Advertisement

Slow receptor acquisition by NK cells regenerated in vivo from transplanted fetal liver or adult bone marrow stem cells

  • Motoi Maeda
    Affiliations
    aTerry Fox Laboratory, BC Cancer Agency, Vancouver, BC, Canada; Departments of bMedical Genetics and cPathology and Laboratory Medicine, University of British Columbia, Vancouver, Canada
    Search for articles by this author
  • Naoyuki Uchida
    Footnotes
    Affiliations
    aTerry Fox Laboratory, BC Cancer Agency, Vancouver, BC, Canada; Departments of bMedical Genetics and cPathology and Laboratory Medicine, University of British Columbia, Vancouver, Canada
    Search for articles by this author
  • Connie J Eaves
    Affiliations
    aTerry Fox Laboratory, BC Cancer Agency, Vancouver, BC, Canada; Departments of bMedical Genetics and cPathology and Laboratory Medicine, University of British Columbia, Vancouver, Canada
    Search for articles by this author
  • Fumio Takei
    Correspondence
    Offprint requests to: Fumio Takei, Ph.D., 601 West 10th Avenue, Vancouver, British Columbia, Canada V5Z 1L3
    Affiliations
    aTerry Fox Laboratory, BC Cancer Agency, Vancouver, BC, Canada; Departments of bMedical Genetics and cPathology and Laboratory Medicine, University of British Columbia, Vancouver, Canada
    Search for articles by this author
  • Author Footnotes
    ∗ The first two authors contributed equally to this work.Dr. Uchida is now at the First Department of Internal Medicine, Ehime University School of Medicine, Shigenobu, Ehime, Japan.

      Abstract

      Objective

      Adult mouse natural killer (NK) cells express two families of MHC class I–specific receptors, namely Ly49 and CD94/NKG2, whereas fetal and neonatal NK cells express only CD94/NKG2. After birth, Ly49+ NK cells slowly increase and CD94/NKG2+ NK cells decrease. The aim of this study was to determine whether murine NK cells develop differently from transplants of fetal liver and adult marrow stem cells and whether the adult marrow microenvironment is critical for NK receptor maturation.

      Materials and methods

      Enriched populations of stem cells were transplanted into adult mice, and the kinetics of NK receptor acquisition was examined. NK cells from osteopetrotic Csf1op/Csf1op mice, in which hematopoiesis within the marrow is severely limited, were also analyzed.

      Results

      NK cells regenerated from both fetal and adult stem cells initially resembled neonatal NK cells in their slow acquisition of Ly49 over several weeks, although the adult stem cell–derived NK cells matured approximately 10 days sooner. NK cells from adult Csf1op/Csf1op mice expressed normal levels of Ly49.

