This information is current as
of June 14, 2017.
Cre Recombinase-Mediated Inactivation of
H-2D d Transgene Expression: Evidence for
Partial Missing Self-Recognition by Ly49A
NK Cells
Vassilios Ioannidis, Jacques Zimmer, Friedrich Beermann
and Werner Held
J Immunol 2001; 167:6256-6262; ;
doi: 10.4049/jimmunol.167.11.6256
http://www.jimmunol.org/content/167/11/6256
Subscription
Permissions
Email Alerts
This article cites 36 articles, 22 of which you can access for free at:
http://www.jimmunol.org/content/167/11/6256.full#ref-list-1
Information about subscribing to The Journal of Immunology is online at:
http://jimmunol.org/subscription
Submit copyright permission requests at:
http://www.aai.org/About/Publications/JI/copyright.html
Receive free email-alerts when new articles cite this article. Sign up at:
http://jimmunol.org/alerts
The Journal of Immunology is published twice each month by
The American Association of Immunologists, Inc.,
1451 Rockville Pike, Suite 650, Rockville, MD 20852
Copyright © 2001 by The American Association of
Immunologists All rights reserved.
Print ISSN: 0022-1767 Online ISSN: 1550-6606.
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
References
Cre Recombinase-Mediated Inactivation of H-2Dd Transgene
Expression: Evidence for Partial Missing Self-Recognition by
Ly49A NK Cells1
Vassilios Ioannidis,* Jacques Zimmer,* Friedrich Beermann,† and Werner Held2*
E
xpression of MHC class I molecules protects cells from
NK cell-mediated killing. This is regulated via the interaction of incompletely defined receptors that activate NK
cell effector function and inhibitory NK cell receptors specific for
MHC class I molecules. The engagement of MHC class I receptors
generates inhibitory signals that abort the induction of the NK
cell’s lytic program. In the absence of MHC class I molecules on
target cells NK cells fail to receive inhibitory signals, which allows
target cell lysis to proceed.
Murine NK cells recognize classical MHC class I molecules via
Ly49 family receptors. In general these receptors are specific for
one and occasionally two class I isotypes, sometimes discriminating alleles thereof (1). The prototype receptor, Ly49A, inhibits NK
cell function upon the interaction with the H-2Dd but not the Db
class I molecule (2). The specificity of Ly49 receptors together
with their combinatorial distribution allows NK cells to react to
subtle changes in a cell’s MHC class I make-up, such as the absence of a single class I allele (3).
The MHC class I environment in which NK cells mature or exist
is an important determinant of their reactivity to target cells. The
requirements for NK cell adaptation to class I were addressed using two experimental systems: MHC-deficient and MHC-different
mouse models. First, although NK cells arise in the absence of
*Ludwig Institute for Cancer Research, University of Lausanne, Epalinges, Switzerland; and †Swiss Institute for Experimental Cancer Research, Epalinges, Switzerland
Received for publication June 28, 2001. Accepted for publication October 1, 2001.
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.
1
This work was supported in part by a grant to W.H. from the Swiss National Science
Foundation. W.H. is the recipient of a Swiss Talents for Academic Research and
Teaching fellowship from the Swiss National Science Foundation.
2
Address correspondence and reprint requests to Dr. Werner Held, Ludwig Institute
for Cancer Research, 155 Ch. des Boveresses, 1066 Epalinges, Switzerland. E-mail
address: [email protected]
Copyright © 2001 by The American Association of Immunologists
MHC class I molecules in ␤2-microglobulin (␤2m)3-deficient
mice, they were found to be impaired in their ability to lyse various
target cells (4 – 6). In radiation bone marrow chimeras, the absence
of class I molecules from either the radioresistant cells or some or
all of the radiosensitive (hematopoietic) cells was sufficient to render NK cells hyporesponsive (4, 5, 7). In addition, MHC differences between mice were used to investigate NK cell function and
tolerance. NK cells that developed in H-2b mice expressing an
H-2Dd transgene (Tg) acquired the ability to reject bone marrow
grafts from H-2b donors. Nontransgenic (non-Tg) and Tg mice
were self-tolerant, as the respective syngeneic bone marrow grafts
were not rejected (8). Thus, NK cells can adapt to the presence of
MHC class I molecules. They acquire tolerance toward MHCidentical cells and reactivity to cells that lack them.
