| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Department of Immunology and Infectious Diseases, Shinshu University Graduate School of Medicine, Matsumoto, Japan; and Laboratory for Cell Signaling, RIKEN Research Center for Allergy and Immunology, Yokohama, Japan
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
A series of NK cell receptors (NKRs) including Ly49 and NKG2 appear at early stages during NK cell development, and the frequencies of Ly49+ NK cells are considered to continue to be up-regulated cumulatively as NK cells mature (4, 5, 6, 7, 8). Because these molecules are critical for the physiological functions of NK cells through inhibiting or promoting NK cell activation (9), generation and selection of the NKR repertoire should be strictly controlled in a developmentally regulated manner (4, 5), and the disturbance of such a control would have profound impacts on the functions of mature NK cells (7, 9, 10, 11).
IFN regulatory factor-2 (IRF-2) is a transcription factor known to be required for efficient NK cell development. With respect to the mechanism of NK cell deficiency in mice lacking this transcription factor (IRF-2–/– mice), IRF-2 was previously implicated in IL-15 responses of immature NK cells using an in vitro cell culture system (12), a speculation requiring confirmation under more physiological conditions. Roles of IL-15 in NK cell development and NKR expression have been firmly established through studies using mice lacking IL-15 or its receptor (13, 14, 15). However, the stages at which IRF-2 and IL-15 function as well as the functional relationship between these two molecules have not been extensively examined, and the mechanism of IRF-2 actions in the generation of mature NK cells remains to be addressed.
In this study we attempted to characterize NK cells present at reduced levels in the periphery and BM in IRF-2–/– mice in reference to those in IL-15–/– mice (13). Phenotypic examination revealed that peripheral NK cells in IRF-2–/– mice were very immature as judged by their reduced expression of CD11b, Dx5, and Ly49 receptors. In contrast, NK cells in these mice developed in the BM up to the relatively mature CD11blowDx5high stage with less compromised Ly49 repertoire, but fully mature CD11bhigh NK cells were virtually absent. We found, in addition, that developing NK cells in the BM were undergoing apoptosis at accelerated frequencies. In contrast, NK cell numbers were much more severely reduced in the BM in IL-15–/– mice even compared with IRF-2–/– mice. Thus, IRF-2 played a unique role in keeping immature CD11blowDx5high NK cells alive, thus allowing their further maturation and the supply of mature NK cells to the periphery, whereas IL-15 was essential for early expansion of NK cells. Functional relationships between IRF-2 and GATA-3, another transcription factor required for mature NK cell generation (16), will be discussed.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
IRF-2–/– mice (17) backcrossed to C57BL/6 mice at least six times were described previously (18). Mice lacking 
2-microglobulin (2m–/– mice) and RAG-1 (RAG-1–/– mice) were purchased from The Jackson Laboratory and were provided by Dr. T. Nakayama (Chiba University, Chiba, Japan), respectively; they were backcrossed to C57/BL6 mice at least six times in addition and used to generate IRF-2–/–2m–/– and IRF-2–/–RAG-1–/– double-mutant mice. IRF-2–/–H-2d mice were generated by intercrossing IRF-2–/– mice with B10.D2 mice (SLC). IL-15–/– mice (13) were purchased from Taconic Farms. As we observed no significant difference between IRF-2+/– and IRF-2+/+ mice, mice with either of these two genotypes were used as controls. All mice were maintained in the animal facility in Shinshu University under specific pathogen-free conditions and used at 8–12 wk of age according to the institutional guideline for animal experimentation.
Abs and reagents
Fluorochrome- and biotin-conjugated mAbs and streptavidins used in this study were purchased from BD Pharmingen, except for FITC-anti-Ly5.2, PE-anti-Ly49I (clone YLI-90), anti-c-Kit, and anti-TCR from eBioscience. Biotin-labeled Abs were developed with PE-Cy7- or PE-Texas Red-streptavidin (eBioscience). Anti-CD16/32 Ab (FcBlock; BD Pharmingen) was used to block Fc-mediated binding of mAbs.
Flow cytometry
BM cells were flushed from the femurs, PBMC were isolated by centrifugation over Percoll (Amersham Biosciences), and liver mononuclear cells were prepared as previously described (19). Cells were stained and analyzed on the Cytomics FC500 flow cytometer (Beckman Coulter). Dead cells were regularly gated out using propidium iodide, and cells within the lymphocyte gate defined by forward and side scatters were analyzed. Data analysis was conducted with RXP software (Beckman Coulter).