      Conclusion

      Maturation of the NK receptor repertoire is a slow process regardless of their stem cell origin or reduced marrow space caused by osteopetrosis.
      Adult mouse NK cells express both the Ly49 family and the CD94/NKG2 heterodimers [
      • Takei F
      • McQueen K.L
      • Maeda M
      • et al.
      Ly49 and CD94/NKG2: developmentally regulated expression and evolution.
      ,
      • Raulet D.H
      • Vance R.E
      • McMahon C.W
      Regulation of the natural killer cell receptor repertoire.
      ]. The former recognize specific classical MHC class I molecules whereas the latter bind the non-classical MHC class I Qa-1b. Co-expression of different members of these receptor families generates diverse receptor repertoires that enable the NK cell population as a whole to recognize subtle changes in MHC class I expression on target cells while maintaining self-tolerance. Unlike adult NK cells, most fetal and neonatal NK cells express CD94/NKG2 but not Ly49 [
      • Toomey J.A
      • Salcedo M
      • Cotterill L.A
      • et al.
      Stochastic acquisition of Qa1 receptors during the development of fetal NK cells in vitro accounts in part but not in whole for the ability of these cells to distinguish between class I-sufficient and class I- deficient targets.
      ]. Later, there is a gradual appearance of NK cells expressing Ly49 and a concomitant decline in NK cells expressing CD94/NKG2 by approximately 50% [
      • Dorfman J.R
      • Raulet D.H
      Acquisition of Ly49 receptor expression by developing natural killer cells.
      ,
      • Kubota A
      • Kubota S
      • Lohwasser S
      • Mager D.L
      • Takei F
      Diversity of NK cell receptor repertoire in adult and neonatal mice.
      ,
      • Salcedo M
      • Colucci F
      • Dyson P.J
      • et al.
      Role of Qa-1(b)-binding receptors in the specificity of developing NK cells.
      ]. Adult bone marrow (BM)–derived NK cell progenitors can differentiate in vitro into NK cells expressing CD94/NKG2 but not Ly49 [
      • Williams N.S
      • Moore T.A
      • Schatzle J.D
      • et al.
      Generation of lytic natural killer 1.1+, Ly-49- cells from multipotential murine bone marrow progenitors in a stroma-free culture: definition of cytokine requirements and developmental intermediates.
      ]. Further differentiation into mature NK cells expressing both CD94/NKG2 and Ly49 requires the presence of OP-9 stromal cells [
      • Williams N.S
      • Klem J
      • Puzanov I.J
      • Sivakumar P.V
      • Bennett M
      • Kumar V
      Differentiation of NK1.1+, Ly49+ NK cells from flt3+ multipotent marrow progenitor cells.
      ]. The acquisition of Ly49 by NK cells developing in vitro is stochastic but seems to follow a pre-determined program [
      • Williams N.S
      • Kubota A
      • Bennett M
      • Kumar V
      • Takei F
      Clonal analysis of NK cell development from bone marrow progenitors in vitro: orderly acquisition of receptor gene expression.
      ,
      • Roth C
      • Carlyle J.R
      • Takizawa H
      • Raulet D.H
      Clonal acquisition of inhibitory Ly49 receptors on developing NK cells is successively restricted and regulated by stromal class I MHC.
      ]. On the other hand, when embryonic stem cells are induced to differentiate into NK cells in vitro, they express CD94/NKG2 but not Ly49, even when cultured in the presence of OP-9 cells [
      • Lian R.H
      • Maeda M
      • Lohwasser S
      • et al.
      Orderly and nonstochastic acquisition of CD94/NKG2 receptors by developing NK cells derived from embryonic stem cells in vitro.
      ]. Therefore, fetal and adult hematopoietic stem cells (HSCs) may be pre-programmed to differentiate into fetal and adult NK cells, respectively, as has also been described for B-lineage, T-lineage, and erythroid differentiation programs [
      • Geiger H
      • Sick S
      • Bonifer C
      • Muller A.M
      Globin gene expression is reprogrammed in chimeras generated by injecting adult hematopoietic stem cells into mouse blastocysts.
      ,
      • Kincade P.W
      • Owen J.J
      • Igarashi H
      • Kouro T
      • Yokota T
      • Rossi M.I
      Nature or nurture? Steady-state lymphocyte formation in adults does not recapitulate ontogeny.
      ]. To test this possibility, we investigated receptor expression on NK cells in adult mice transplanted with HSCs isolated from fetal liver (FL) or adult BM. We also examined NK cells from adult Csf1op/Csf1op mice in which differentiation within the BM space is severely reduced by the osteopetrosis characteristic of their BM [
      • Begg S.K
      • Radley J.M
      • Pollard J.W
      • Chisholm O.T
      • Stanley E.R
      • Bertoncello I
      Delayed hematopoietic development in osteopetrotic (op/op) mice.
      ]. Our results show that the appearance of NK cells with a mature receptor phenotype is slow regardless of the source of HSCs but is not impeded when the marrow is highly osteopetrotic.