Due to the generalized effect on NK cell function in the absence
of MHC class I, MHC-deficient mice were not suited to address the
function of NK cell subsets that are defined by the expression of
specific receptors for MHC class I. Indeed, the rejection of MHCdifferent but not MHC-deficient bone marrow cells could be prevented by the enforced expression of an appropriate inhibitory
MHC receptor on all NK cells (9 –11). Further, it was shown that
NK cell self-tolerance was ensured via the inhibitory effect of
MHC receptors (12–14); e.g., the inhibitory Ly49A receptor prevented H-2d NK cells from attacking syngeneic cells. In contrast,
self-tolerance of H-2b Ly49A⫹ NK cells was independent of
Ly49A function (12, 13). Because Ly49A⫹ NK cells from H-2b
mice killed ␤2m-deficient target cells it was concluded that selftolerance was ensured by the inhibition through H-2b-specific receptors (13). In agreement with this thesis, in a panel of human NK
cell clones, each clone expressed at least one inhibitory receptor
specific for self-MHC (15). These results suggest that autoaggressive clones are not present in the NK cell compartment. However,
they did not rule out the existence of additional tolerance mechanisms. Autoaggressive NK cells may be refractory to cloning.
Abbreviations used in this paper: ␤2m, ␤2-microglobulin; Tg, transgene/transgenic;
loxDd, H-2Dd gene flanked with loxP sites; bc, backcross; ER, estrogen receptor;
CD3-Cy, CD3-CyChrome.
3
0022-1767/01/$02.00
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
We have established H-2Dd-transgenic (Tg) mice, in which H-2Dd expression can be extinguished by Cre recombinase-mediated
deletion of an essential portion of the transgene (Tg). NK cells adapted to the expression of the H-2Dd Tg in H-2b mice and acquired
reactivity to cells lacking H-2Dd, both in vivo and in vitro. H-2Dd-Tg mice crossed to mice harboring an Mx-Cre Tg resulted in
mosaic H-2Dd expression. That abrogated NK cell reactivity to cells lacking Dd. In Dd single Tg mice it is the Ly49Aⴙ NK cell
subset that reacts to cells lacking Dd, because the inhibitory Ly49A receptor is no longer engaged by its Dd ligand. In contrast,
Ly49Aⴙ NK cells from Dd ⴛ MxCre double Tg mice were unable to react to Dd-negative cells. These Ly49Aⴙ NK cells retained
reactivity to target cells that were completely devoid of MHC class I molecules, suggesting that they were not anergic. Variegated
Dd expression thus impacts specifically missing Dd but not globally missing class I reactivity by Ly49Aⴙ NK cells. We propose that
the absence of Dd from some host cells results in the acquisition of only partial missing self-reactivity. The Journal of Immunology, 2001, 167: 6256 – 6262.
The Journal of Immunology
6257
Consistent with this possibility, murine NK cell clones lacking
known self-MHC-specific inhibitory receptors were detected using
a single cell PCR assay (16). Evidence for the existence of potentially autoaggressive NK cells was also obtained using a class I
MHC mosaic mouse, where H-2b cells coexisted with H-2b cells
expressing H-2Dd (17). NK cells in these mice were unable to react
to H-2b⫹Dd⫺ cells in vitro and in vivo. However, upon the removal of H-2b⫹Dd⫺ cells in vitro the reactivity of H-2b⫹Dd⫹ NK
cells to H-2b⫹Dd⫺ target cells was observed. Based on these findings it was proposed that NK cell tolerance may also be ensured by
NK cell silencing (17, 18).
To further address the function of NK cells in relation to variable patterns of MHC class I expression we have engineered a
genetic “off switch” for an H-2Dd Tg using Cre recombinase. The
shut down of H-2Dd expression in a fraction of cells prevented NK
cell reactivity to H-2b⫹Dd⫺ cells. This was associated with the
inability of the Ly49A⫹ NK cell subset to perform missing Dd
recognition. However, these Ly49A⫹ NK cells were able to react
to target cells, which were completely lacking MHC class I molecules, suggesting that they were not silent but rather specifically
unable to react to cells lacking Dd. These findings are discussed in
the context of the adaptation of NK cells to self-MHC class I.