Natural cytotoxicity and IFN-
production
Splenic and BM NK cells were enriched from control RAG-1–/– and IRF-2–/–RAG-1–/– double-mutant mice by depleting TCR+, CD19+, TER119+, and I-Ab+ cells using MACS (AutoMACS; Miltenyi Biotec), and natural cytotoxicity on YAC-1 cells was examined as previously described (20). Because the purity of NK cells was not always identical in the preparations from control RAG-1–/– and IRF-2–/–RAG-1–/– double-mutant mice due to the reduced numbers of NK cells in the latter, killer assays were conducted with the defined NK cells to target cell ratios. IFN- production by NK cells was examined on the flow cytometer by staining spleen or BM cells cultured for 16 h in the presence of recombinant human IL-2 (1000 U/ml; provided by Ajinomoto) and murine recombinant IL-12 (5 ng/ml; R&D Systems). Surface-stained cells for NK1.1, CD3, and CD11b were subjected to cytoplasmic staining for IFN- using a Cytofix/Cytoperm Plus kit (BD Pharmingen) according to the instructions supplied by the manufacturer.
Radiation BM chimeras
Radiation BM chimeras were established by reconstituting irradiated B6-Ly5.1 mice (purchased from Sankyo) with RBC-depleted BM cells prepared from IRF-2–/– or control mice bearing Ly5.2 marker as previously described (18) and were analyzed 8–10 wk later. In all chimeras, the percentages of donor-derived Ly5.1– cells within splenic and BM NK cells exceeded 95%.
In vivo proliferation and apoptosis
In vivo proliferation of NK cells was examined as previously described (6). Briefly, BrdU (Wako) solution in PBS was injected i.p. (3 mg/head), and 3 h later, BM cells were isolated and stained using a BrdU flow kit (BD Pharmingen) according to the instructions provided by the supplier. To detect cells undergoing apoptosis, NK1.1+CD3– cells were electrically gated and examined with PE-labeled annexin V using an Apoptosis Detection Kit (BD Pharmingen). Already dead cells positive for both 7-aminoactinomycin and annexin V were excluded from the analyses.
Allogenic BM graft rejection
Lethally irradiated (9.0 Gy) IRF-2–/– or control mice were injected i.v. with 3 x 106 RBC-depleted BM cells prepared from BALB/c mice. Eight days later, spleens were isolated and fixed in Bouin’s solution.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
As previously reported (12), IRF-2–/– mice exhibited a reduction of NK cells defined as NK1.1+CD3– cells in spleen and liver (Fig. 1, A and B). We examined the developmental stages of the residual NK cells in the periphery in IRF-2–/– mice according to the developmental framework proposed by Yokoyama et al. (2, 6), in which developmental intermediates of NK cells up-regulated the expression of Dx5 first and then that of CD11b, followed by the acquisition of CD43. As shown in Fig. 1, A and B, splenic and hepatic NK cell populations in control animals contained small numbers of immature NK cells (CD43–CD11blow) together with the prevailing numbers of fully mature NK cells (CD43+CD11bhighDx5high), an observation consistent with the report by Kim et al. (6). In contrast, the number of not only CD11bhighDx5high, but also that of CD11blowDx5high, NK cells was greatly reduced, and virtually no NK cells were positive for CD43 in spleen and liver in IRF-2–/– mice (Fig. 1, A and B). Notably, the absolute numbers of CD11blow immature NK cells in individual spleens in IRF-2–/– mice were not dramatically altered, whereas those of CD11bhigh mature NK cells were 100-fold reduced, on the average, compared with those in control animals (Fig. 1C). It was of note that more than half the splenic and hepatic NK cells in IRF-2–/– mice were Dx5low in contrast to control NK cells, >80% of which were Dx5high. These observations indicated that IRF-2 deficiency affected not only phenotypically mature NK cell production, but also immature NK cells qualitatively.