      Materials and methods

      Mice

      Osteopetrotic Csf1op/Csf1op and littermate control (+/+ or +/Csf1op) mice were obtained directly from the Jackson Laboratory (Bar Harbor, ME, USA) and fed soft chow. C57Bl/6J (Ly5.2) and C57Bl/6J-Pep 3b (Ly5.1), also originally from the Jackson Laboratory, were bred and maintained in our animal facility as described [
      • Miller C.L
      • Rebel V.I
      • Lemieux M.E
      • Helgason C.D
      • Lansdorp P.M
      • Eaves C.J
      Studies of W mutant mice provide evidence for alternate mechanisms capable of activating hematopoietic stem cells.
      ].

      Cell lines, monoclonal antibodies, and flow cytometry

      RMA and RMA-S have been described [
      • Lian R.H
      • Freeman J.D
      • Mager D.L
      • Takei F
      Role of conserved glycosylation site unique to murine class I MHC in recognition by Ly-49 NK cell receptor.
      ]. The monoclonal antibodies (mAbs) YE1/48 (anti-Ly49A), 5E6 (anti-Ly49C/I), 4D11 (anti-Ly49G2), anti-FcRγ (2.4G2), anti-Ly5.1 (A20-1.7), FITC-conjugated anti-Ly5.2 (AL14A2), biotin-conjugated anti- Mac1 (M1/70), anti-Gr1 (RB6-8C5), anti-B220 (RA3-6B2), and anti-Ly1 (53-7.3) have been described [
      • Lian R.H
      • Maeda M
      • Lohwasser S
      • et al.
      Orderly and nonstochastic acquisition of CD94/NKG2 receptors by developing NK cells derived from embryonic stem cells in vitro.
      ,
      • Miller C.L
      • Rebel V.I
      • Helgason C.D
      • Lansdorp P.M
      • Eaves C.J
      Impaired steel factor responsiveness differentially affects the detection and long-term maintenance of fetal liver hematopoietic stem cells in vivo.
      ]. All other antibodies and fluorochrome-conjugated streptavidin were purchased from BD Biosciences (Mississauga, Ontario, Canada). Flow cytometry for cell sorting and analysis was done as described [
      • Lian R.H
      • Maeda M
      • Lohwasser S
      • et al.
      Orderly and nonstochastic acquisition of CD94/NKG2 receptors by developing NK cells derived from embryonic stem cells in vitro.
      ].

      Transplantation

      Single cell suspensions of E14 liver and 16–20-week-old BM cells were prepared as described [
      • Uchida N
      • Leung F.Y
      • Eaves C.J
      Liver and marrow of adult mdr-1a/1b(-/-) mice show normal generation, function, and multi-tissue trafficking of primitive hematopoietic cells.
      ]. For FL HSC enrichment, cells were stained with a cocktail of mAbs to Gr1, B220, Ly1, Ter119, interleukin (IL)-7 receptor α-chain, IL-2 receptor beta-chain, and NK1.1, and unstained lineage-negative (lin) cells were isolated by cell sorting. For adult mouse BM cells, anti-Mac1 was also added to the mAb cocktail. Sorted lin FL cells (0.93–1.98×106 cells/mouse) or lin BM cells (1.08–1.43×106 cells/mouse) were transplanted intravenously into irradiated (900 cGy) Pep3b mice as described [
      • Uchida N
      • Leung F.Y
      • Eaves C.J
      Liver and marrow of adult mdr-1a/1b(-/-) mice show normal generation, function, and multi-tissue trafficking of primitive hematopoietic cells.
      ].

      Analysis of NK cells

      Donor-derived NK cells (Ly5.2+ NK1.1+ CD3) were analyzed for CD94/NKG2 and Ly49 expression by flow cytometry. For cytotoxicity assays, splenocytes were cultured with IL-2 as described [
      • Lian R.H
      • Freeman J.D
      • Mager D.L
      • Takei F
      Role of conserved glycosylation site unique to murine class I MHC in recognition by Ly-49 NK cell receptor.
      ] and recipient NK cells were removed by panning with anti-Ly5.1 mAbs. Donor-derived NK cells thus isolated were cultured with IL-2 for an additional 2 days and used for cytotoxicity assays as described [
      • Lian R.H
      • Freeman J.D
      • Mager D.L
      • Takei F
      Role of conserved glycosylation site unique to murine class I MHC in recognition by Ly-49 NK cell receptor.
      ].