For mouse typing, whole blood was stained with FITC-labeled anti-H2
mAbs: H-2Kd (SF1.1.1.1), Dd (34.2.12), and Db (KH95). A PE-labeled
mAb (PK136) to the NK1.1 Ag was used to confirm the B6 origin of the
NK gene complex (all mAbs were obtained from BD PharMingen, San
Diego, CA). Erythrocytes were removed with FACS lysis solution according to the manufacturer’s recommendation (BD Biosciences, San
Jose, CA).
Spleen, lymph node, and bone marrow cell suspensions (usually 106
cells) were incubated with 24G2 (anti-CD16/32) hybridoma supernatant to
reduce background. Cells were further stained with a mixture of NK1.1-PE,
CD3-CyChrome (CD3-Cy), and biotinylated anti-H-2 mAbs. Alternatively, NK1.1-PE and CD3-Cy were used in conjunction with appropriate
combinations of FITC-labeled and biotinylated anti-Ly49 mAbs: Ly49A
(JR9-318) kindly provided by J. Roland (Institut Pasteur, Paris, France),
Ly49C/I (SW5E6), Ly49G2 (4D11), and Ly49D (4E5). All mAbs were
purchased from BD PharMingen, except JR9-318, which was labeled in
this lab. Biotinylated mAbs were revealed using streptavidin-allophycocyanin (Molecular Probes, Eugene, OR). In general 5 ⫻ 104–105 cells were
analyzed on a FACSCalibur flow cytometer using CellQuest software (BD
Biosciences).
Materials and Methods
Cytotoxicity assays
An oligo containing two loxP sites (GenBank accession no. M10494, nucleotides 14 – 47), an internal SpeI site, and external XbaI sites was cloned
into the XbaI site of pSK II (Stratagene, La Jolla, CA). The 1.8-kb XbaI
fragment from pDd1 (19) was then inserted into the loxP-flanked SpeI site,
which destroyed both SpeI and XbaI sites. The loxP-1.8-kb Dd1-loxP fragment was removed from the vector (using the external XbaI sites) and
reinserted into pDd1. The 8.1-kb loxDd construct was liberated from vector
sequence using EcoRI digestion and injected into fertilized (C57BL/6 ⫻
DBA/2)F2 oocytes. Six Tg founder mice termed loxDd Tg were identified
by PCR using the following primers: P1, 5⬘-CTGCCCACGCCCACT
GTCT-3⬘; and P2, 5⬘-GACCATTCACCACTGTAGG-3⬘. This resulted in
a 283-bp fragment for the loxDd Tg and 243 bp for wild-type H-2Db or
H-2Dd alleles. Tg lines were established from five founders by backcrossing to C57BL/6 (B6) (H-2b) mice. For the experiments shown in this work
Dd-Tg mice (line 28) B6 backcross (bc) 2– 4 were used.
Mx-1-Cre (20) B6 bc 1 and CMV-Cre estrogen receptor (ER) (21) Tg
mice B6 bc 3 were kindly provided by M. Aguet (Swiss Institute for Experimental Cancer Research, Epalinges, Switzerland) and P. Chambon (Institut de Génétique et de Biologie Moléculaire et Cellulaire, Strasbourg,
France), respectively. LoxDd ⫻ Cre double Tg mice were obtained by
breeding. Double Tg offspring were identified by flow cytometry for H-2Dd
and by Southern-blot of tail DNA probed with a 700-bp MluI fragment
derived from Cre cDNA (kindly provided by M. Aguet, Swiss Institute for
Experimental Cancer Research).
B6 (H-2b) and B10.D2 (H-2d) mice were obtained from Harlan Olac
(Zeist, The Netherlands). ␤2m-deficient mice were purchased from The
Jackson Laboratory (Bar Harbor, ME). Ly49A-Tg mice (B6 bc 8) were
described before (9). All mice were ⬎6 wk of age when used for experiments. They were homozygous for a B6-derived NK gene complex and had
a defined MHC based on flow cytometric analysis using appropriate Abs.