|
production of phenotypically immature NK cells in IRF-2–/– mice
Phenotypically immature NK cells enriched from the spleen (Fig. 2A) or BM (Fig. 2B) of IRF-2–/–RAG-1–/– double-mutant mice were similarly cytotoxic to NK-sensitive YAC-1 cells as those from RAG-1–/– mice. In contrast, combined stimulation with IL-2 and IL-12 in vitro induced IFN--producing NK cells less efficiently from splenocytes derived from IRF-2–/– mice (16.2%) than from control splenocytes (26.0%; Fig. 2C). Although the majority of IFN--producing NK cells in the control cultures were CD11bhigh, considerable fractions of CD11blow NK cells also producing IFN-, in agreement with a previous report (6). However, IFN- producers were less prominent in the CD11blow fraction in IRF-2–/– mice than in the corresponding fraction in control mice (13.5/89.7 vs 4.7/21.0; Fig. 2C). A similar reduction in IFN- production was observed in CD11blow NK cells in the BM of IRF-2–/– mice (Fig. 2D). Furthermore, rejection of allogenic BM grafts in IRF-2–/– mice was highly inefficient compared with that in controls, as revealed by the elevated numbers of CFUs in the spleen in the former (Fig. 2E). This observation was consistent with a previous description that rejection of the MHC-negative tumor RMA-S was strongly impaired in IRF-2–/– mice (12). Thus, the lack of IRF-2 affected the physiological functions of peripheral NK cells in vivo despite the normal natural cytotoxicity.
|
We observed that CD51 and c-Kit were not yet down-regulated on the majority of NK cells in the BM in IRF-2–/– mice (Fig. 3A), and these cells lacked the fully mature CD43+CD11bhigh cells, as found in our preliminary analyses (18) (Fig. 3A), indicative of the immature nature of these NK cells (2, 6). Quantitatively, the numbers of CD11blow immature NK cells were reduced in the BM of IRF-2–/– mice to only half the number in control mice (Fig. 3B; 5.9 vs 10.5 x 104 cells/femur), whereas those of CD43+CD11bhigh fully matured NK cells were 50-fold reduced in the absence of IRF-2 (Fig. 3B; 0.92 vs 43.8 x 104 cells/femur), indicating that generation from the committed progenitors and expansion of immature NK cells were not severely impaired, but further maturation was drastically impaired in IRF-2–/– mice. On the average, CD51+ and c-Kit+ cells constituted 62.6 and 71.3% of the CD11blow NK cells in BM NK cells in IRF-2–/– mice (n = 3), respectively, whereas CD11blow NK cells in control mice were only 30.9 and 35.5% positive for CD51 and c-Kit, respectively (n = 3). Thus, the absolute numbers of CD51+CD11blow and c-Kit+CD11blow NK cells were calculated to be 3.7 and 4.2 x 104 cells/femur, respectively, for IRF-2–/– mice and 3.2 and 3.7 x 104 cells/femur, respectively, for control mice. These values were not dramatically different between IRF-2–/– and control mice, suggesting that IRF-2 deficiency mainly affected the transition from the CD51/c-Kit+ to the CD51/c-Kit– stage within CD11blow NK cells. This was a situation contrasting with that in IL-15–/– mice, in which the total numbers of NK cells in the BM were 40- and 10-fold more severely reduced than in control and IRF-2–/– mice, respectively (Fig. 3B). Because practically all NK cells in the BM in IL-15–/– mice were CD11blow (S. Taki, unpublished observation), these results indicated that IL-15 was required for early expansion, whereas IRF-2 acted later for the maturation of NK cells. Importantly, the majority of CD11blow immature NK cells in the BM in IRF-2–/– mice were Dx5high (Fig. 3A) in contrast with those in the spleen and liver that stayed at the Dx5low stage (Fig. 1, A and B). Because Dx5 were shown to be one of the earliest markers acquired during NK cell development (6, 8), these results indicated unexpectedly that NK cell development advanced up to the CD11blowDx5high stage in the BM, but these cells did not efficiently contribute to the peripheral NK cell pool.
|
Similar differences in surface phenotypes between BM and splenic NK cells were observed in radiation BM chimeras established by reconstituting B6-Ly5.1 mice with BM cells from control and IRF-2–/– mice (control and IRF-2–/– chimeras, respectively; Fig. 4). Thus, Dx5low NK cells were dominant in the spleen in IRF-2–/– chimeras (Fig. 4A), whereas Dx5high NK cells were prevailing in the BM (Fig. 4B). In addition, the percentages of CD43+ and CD43–CD11bhigh NK cells derived from the donors were much lower in the spleen in IRF-2–/– chimeras than in control chimeras (Fig. 4A). Another series of chimeras generated with the converse combination of donors and recipients showed NK cell development indistinguishable from that in normal mice (S. Taki, unpublished observations), excluding the contribution of the BM environment to the NK cell phenotype of IRF-2–/– mice. These results indicated that IRF-2 deficiency within hemopoietic cells was responsible not only for the lack of mature NK cells in BM and spleen, but also for the less mature phenotypes of NK cells in spleen than in BM in IRF-2–/– mice.