      Statistical analyses

      Data are presented as mean±SD. The paired Student's t-test was used to compare frequencies of NK receptors on NK cells after transplantation. A p value less than 0.05 was considered statistically significant in two-tailed Student's t-tests of compared datasets.

      Results

      When lin FL and adult BM cells from B6 mice were injected into irradiated congenic Pep3b mice, donor-derived NK cells were generated with similar kinetics (Fig. 1A). The earliest point when sufficient donor-derived NK cells could be recovered for analysis of receptor expression was 10 days after transplantation. At that time, most of the FL-derived NK cells expressed CD94/NKG2 but not Ly49. Over the subsequent 40 days, the frequency and hence absolute number of CD94/NKG2+ NK cells declined and Ly49+ NK cells became more prevalent (Fig. 1B). NK cells derived from adult BM HSCs showed more accelerated changes in NK receptor expression than those generated from FL HSCs (Fig. 1B). By day 10, for example, the frequency of Ly49G2+ cells (40±10%, n = 6) in recipients of adult BM HSCs was significantly higher (p<0.05) than that in recipients of FL HSCs (6.4±1.5%, n = 6) and was already similar to the frequency of Ly49G2+ cells in adult NK populations. The frequencies of Ly49A+ and Ly49C/I+ NK cells were also higher in recipients of BM HSCs than in recipients of FL HSCs (3.9±1.2% vs. 0.8±0.2% for Ly49A and 12.0±1.5% vs. 5.5±1.7% for Ly49C/I, n = 6, p<0.05). Conversely, the frequency of CD94/NKG2+ cells was lower in recipients of BM HSCs as compared to recipients of FL HSCs (66±4% vs. 80±1%, n = 6, p<0.05). The percentages of Ly49C/I+ and Ly49A+ NK cells on day 10 were significantly lower than in the normal adult and, only after another 40 days, did a population of NK cells with the normal adult receptor repertoire become apparent. In general, the maturation of NK cells derived from BM HSCs was faster than that of FL HSC-derived NK cells by about 10 days. Both FL and adult BM HSC-derived NK cells killed MHC class I negative RMA-S cells but not MHC class I positive RMA cells, indicating that they were functional and able to recognize MHC class I on target cells (Fig. 1C). These results show that FL HSCs differentiate into adult NK cells in adult recipients. On the other hand, even from adult BM HSCs, the generation of the full NK cell receptor repertoire takes several weeks.
      Figure thumbnail gr1
      Figure 1Analysis of developing donor-derived NK cells (Ly5.2+ NK1.1+ CD3−) from FL or adult BM HSCs after transplantation. (A) Sequential analysis of reconstituted donor-derived NK cells from FL (93,000 cells/mouse) or adult BM (108,000 cells/mouse) HSCs after transplantation. Total number of donor-derived cells per spleen (▴,×106 cells), total number of donor-derived NK cells per spleen (■,×104 cells), and the proportion of donor-derived NK cells against whole donor-derived cells in spleen (♦, %) are shown. Mean and SD values of three recipients at each time point are shown. The results were reproducible in two independent experiments. (B) Sequential analysis of NK receptor repertoires (♦, Ly49A; □, Ly49C/I; ▴, Ly49G2; ◊, NKG2 on developing donor-derived NK cells from FL or adult BM HSCs after transplantation. The expression of CD94 (not shown) on the donor-derived NK cells from both FL and adult BM HSCs at every time point was identical to that of NKG2. Mean values±SD of six recipients from two independent experiments at each time point are shown. For comparison with normal adult NK cells, mean values and SD of each NK cell receptor on NK cells from three 20 week- old C57/BL6J mice are shown in the center. On day 10, the percentages of BM-derived NK cells expressing Ly49 were higher than those of FL-derived cells for all Ly49 receptors tested, whereas the percentage of CD94/NKG2+ cells was lower for BM-derived NK cells than FL-derived cells. Two-tailed t-test indicated that all the differences are statistically significant (p<0.05). On the subsequent days, only the differences in Ly49G expression on day 20 and Ly49C/I on day 50 were statistically significant (). (C) Functional maturation of donor-derived NK-LAK cells at 20 days after transplantation. The purity of donor-derived NK-LAK cells after negative panning using the recipient's type anti-Ly5.1 mAbs used in these cytotoxicity assay series was 93±4%. The specific cytotoxicity of donor-derived NK cells (▴, from FL HSCs; •, from adult BM HSCs) against RMA (MHC-class I positive C57/BL6 murine T lymphoma cell line) and RMA-S (MHC-class I negative cell line derived from RMA due to TAP2 mutation) at various effector/target ratios is shown. Results shown are representative of six independent experiments.