In all experiments appropriate littermate mice were used as controls.
Polymerase chain reaction
Recombination of the loxDd Tg was analyzed using a semiquantitative
PCR approach. Starting with ⬃1 ␮g, serial (10-fold) dilutions of lymph
node or tail DNA were used as a template for PCR with the following
primers: P2, 5⬘-GACCATTCACCACTGTAGG-3⬘; and P3, 5⬘-TGGGAA
GAAACAGGAGGAG-3⬘. This results in a 420-bp amplification product
for recombined loxDd. In nonrecombined loxDd and non-Tg mice the two
primers are separated by ⬃2.2 kb and no PCR product is detected.
Thy-1-specific PCR was performed as a control: Thy-1, 5⬘ (775–795)
CCATCCAGCATGAGTTCAGCC; and Thy-1, 3⬘ (1432–1451) GCATC
CAGGATGTGTTCTGA. This results in a 676-bp amplification product.
PCR was performed using 1 ␮M of each primer and 200 nM dNTPs.
After denaturing for 3⬘ at 92°C PCR was performed at 92°C for 1 min,
55°C for 1 min, and 72°C for 1 min for 40 cycles. Comparable data were
Flow cytometry
NK cell-mediated lysis of lymphoblast targets was done as described in
Ref. 22. Briefly, erythrocyte-depleted spleen cells were passed over a nylon
wool column. Nonadherent cells were cultured in complete DMEM plus
500 ng/ml recombinant human IL-2 (a gift from Glaxo Wellcome, Geneva,
Switzerland). At day 3 adherent and nonadherent cells were harvested and
used as effectors. T cell blasts were obtained by culturing spleen cells (2 ⫻
106 cells/ml) in DMEM plus 2.5 ␮g/ml Con A (Sigma-Aldrich, Buchs,
Switzerland). After 2 days dead cells were removed by Ficoll density gradient centrifugation (Pharmacia Biotech, Uppsala, Sweden). Cells (3 ⫻
106) were labeled with 50 ␮Ci of 51Cr for 1 h at 37°C. After three washes,
5 ⫻ 103 labeled target cells were mixed with effector cells in duplicate at
various E:T ratios in 96-well U-bottom plates. Supernatants were harvested
after 4 h and 51Cr release into the supernatant was determined using a
gamma counter. The percentage of NK1.1⫹CD3⫺ cells in the effector cell
population was estimated using flow cytometry.
Detection of intracellular IFN-␥
Total splenocytes were cultured (10 ml at 2 ⫻ 106 cells/ml) in the presence
of human IL-2 as above. After 4 days, nonadherent cells were discarded
and adherent cells (containing normally 30 –50% NK cells) were detached
using PBS/EDTA. After washing, 1.5 ⫻ 106 cells were plated in a 24-well
culture plate in a final volume of 2 ml in the presence of IL-2. Stimulator
cells (2 ⫻ 106) were added: C1498 (H-2b, thymoma), Dd-transfected
C1498 (kindly provided by W. Seaman, University of California, San Francisco, CA). Brefeldin A (10 ␮g/ml; Sigma-Aldrich) was added 3 h after the
onset of cultures. After overnight culture, the cells were harvested, washed,
and surface-labeled as described above using biotinylated anti-NK1.1 followed by a mixture of CD3-Cy, Ly49A-FITC, and streptavidin-allophycocyanin to reveal the biotinylated mAb. After two washes in PBS 5%
FCS, cells were fixed and permeabilized according to the supplier’s instructions (BD PharMingen). The cells were then stained intracellularly
with the PE-conjugated anti-IFN-␥ (XMG1.2) or an isotype-matched control mAb (BD PharMingen). After washing, the samples were analyzed
using four-color flow cytometry as above.
Bone marrow graft rejection
Recipient mice were lethally irradiated (960 rad from a 137Cs source) and
reconstituted 24 h later by i.v. inoculation of 106 (H-2b) or 5 ⫻ 106
(␤2m⫺/⫺) bone marrow cells. To deplete NK cells some recipient mice
received 100 ␮g of purified mAb PK136 i.p. 24 h before irradiation. Five
days after irradiation the mice were injected with 3 ␮Ci of [125I]UdR (Amersham, Dübendorf, Switzerland) i.p. Twenty-four hours later the incorporation of radioactivity into the spleen was measured using a gamma
counter.