|
We further characterized NK cells in IRF-2–/– mice by examining the expression of NKRs that were critical for physiological functions of NK cells (9, 21). As shown in Fig. 5A, Ly49 receptor expression by splenic NK cells was much less frequent in IRF-2–/– mice than in control mice. On the contrary, NK cells expressing NKG2A/C/E were more frequent in IRF-2–/– than in control mice. Another NKG2 molecule, NKG2D, was expressed by all NK cells regardless of IRF-2 deficiency (S. Nakajima, unpublished observation). Notably, in contrast, the frequencies of Ly49 receptors, except for Ly49I, on IRF-2-deficient NK cells in the BM were significantly higher than those in control mice, and Ly49I- and NKG2A/C/E-expressing NK cells were less frequent in the BM in IRF-2–/– mice than in control mice (Fig. 5A). Because the majority of NK cells in IRF-2–/– mice were CD11blow, we also compared Ly49 expression between control and IRF-2-deficient CD11blow NK cells (Fig. 5B). Even in these cases, the frequencies of Ly49+ NK cells in the spleen were affected in a manner similar to that in whole NK cells, albeit with a reduced severity, probably reflecting the immature nature of CD11blow splenic NK cells in IRF-2–/– mice, as indicated by the Dx5low phenotype (Fig. 1). These results indicated that both the lack of CD11bhigh NK cells and the immaturity of CD11blow NK cells contributed to the reduced Ly49+ NK cell frequencies in the spleen in IRF-2–/–mice. As previously shown, CD11blow immature NK cells expressed Ly49 receptors and NKG2A/C/E at lower and higher frequencies, respectively, compared with CD11bhigh mature NK cells (6), and Dx5high NK cells in the BM expressed Ly49 receptors at higher frequencies than Dx5low NK cells (8). Our current findings hence further strengthened the idea that splenic NK cells were less mature than those in the BM in IRF-2–/– mice. In contrast, in IL-15–/– mice, NK cells bearing Ly49 A, G2, and I were reduced substantially in the BM (Fig. 5B) as well as in the spleen (15) (S. Taki, unpublished observations). These phenotypes indicated that NK cells in the BM in IL-15–/– mice were less mature than those in IRF-2–/– mice.
|
The inefficient contribution of BM Ly49+ NK cells to the peripheral NK cell population in IRF-2–/– mice (Fig. 5A) raised the possibility that MHC class I-mediated self tolerance (7, 22) might select negatively Ly49+ cells. As summarized in Table I, the presence of H-2d molecules, ligands for Ly49A and G2 receptors (23), significantly reduced the frequencies of NK cells expressing these Ly49 receptors in the BM regardless of the presence of IRF-2. Consistent with the hypothesis that class I MHC-mediated self tolerance acted to reduce the frequencies of NK cells expressing multiple Ly49 receptors (24, 25, 26), the prevalence of Ly49A+G2+ double-positive (DP) cells was lower in mice expressing H-2d molecules than in 2m–/– mice regardless of whether IRF-2 was present (Table I). This decrease was not due merely to the reduced frequencies of each Ly49 receptor, because the deviation of the observed Ly49A+G2+ DP cell frequencies from the values expected assuming random distribution of each Ly49 receptor was lower in mice with the H-2d background than in 2m–/– mice (Table I). Thus, even in the absence of IRF-2, class I-dependent counterselection against the developing Ly49A+G2+ DP cells occurred normally in the BM.