      Since the site of hematopoiesis shifts from the FL to the BM around birth, we asked whether the BM microenvironment is critical for the maturation of NK cells with an adult phenotype. Accordingly, we examined NK cells from adult osteopetrotic Csf1op/csf1op mice, which have a defective BM compartment due to an osteoclast deficiency. Murine NK cells do not acquire an adult NK receptor repertoire until ~5 weeks after birth [
      • Dorfman J.R
      • Raulet D.H
      Acquisition of Ly49 receptor expression by developing natural killer cells.
      ,
      • Salcedo M
      • Colucci F
      • Dyson P.J
      • et al.
      Role of Qa-1(b)-binding receptors in the specificity of developing NK cells.
      ], and the BM starts to develop in Csf1op/Csf1op mice after the mice are 5 weeks old [
      • Begg S.K
      • Radley J.M
      • Pollard J.W
      • Chisholm O.T
      • Stanley E.R
      • Bertoncello I
      Delayed hematopoietic development in osteopetrotic (op/op) mice.
      ]. Therefore, we analyzed 5-week-old Csf1op/Csf1op mice in which NK cells are thought to be generated, at least in part, in extramedullary sites [
      • Begg S.K
      • Radley J.M
      • Pollard J.W
      • Chisholm O.T
      • Stanley E.R
      • Bertoncello I
      Delayed hematopoietic development in osteopetrotic (op/op) mice.
      ]. Csf1op/Csf1op mice are significantly smaller than either wild-type or heterozygous control mice of the same age and are found to have sixfold fewer hematopoietic (CD45+) cells and 12-fold fewer NK cells in their BM at 5 weeks of age (n = 4) as compared to controls (n = 4). The number of NK cells present in the spleen was also significantly lower (p<0.05) in the Csf1op/Csf1op mice as compared to controls (4.7±0.5×105 cells/spleen, n = 4 vs 29±8×105 cells/spleen, n = 4). However, the expression of NK receptors on NK cells from the spleen of Csf1op/Csf1op mice was not affected by the osteopetrotic state of their BM (Table 1).
      Table 1.Expression of Ly49 and CD94/NKG2 on NK cells from osteopetrotic Csf1op/Csf1op mice
      Control mice (n = 4)Csf1op/Csf1op mice (n = 4)
      Total splenocytes (× 107 cells)8.8±0.913.2±0.21
      Total splenic NK cells (× 105 cells)29±84.7±0.5
      Total BM cells (× 106 cells)
      Only CD45+ hematopoietic cells were counted.
      22±43.4±1.1
      Total BM NK cells (× 104 cells)39±133.2±1.1
      YE1/48 positive cells (%)
      Due to allelic differences between B6 mice and Csf1op/Csf1op mice in the NK gene complex, Ly49 molecules in Csf1op/Csf1op mice recognized by the YE1/48 (anti-Ly49AB6) and 4D11 (anti-Ly49GB6) mAbs have not been identified. The 5E6 (anti-Ly49C/IB6) mAb did not bind to NK cells from Csf1op/Csf1op or wild-type mice. Five week-old mice were analyzed.
      5.9±0.56.3±1.2
      4D11 positive cells (%)
      Due to allelic differences between B6 mice and Csf1op/Csf1op mice in the NK gene complex, Ly49 molecules in Csf1op/Csf1op mice recognized by the YE1/48 (anti-Ly49AB6) and 4D11 (anti-Ly49GB6) mAbs have not been identified. The 5E6 (anti-Ly49C/IB6) mAb did not bind to NK cells from Csf1op/Csf1op or wild-type mice. Five week-old mice were analyzed.
      24±0.621±1.3
      CD94 positive cells (%)67±2.869±1.3
      NKG2 positive cells (%)69±2.472±1.4
      Only CD45+ hematopoietic cells were counted.
      Due to allelic differences between B6 mice and Csf1op/Csf1op mice in the NK gene complex, Ly49 molecules in Csf1op/Csf1op mice recognized by the YE1/48 (anti-Ly49AB6) and 4D11 (anti-Ly49GB6) mAbs have not been identified. The 5E6 (anti-Ly49C/IB6) mAb did not bind to NK cells from Csf1op/Csf1op or wild-type mice. Five week-old mice were analyzed.