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
Transgene construction and transgenic mice
obtained when a fixed amount of DNA template was analyzed using variable numbers (25, 30, or 35) of PCR cycles. PCR products were run on
agarose gel containing ethidium bromide and visualized under UV.
6258
ACQUISITION OF MISSING SELF-REACTIVITY
Results
Modification of the H-2Dd gene
An 8-kb EcoRI fragment containing the entire H-2Dd gene (19)
was modified by introducing two loxP sites into noncoding regions. The DNA sequence, which is flanked by two loxP sites, is
deleted upon the action of the Cre recombinase. Because the two
loxP sites in the H-2Dd gene flank the promoter region and the
exons that encode the ␣1 and ␣2 domains, de novo expression of
H-2Dd will be prevented following Cre-mediated recombination
(Fig. 1).
Generation and characterization of Dd-Tg mice
NK cells from Dd-Tg mice react to the absence of H-2Dd
Short-term (3– 4 days) IL-2-activated NK cells from Dd Tg (H2b⫹Dd⫹) mice killed H-2b lymphoblast targets. They did not kill
syngeneic (H-2b⫹Dd⫹) lymphoblast targets, indicating self-tolerance. Moreover, Tg and non-Tg NK cells killed MHC class I-deficient (due to a disruption of the ␤2m gene) target cells equally
well (Fig. 3A). These results suggest that Dd-Tg NK cells have
adapted to the presence of the class I Tg and react to its absence in
agreement with missing self-recognition. Thus, the introduction of
two loxP sites into the H-2Dd Tg did not alter the expected effects
of Dd Tg expression on NK cell reactivity.
Generation and analysis of Dd ⫻ Cre double Tg mice
The presence of two loxP sites in the Dd Tg offered the opportunity
to manipulate H-2Dd expression in vivo using Cre-Tg mice. Dd-Tg
FIGURE 1. Schematic representation and modification of the H-2Dd
gene. Exons (filled boxes) and introns (open boxes) of an 8-kb EcoRI (R)
fragment containing the H-2Dd gene are shown schematically. LoxP sites
(filled triangles), introduced into two XbaI sites (X), flank the promoter
region and exons encoding the leader peptide (L) as well as the ␣1 and ␣2
domains of the H-2Dd protein. Cre recombinase action deletes the intervening sequence, which will prevent H-2Dd expression. PCR primers P1
and P2 are used to identify mice harboring the modified H-2Dd Tg based
on a longer amplification product (283 bp) compared with a wild-type
H-2Dd or Db alleles (243 bp). Primers P3 and P2 are used to detect recombination of the floxed Dd Tg. Recombination results in a 420-bp PCR
product while the two primers are separated by ⬃2.2 kb when the Dd Tg
is not recombined.
FIGURE 2. H-2Dd expression in Dd-Tg mice. A, Histograms show
lymph node cells from the indicated strains of mice stained with Abs specific for H-2Dd (thick line) and control H-2Kd (thin line). Numbers indicate
the percentage of H-2Dd-positive cells in the respective strains. H-2Dd cell
surface levels are compared between B10. D2 (H-2d) and Dd-Tg mice,
using the mean fluorescence intensity (MFI) of H-2Dd staining with mAb
34.2.12. B, Lymph node cells from the indicated Dd ⫻ Cre double Tg mice
were stained as above. Numbers indicate the percentage of H-2Dd-high
cells in the respective strains. C, upper panel, Density plots show CD3 vs
NK1.1 expression among total splenocytes of the indicated types of mice.
