, α-galactosylceramide could do so (46). Because the α-anomeric form of glycolipids are not natural constituents of mammals, other hydrophobic ligands with certain degrees of heterogeneity might be associated with some fractions of CD1d molecules and might shape autoreactive cell pools in each organ. In vivo natural ligands of CD1d that are capable of stimulating lymphocytes should be clarified.

Athymic nude rats had a relatively high proportion of TRAV14 transcripts with a variable CDR3 region and a different kind of invariant TRAV-J transcript was found (H.-Z. Chen and A. Matsuura, unpublished observation). Our study revealed that NKT populations in rats were more variable than in mice, suggesting they have novel roles in nature.

Copyright © 2000 by The American Association of Immunologists

2000