      Discussion

      Following transplantation of HSCs from FL or adult BM into adult mice, donor-derived NK cells developing in the recipients initially resemble neonatal NK cells expressing CD94/NKG2 but not Ly49. Subsequently, Ly49G+ NK cells appear first, followed by Ly49C/I+, and finally Ly49A+ cells, while the frequency of CD94/NKG2+ NK cells declines. This pattern of changes in the NK receptor phenotype is highly reproducible and is very similar to that seen in neonatal mice [
      • Dorfman J.R
      • Raulet D.H
      Acquisition of Ly49 receptor expression by developing natural killer cells.
      ,
      • Salcedo M
      • Colucci F
      • Dyson P.J
      • et al.
      Role of Qa-1(b)-binding receptors in the specificity of developing NK cells.
      ]. Kim et al. classified NK cell development in vivo into five stages and suggested that CD94/NKG2 expression precedes Ly49 in the normal steady-state development of NK cells in adult mice [
      • Kim S
      • Iizuka K
      • Kang H.S
      • et al.
      In vivo developmental stages in murine natural killer cell maturation.
      ]. Our results are consistent with this view. However, it takes more than 7 weeks for donor-derived NK cells to acquire the receptor repertoire of normal adult NK cells even when adult BM HSCs are transplanted. Our findings that FL HSCs injected into irradiated adult mice differentiate into adult NK cells suggested that the microenvironment in which NK cells develop might dictate the way NK cells mature. However, even when adult HSCs were transplanted into adult mice, the acquisition of the adult NK receptor repertoire took several weeks with very reproducible kinetics, favoring the concept that NK cell development is pre-programmed.
      Our analysis of NK cells from Csf1op/Csf1op mice shows that severe osteopetrosis in these mice does not affect the expression of Ly49. It should be noted that Csf1op/Csf1op mice do have some BM tissue, albeit very much reduced as compared to their wild-type littermates, and it is difficult to determine what proportions of splenic NK cells in Csf1op/Csf1op mice are generated in the BM vs extramedullary sites. Nevertheless, our results suggest that the presence of an osteopetrotic BM microenvironment does not preclude the acquisition of Ly49 receptors by NK cells. It is interesting to note that our results are different from previous studies with estrogen-induced osteopetrotic mice, which had low numbers of Ly49+ NK1.1+ cells [
      • Puzanov I.J
      • Bennett M
      • Kumar V
      IL-15 can substitute for the marrow microenvironment in the differentiation of natural killer cells.
      ]. It is possible that the effects of estrogen on Ly49 expression may be unrelated to osteopetrosis.
      Interestingly, clinical human allogenic BM transplant recipients also showed a similarly slow reconstitution of the normal receptor repertoire [
      • Shilling H.G
      • Young N
      • Guethlein L.A
      • et al.
      Genetic control of human NK cell repertoire.
      ,
      • Shilling H.G
      • McQueen K.L
      • Cheng N.W
      • Shizuru J.A
      • Negrin R.S
      • Parham P
      Reconstitution of NK cell receptor repertoire following HLA-matched hematopoietic cell transplantation.
      ]. On initial engraftment, very few of the NK cells produced expressed the killer cell Ig-like receptor (KIR), while the majority were CD94/NKG2A+. During the subsequent 6–9 months, there was a gradual appearance of NK cells expressing KIR and a corresponding decline in NK cells expressing CD94/NKG2A in the majority of patients. However, in patients who exhibited a depressed frequency of KIR-expressing NK cells and a concomitantly high frequency of CD94/NKG2A cells for the first year after transplantation, the frequency of KIR-expressing NK cells did not reach donor levels until after 3 years. It was proposed that the rate of KIR acquisition may be influenced by the conditioning regimen of the patients, including radiation and chemotherapy. Our experiments suggest that slow changes in receptor expression may not be due to radiation-damaged stroma. The transplantation of B6 FL HSCs into c-kit-mutant (W41/W41) mice [
      • Miller C.L
      • Rebel V.I
      • Lemieux M.E
      • Helgason C.D
      • Lansdorp P.M
      • Eaves C.J
      Studies of W mutant mice provide evidence for alternate mechanisms capable of activating hematopoietic stem cells.
      ] that received only 400 cGy radiation resulted in similar kinetics of change in NK receptor expression (data not shown). Although human and murine NK cells express very different receptors (KIR and Ly49, respectively) for the recognition of MHC class I, they share similar patterns of receptor acquisition as they develop from transplanted HSCs. The reason for the slow receptor acquisition is still unknown.