Numbers indicate the percentage of CD3⫺NK1.1⫹ cells among total spleen
cells. The lower panel illustrates Ly49A vs Ly49C/I usage among gated
CD3⫺NK1.1⫹ NK cells of the above mice. Numbers indicate the percentage of cells in the respective quadrants among total NK cells.
mice were thus crossed with Mx-Cre-Tg mice. The Mx promoter
allows the induction of Cre expression via an IFN-␣␤-inducible
regulatory element (20). Second, Dd-Tg mice were crossed with
CMV-CreER-Tg mice, which constitutively express a fusion protein of Cre recombinase with the ligand binding domain of the
human ER driven by the CMV promoter (21). The activity of the
fusion protein can be induced upon the administration of the synthetic ligand 4-hydroxytamoxifen.
Without deliberate induction of Cre expression or of its activity
we have assessed H-2Dd expression in double Tg mice using flow
cytometry. Approximately one-third of PBL or lymph node cells in
young adult Dd ⫻ Mx-Cre double Tg mice were completely
H-2Dd-negative (Fig. 2B). A similar fraction of cells expressed
H-2Dd at the same level as Dd single Tg mice while the remaining
cells expressed intermediate H-2Dd levels. These latter cells may
harbor a partial recombination of the multicopy Tg locus. No or
very little H-2Dd loss (⬍1% of cells) had occurred in lymphoid
cells of Dd ⫻ CMV-CreER mice.
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
The loxP-modified H-2Dd gene was injected into fertilized (B6 ⫻
DBA/2) F2 oocytes to generate Tg mice. Five Tg lines were initially established by backcrossing to B6 mice. Offspring of H-2b
haplotype and with a NK gene complex of B6 origin were selected
as detailed in Materials and Methods. The different lines expressed
H-2Dd on lymphoid cells at levels slightly below to twice as high
as compared with B10.D2 (H-2d) mice (data not shown). One Tg
line (28) termed hereafter Dd Tg, that expressed H-2Dd at levels
similar (⬇80%) to B10.D2 (H-2d) mice, was analyzed in more
detail (Fig. 2A).
The Journal of Immunology
Recombination of the loxDd Tg was confirmed at the level of
genomic DNA. Lymph node cell-derived DNA was subjected to
PCR amplification using primers specific for the H-2Dd gene,
which are normally separated by ⬃2.2 kb of intervening sequence
(p2–p3, as indicated in Fig. 1). Following recombination of the
H-2Dd gene these primers were expected to yield an amplification
product of 420 bp. Indeed, a PCR product specific for a recombined Dd Tg was exclusively observed in lymphoid tissue of Dd ⫻
Mx-Cre double Tg but not of Dd single Tg or Dd ⫻ CMV-CreER
double Tg mice (Fig. 4A), consistent with the cytometric analysis.
This assay allowed us to determine whether Dd recombination has
also occurred in nonlymphoid tissues. Indeed, Dd recombination
was observed in tail DNA of both Dd ⫻ Mx-Cre and Dd ⫻ CMVCreER mice (Fig. 4B). The two Dd ⫻ Cre double Tg mice thus
displayed distinct patterns of H-2Dd expression. Mx-Cre led to
H-2Dd loss on some but not all lymphocytes (hematopoietic origin) and in tail (nonhematopoietic origin). In contrast, lymphocytes in Dd ⫻ CMV-CreER mice were homogeneously H-2Dd
positive while Dd recombination was evident in tail. These findings
indicated considerable leakiness of the two Cre induction systems,
yet they clearly established the functionality of the H-2Dd off
switch.
Lack of missing Dd recognition by NK cells from Dd ⫻ Cre
double Tg mice in vivo
The partial loss of H-2Dd expression had no significant effects on
the abundance of NK cells or the representation of several MHC
class I-specific Ly49 receptors (Fig. 2C and data not shown). The
impact of the H-2Dd expression patterns on NK cell reactivity was
tested using bone marrow graft rejection experiments. Lethally irradiated Dd ⫻ Cre double Tg mice were challenged with bone
FIGURE 4. Recombination of the Dd Tg in Dd ⫻ Cre double Tg mice.
Recombination of the Dd Tg in different tissues was assessed using a semiquantitative PCR approach. Genomic DNA was extracted from lymph node
cells (A) or tail (B) and serially (10-fold) diluted before amplification with
H-2D- or Thy-1-specific primers. Primers specific for the H-2D gene are
normally separated by ⬃2.2 kb of intervening sequence (P2–P3; see Fig.