|
2m–/– mice), although Ly49I was known to bind H-2d, albeit weakly (7, 23). These observations suggested that the selection against Ly49I+ NK cells had not been established yet in the BM, possibly because of the relatively late generation of Ly49I+ NK cells during NK cell development, in line with the models in which various Ly49 receptors were expressed sequentially but not simultaneously (4, 5). This idea was consistent with the observation that Ly49I+ NK cells in the BM were, in contrast to Ly49A+ or G2+ NK cells, less frequent in IRF-2–/– mice that lacked mature NK cells than in control mice regardless of the MHC genotype (Fig. 5A and Table 1). Activated phenotypes and accelerated apoptosis of IRF-2-deficient NK cells in vivo
We further characterized NK cells in BM in IRF-2–/– mice. It was found that BM NK cells in IRF-2–/– mice were apparently larger than those in control mice, as judged from the forward scatter profiles (Fig. 6A). In addition, the frequencies of CD11blow immature NK cells in the BM expressing CD69 were elevated in IRF-2–/– mice compared with those in control mice (Fig. 6B). These observations indicated that BM immature NK cells in IRF-2–/– mice were hyperactivated. However, this hyperactivated state of immature NK cells was not associated with accelerated proliferation, because the frequencies of NK cells acutely incorporating BrdU in BM in IRF-2–/– mice were comparable to those in control mice (Fig. 6, C and D). In contrast, we observed that NK cells undergoing apoptosis, as visualized with annexin V, were 2–3 times more abundant in IRF-2–/– mice than in controls (Fig. 6E). Collectively, these results suggested that CD11blowDx5high NK cells were generated in IRF-2–/– mice at an efficiency comparable to that in control mice, whereas they were lost at accelerated frequencies within the BM via apoptosis.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Remarkably, NK cells in BM were more mature than those in spleen and liver in IRF-2–/– mice, because the latter expressed less Dx5 than the former (Fig. 1, A and B, and Fig. 3A). The severer distortion of the NKR repertoire in spleen than in BM in IRF-2–/– mice (Fig. 5A) supported this view. We propose a model for the role of IRF-2 in the generation of peripheral NK cells (Fig. 7) in which IRF-2 is dispensable for generating relatively mature CD11blowDx5high NK cells that are expressing Ly49 receptors, whereas these cells cannot be efficiently exported to the periphery, because their premature apoptosis within the BM was accelerated in the absence of IRF-2 through a mechanism as yet to be established (see below). Selection of the Ly49 repertoire mediated by non-MHC and MHC factors seemed to proceed normally even in the BM in IRF-2–/– mice, at least for Ly49A and G2 (Table I), suggesting that the paucity of NK cells bearing Ly49 receptors (Fig. 5A) was not due to the aberrant NKR selection. Lowin-Kropf and Held (28) previously reported, in contrast, that forced overexpression of Ly49A had a positive impact on NK cell development. However, BM NK cells with Ly49A+ and G2+ cells of slightly elevated frequencies in IRF-2–/– mice (Fig. 5A) appeared to have no developmental advantage. Thus, the influence of IRF-2 deficiency was dominant over the positive effects, if any, of Ly49 expression.
|
Despite the distinct phenotypic immaturities, NK cells in spleen as well as BM in IRF-2–/– mice were normally cytotoxic to YAC-1 cells (Fig. 2, A and B). This rather unexpected observation was, however, not peculiar to IRF-2–/– mice. It was also observed previously that natural cytotoxicity was unaffected or was only slightly reduced despite the impaired Ly49 expression and/or maturation defects in NK cells in mutant mice lacking PU.1, Id2, T-bet, or GATA-3 (16, 27, 30, 31). Furthermore, perforin mRNA was detected even in very primitive Dx5low NK cells in BM (8), and very immature NK cells in IL-15–/– mice have recently been shown to be cytotoxic, albeit at a reduced level (32). These observations by us and others suggested that acquisition of natural cytotoxicity was completed at early stages (Dx5low) of development, in contrast with IFN- production, which continued to be elevated at the transition from the CD11blow to the CD11bhigh stage (Fig. 2, C and D), and hence was not sufficient to define fully mature NK cells. An interesting possibility in this regard is that the severely impaired rejection of allogenic BM grafts (Fig. 2E) and the strong defect in MHC-negative tumor rejection in vivo (12) in IRF-2–/– mice might reflect defects in the maturation-associated expression and/or down-regulation of functional molecules in addition to the reduction of NK cell numbers (Fig. 1C) and IFN- production (Fig. 2C). This idea, however, is merely a speculation at this time, and it is impossible to exclude an alternative possibility that cells other than NK cells, which might contribute to allogenic BM rejection, were also affected in IRF-2–/– mice.