      Acknowledgements

      This work was supported by the National Cancer Institute of Canada. N.U. was a recipient of a postdoctoral fellowship from the Canadian Institutes of Health Research. We thank Kristin J. Lyons for technical assistance.

      References

        • Takei F
        • McQueen K.L
        • Maeda M
        • et al.
        Ly49 and CD94/NKG2: developmentally regulated expression and evolution.
        Immunol Rev. 2001; 181: 90-103
        • Raulet D.H
        • Vance R.E
        • McMahon C.W
        Regulation of the natural killer cell receptor repertoire.
        Annu Rev Immunol. 2001; 19: 291-330
        • Toomey J.A
        • Salcedo M
        • Cotterill L.A
        • et al.
        Stochastic acquisition of Qa1 receptors during the development of fetal NK cells in vitro accounts in part but not in whole for the ability of these cells to distinguish between class I-sufficient and class I- deficient targets.
        J Immunol. 1999; 163: 3176-3184
        • Dorfman J.R
        • Raulet D.H
        Acquisition of Ly49 receptor expression by developing natural killer cells.
        J Exp Med. 1998; 187: 609-618
        • Kubota A
        • Kubota S
        • Lohwasser S
        • Mager D.L
        • Takei F
        Diversity of NK cell receptor repertoire in adult and neonatal mice.
        J Immunol. 1999; 163: 212-216
        • Salcedo M
        • Colucci F
        • Dyson P.J
        • et al.
        Role of Qa-1(b)-binding receptors in the specificity of developing NK cells.
        Eur J Immunol. 2000; 30: 1094-1101
        • Williams N.S
        • Moore T.A
        • Schatzle J.D
        • et al.
        Generation of lytic natural killer 1.1+, Ly-49- cells from multipotential murine bone marrow progenitors in a stroma-free culture: definition of cytokine requirements and developmental intermediates.
        J Exp Med. 1997; 186: 1609-1614
        • Williams N.S
        • Klem J
        • Puzanov I.J
        • Sivakumar P.V
        • Bennett M
        • Kumar V
        Differentiation of NK1.1+, Ly49+ NK cells from flt3+ multipotent marrow progenitor cells.
        J Immunol. 1999; 163: 2648-2656
        • Williams N.S
        • Kubota A
        • Bennett M
        • Kumar V
        • Takei F
        Clonal analysis of NK cell development from bone marrow progenitors in vitro: orderly acquisition of receptor gene expression.
        Eur J Immunol. 2000; 30: 2074-2082
        • Roth C
        • Carlyle J.R
        • Takizawa H
        • Raulet D.H
        Clonal acquisition of inhibitory Ly49 receptors on developing NK cells is successively restricted and regulated by stromal class I MHC.
        Immunity. 2000; 13: 143-153
        • Lian R.H
        • Maeda M
        • Lohwasser S
        • et al.