1). Upon recombination and deletion of intervening sequence, P2 is separated
from P3 by ⬃420 bp. The control Thy-1 amplification product is 676 bp.
marrow cells from H-2b (Dd-negative) mice. The proliferation of
bone marrow stem cells (indicating acceptance of the graft and
thus the absence of a NK cell reaction against it) was read out by
[125I]UdR incorporation into spleens of recipient mice. As shown
in Fig. 5, Dd (H-2bDd) single Tg mice rejected H-2b bone marrow
FIGURE 5. Bone marrow graft rejection by Dd-Tg mice. The indicated
strains of mice were lethally irradiated before the infusion of 1 ⫻ 106 H-2b
(B6) (left) or 5 ⫻ 106 MHC class I-deficient (␤2m⫺/⫺) bone marrow cells
(right). Some B10.D2 mice were depleted of NK1.1⫹ cells using mAb
PK136. After 5 days the mice were injected with 3 ␮Ci of [125I]UdR i.p.
Twenty-four hours later the incorporation of radioactivity into the spleen
was measured. Open circles represent mean values (⫾ SD) from three or
more mice. Values for individual mice are shown when fewer recipients
were used. High incorporation of label indicates acceptance of the graft and
thus the absence of a NK cell-mediated rejection. Low values are indicative
of graft rejection.
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
FIGURE 3. Reactivity of NK cells from Dd-Tg mice. Nylon wool nonadherent spleen cells were cultured for 3 days in the presence of IL-2. The
cytotoxic activity of these preparations was tested in standard 51Cr release
assays using Con A-activated lymphoblasts derived from ␤2m-deficient
(open squares), H-2b (B6) (open circles), or H-2bDd (closed circles) spleen
cells. A, Effector cells were derived from Dd-Tg (H-2bDd) or B6 (H-2b)
mice. B, Effector cells were from Dd ⫻ Mx-Cre or Dd ⫻ CMV-CreER
double Tg mice. The percentages of NK cells in the respective effector cell
cultures were consistently 10 –20% of total cells. Graphs show means
(⫾ SD) compiled from three to seven independent experiments. For ␤2mdeficient targets only the mean is shown.
6259
6260
ACQUISITION OF MISSING SELF-REACTIVITY
FIGURE 6. Reactivity of NK cells from Ly49A ⫻ Dd ⫻ Mx-Cre triple
Tg mice. Nylon wool nonadherent spleen cells from the indicated mouse
strains were cultured for 3 days in the presence of IL-2. The cytotoxic
activity of these preparations was tested against Con A-activated target
cells derived from H-2b (B6) (䡬) and H-2bDd (䢇) or ␤2m-deficient spleen
cells (□). Graphs show mean percentage of specific lysis (⫾ SD) compiled
from four values obtained in two independent experiments. For ␤2m-deficient targets only the mean is shown.
FIGURE 7. Reactivity of Ly49A NK cells. Splenocytes from B6 mice
were cultured for 4 days in the presence of IL-2. Plastic adherent cells were
harvested and exposed overnight to the indicated tumor cell lines before
cell surface and intracellular staining. Dot plots show intracellular IFN-␥⫹
vs Ly49A among gated (NK1.1⫹CD3⫺) NK cells. Numbers indicate the
percentages of NK cells that produce IFN-␥ among the Ly49A⫹ or
Ly49A⫺ NK cell subsets.
NK cells from Dd ⫻ CMV-CreER mice acquire missing Dd
recognition in vitro
We next determined whether H-2Dd recombination in double Tg
mice influenced NK cell reactivity in vitro. Short-term (3 days)
IL-2-activated NK cells from Dd ⫻ Mx-Cre mice were unable to
kill H-2b lymphoblasts (Fig. 3B). MHC class I-deficient targets
were readily killed, consistent with the bone marrow rejection experiments. However, in contrast to the in vivo data, NK cells from
Dd ⫻ CMV-CreER-Tg mice killed H-2b target cells as efficiently as
NK cells from Dd single Tg mice (Fig. 3B). The in vitro culture may
thus enable missing Dd reactivity perhaps due to an IL-2-dependent
reactivation of silenced NK cells, as suggested before (23). In addition, the separation from nonlymphoid cells, some of which lack
H-2Dd, may be required to restore reactivity in vitro. Indeed, missing
Dd reactivity by NK cells from mosaic mice was restored upon the
removal of hematopoietic cells lacking H-2Dd (17).