NK cells in IL-15–/– mice were defective in expressing Ly49A, G2 and I not only in the spleen as recently shown by Kawamura et al. (15) but also in the BM as observed in this study (Fig. 5C). Such a coordinate impairment of Ly49 expression between splenic and BM NK cells contrasted sharply to the differential impairment in Ly49 expression observed between BM and splenic NK cells in IRF-2–/– mice (Fig. 5, A and B). Thus, defective NK cell maturation, instead of the loss of Ly49+ NK cells like that in IRF-2–/– mice, seemed to be responsible for the distorted Ly49 repertoire of splenic NK cells in IL-15–/– mice (15). In addition to the Ly49-inducing activity, IL-15 was obviously critical also for early expansion of NK cells in the BM (Fig. 3B). Because the majority of proliferating NK cells in vivo were CD43–Dx5+CD11blow (6) and the percentages of cells of this phenotype were higher in BM in IRF-2–/– mice than in control mice by a factor of 
1.5 (Fig. 3A), the unaltered percentages of proliferating NK cells (Fig. 6, C and D) indicated that NK cell proliferation was slightly less efficient in the absence of IRF-2. This partial impairment of proliferation might be due to the accelerated apoptosis (Fig. 6E) and could contribute in part to the severe loss of mature NK cells in the periphery. In contrast, we observed that the numbers of NK cells in the BM in IRF-2–/–mice were 10-fold more than those in IL-15–/– mice (Fig. 3B). We thus envisage that IL-15-mediated expansion takes place before IRF-2 acts (see Fig. 7) in a manner independent of IRF-2, and NK cell deficiency in IRF-2–/– mice is not due to the impaired IL-15 responses of immature NK cells, contrary to a previous hypothesis (12). The inefficient NK cell expansion from IRF-2-deficient BM cells in IL-15-supplemented cultures observed in the previous report (12) might be accounted for by accelerated apoptosis (Fig. 6E) and/or the low numbers of NK cells within BM cell preparations (Fig. 3B).
With respect to the regulation of cell survival, IRF-2 was previously shown to have an oncogenic potential as it transformed NIH-3T3 cells in vitro upon overexpression probably by repressing the proapoptotic actions of IRF-1 (33), another IRF family transcription factor shown to be required for NK cell development (34, 35). However, because IRF-1 acted in the BM microenvironment in a stem cell extrinsic manner during NK cell development (35), repression of IRF-1 actions by IRF-2 that was functioning within hemopoietic cells (12) (Fig. 4) appeared to be untenable in this case. IRF-2 was also implicated in the transcriptional activation of the cell cycle-regulated histone H4 gene, thereby potentially participating in cell cycle control (36). Thus, IRF-2 deficiency might result in cell cycle arrest, leading to the acceleration of apoptosis of immature NK cells independently of IRF-1. In relation to the antiapoptotic function of IRF-2 in developing NK cells, a recent report showed massive apoptosis of hepatic Kupffer cells occurring in IRF-2–/– mice treated with LPS in vivo (37).
It was of a great interest that the phenotypes of NK cells in IRF-2–/– mice closely resembled those in chimeric mice generated with GATA-3-deficient hemopoietic stem cells (16). Thus, in these chimeras, fully mature CD43+CD11bhigh NK cells were greatly reduced, and IFN- production by splenic NK cells was substantially lowered, whereas these cells exhibited virtually unaltered natural cytotoxicity. Moreover, NK cells expressing Ly49C/I and D were infrequent in these chimeras compared with those in controls. These strikingly common, albeit not completely overlapping, phenotypes suggested that GATA-3 and IRF-2 acted at similar stages of NK cell maturation. Nevertheless, the mechanisms of action of GATA-3 and IRF-2 did not seem to be identical, because the presence of Dx5high NK cells in spleen and the accumulation of phenotypically immature NK cells in BM were observed in GATA-3–/– chimeras (16), but not in IRF-2–/– mice (Figs. 1A and 3B). These dissimilar phenotypes suggested that GATA-3 was, unlike IRF-2, critical for the maturation, but not for the survival, of CD11blowDx5high immature NK cells. Furthermore, because we observed unaltered expression of GATA-3 mRNA in purified IRF-2-deficient NK cells (S. Nakajima, unpublished observation), it seemed unlikely that GATA-3 was a direct target of IRF-2. We speculate that IRF-2 keeps developing NK cells surviving for a certain period of time by repressing apoptosis, thereby allowing GATA-3 to induce maturation. Although additional studies are certainly required for directly testing this hypothesis, IRF-2 deficiency pointed to a previously unrecognized checkpoint critical for establishment of the normal peripheral NK cell population. The target gene(s) of IRF-2 relevant to NK cell maturation is currently unknown, as are those of many other transcription factors required for efficient NK cell generation (16, 30, 31, 38, 39). We believe that identification of the target genes for IRF-2 relevant in NK cell maturation should contribute to unveiling the complicated gene regulatory network underlying NK cell maturation.