        Orderly and nonstochastic acquisition of CD94/NKG2 receptors by developing NK cells derived from embryonic stem cells in vitro.
        J Immunol. 2002; 168: 4980-4987
        • Geiger H
        • Sick S
        • Bonifer C
        • Muller A.M
        Globin gene expression is reprogrammed in chimeras generated by injecting adult hematopoietic stem cells into mouse blastocysts.
        Cell. 1998; 93: 1055-1065
        • Kincade P.W
        • Owen J.J
        • Igarashi H
        • Kouro T
        • Yokota T
        • Rossi M.I
        Nature or nurture? Steady-state lymphocyte formation in adults does not recapitulate ontogeny.
        Immunol Rev. 2002; 187: 116-125
        • Begg S.K
        • Radley J.M
        • Pollard J.W
        • Chisholm O.T
        • Stanley E.R
        • Bertoncello I
        Delayed hematopoietic development in osteopetrotic (op/op) mice.
        J Exp Med. 1993; 177: 237-242
        • Miller C.L
        • Rebel V.I
        • Lemieux M.E
        • Helgason C.D
        • Lansdorp P.M
        • Eaves C.J
        Studies of W mutant mice provide evidence for alternate mechanisms capable of activating hematopoietic stem cells.
        Exp Hematol. 1996; 24: 185-194
        • Lian R.H
        • Freeman J.D
        • Mager D.L
        • Takei F
        Role of conserved glycosylation site unique to murine class I MHC in recognition by Ly-49 NK cell receptor.
        J Immunol. 1998; 161: 2301-2306
        • Miller C.L
        • Rebel V.I
        • Helgason C.D
        • Lansdorp P.M
        • Eaves C.J
        Impaired steel factor responsiveness differentially affects the detection and long-term maintenance of fetal liver hematopoietic stem cells in vivo.
        Blood. 1997; 89: 1214-1223
        • Uchida N
        • Leung F.Y
        • Eaves C.J
        Liver and marrow of adult mdr-1a/1b(-/-) mice show normal generation, function, and multi-tissue trafficking of primitive hematopoietic cells.
        Exp Hematol. 2002; 30: 862-869
        • Puzanov I.J
        • Bennett M
        • Kumar V
        IL-15 can substitute for the marrow microenvironment in the differentiation of natural killer cells.
        J Immunol. 1996; 157: 4282-4285
        • Kim S
        • Iizuka K
        • Kang H.S
        • et al.
        In vivo developmental stages in murine natural killer cell maturation.
        Nat Immunol. 2002; 3: 523-528
        • Shilling H.G
        • Young N
        • Guethlein L.A
        • et al.
        Genetic control of human NK cell repertoire.
        J Immunol. 2002; 169: 239-247
        • Shilling H.G
        • McQueen K.L
        • Cheng N.W
        • Shizuru J.A
        • Negrin R.S
        • Parham P
        Reconstitution of NK cell receptor repertoire following HLA-matched hematopoietic cell transplantation.
        Blood. 2003; 101: 3730-3740