|
Acknowledgments |
|---|
|
Disclosures |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Footnotes |
|---|
1 This work was supported in part by a Grant-in-Aid for Scientific Research on Priority Area (16043222) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan, and a Grant-in-Aid (15390159) from the Japan Science Promotion Society. 
2 Address correspondence and reprint requests to Dr. Shinsuke Taki, Department of Immunology and Infectious Diseases, Shinshu University Graduate School of Medicine, 3-1-1 Asahi, Matsumoto 390-8621, Japan. E-mail address: shin-t{at}sch.md.shinshu-u.ac.jp
3 Abbreviations used in this paper: BM, bone marrow; 2m, 2-microglobulin; DP, double positive; IRF, IFN regulatory factor; NKR, NK receptor.
Received for publication November 22, 2004. Accepted for publication March 1, 2005.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
-chain in NK cell development and Ly-49 induction. J. Immunol. 171: 5085-5090. production, and liver-specific homing of NK cells. Immunity 19: 701-711.[Medline]
2-microglobulin– and TAP-1– mice. J. Immunol. 159: 5219-5225.[Abstract]
14i NKT cells. Immunity 20: 477-494.[Medline]
-chain-dependent cytokines in the generation, differentiation, and maturation of NK cell precursors and peripheral NK cells in vivo. J. Immunol. 174: 1213-1221./+ (NK1+ T) cells, natural killer cells, and intestinal intraepithelial T cells. J. Exp. Med. 187: 967-972.This article has been cited by other articles:
|
|
 |
|
  S.-i. Ohno, T. Sato, K. Kohu, K. Takeda, K. Okumura, M. Satake, and S. Habu Runx proteins are involved in regulation of CD122, Ly49 family and IFN-{gamma} expression during NK cell differentiation Int. Immunol., January 1, 2008; 20(1): 71 - 79. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Arakura, S. Hida, E. Ichikawa, C. Yajima, S. Nakajima, T. Saida, and S. Taki Genetic Control Directed toward Spontaneous IFN-{alpha}/IFN-beta Responses and Downstream IFN-{gamma} Expression Influences the Pathogenesis of a Murine Psoriasis-Like Skin Disease J. Immunol., September 1, 2007; 179(5): 3249 - 3257. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Pascal, N. R. Nathan, E. Claudio, U. Siebenlist, and S. K. Anderson NF-{kappa}B p50/p65 Affects the Frequency of Ly49 Gene Expression by NK Cells J. Immunol., August 1, 2007; 179(3): 1751 - 1759. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Cuesta, Q. M. Nhu, E. Zudaire, S. Polumuri, F. Cuttitta, and S. N. Vogel IFN Regulatory Factor-2 Regulates Macrophage Apoptosis through a STAT1/3- and Caspase-1-Dependent Mechanism J. Immunol., March 15, 2007; 178(6): 3602 - 3611. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. O. Richards, X. Chang, B. W. Blaser, M. A. Caligiuri, P. Zheng, and Y. Liu Tumor growth impedes natural-killer-cell maturation in the bone marrow Blood, July 1, 2006; 108(1): 246 - 252. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||
) and IRF-2–/– (•) mice. Each symbol represents the value for an individual mouse. Horizontal bars are the means.
). Each symbol represents the value from an individual mouse. Horizontal bars are the means.
) and IRF-2–/– (
) mice were analyzed for the expression of the indicated NKRs. B, Only CD11blow NK1.1+CD3– cells were analyzed. Each column represents the mean and SD of the percentages of NK cells bearing a given NKR in four animals, except for the data for three IRF-2–/– mice in B. Statistically significant differences are marked: 
) mice were examined for the expression of the indicated Ly49 receptors. The means and SDs of three animals for each genotype are shown. Significant difference between control and IL-15–/– mice; p < 0.02 for Ly49A and p < 0.002 for the others.