Ly49A⫹ NK cells from Dd ⫻ Mx-Cre mice perform missing
class I but not missing Dd recognition
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
Missing Dd recognition by NK cells from H-2bDd mice is mediated
by the Ly49A⫹ NK cell subset (12). This suggested that in Dd ⫻
Mx-Cre-Tg mice, the inability to perform missing Dd reactions
was due to a malfunction of the Ly49A⫹ NK cell subset. This
further raised the issue of whether these Ly49A⫹ cells were completely nonreactive (anergic) or whether they were selectively nonreactive to cells lacking Dd. To address these questions we crossed
Dd ⫻ Mx-Cre mice to mice expressing a Tg Ly49A on all their NK
cells. Like Dd ⫻ Mx-Cre double Tg (Fig. 3), NK cells from
Ly49A ⫻ Dd ⫻ Mx-Cre triple Tg mice showed minimal cytotoxic
activity to H-2b lymphoblasts (Fig. 6). However, NK cells from
triple Tg mice efficiently killed ␤2m-deficient lymphoblasts or
YAC-1 target cells (Fig. 6 and data notshown). Ly49A⫹ NK cells
in Dd ⫻ Mx-Cre mice are thus not anergic but rather selectively
impaired in their reactivity to cells lacking Dd.
The selective failure of Ly49A⫹ NK cells to react to missing Dd
may be related to a failure of the Ly49A receptor to inhibit NK
cells in a Dd mosaic situation. To address this issue we have exposed NK cells to tumor cells and determined the production of
IFN-␥ at the single cell level using intracellular flow cytometry
(Fig. 7) (24). A significant fraction of Ly49A⫹ NK cells from B6
mice (25.6 ⫾ 6.0%) produced IFN-␥ following overnight exposure
to the syngeneic (H-2b) tumor cell line C1498 (Table I). Even
though values were variable between experiments, significantly
more Ly49A⫹ NK cells (36.7 ⫾ 7.3%, p ⬍ 0.02) from Dd-Tg mice
produced IFN-␥ in response to C1498 (Table I). This assay thus
appropriately indicates a contribution of missing Dd reactivity to
IFN-␥ production by Ly49A⫹ NK cells from H-2bDd mice.
Ly49A⫹ NK cells from Dd ⫻ Mx-Cre mice behaved essentially
like those from B6; i.e., a similarly small fraction (26.0 ⫾ 9.6%)
of Ly49A⫹ NK cells produced IFN-␥ (Table I). These data are
compatible with the above lysis assays and suggest that Ly49A⫹
NK cells from Mx-Cre mice behave similarly to those from B6
mice: they do not react to missing Dd, yet are able to produce
IFN-␥ in response to C1498 target cells. The exposure to Dd-transfected C1498 cells reduced the percentage of IFN-␥-expressing
Ly49A Dd ⫻ Mx-Cre NK cells ⬃4-fold as compared with C1498
cells (Table I). The capacity of Ly49A to function as an inhibitory
receptor is thus intact in Dd ⫻ Mx-Cre mice. It is interesting to
note that while Ly49A reduced the fraction of IFN-␥-expressing
cells in response to C1498 Dd targets, the mean fluorescence intensity of IFN-␥ staining in the residual Ly49A⫹IFN-␥⫹ NK cells
was not significantly lowered (data not shown). Therefore, this
single cell assay suggests that the engagement of the inhibitory
Ly49A receptor affects NK cell effector function in an all-or-none
fashion. IFN-␥ production by Ly49A⫹ NK cells was independent
from the coexpression of the Dd-specific activating Ly49D receptor
(data not shown), which is able to initiate IFN-␥ production (25).
The responsiveness to C1498 and C1498 Dd cells was not significantly different among Ly49A-negative NK cells or NK cells
expressing Ly49C/I (Table I), which are inhibitory receptors specific for H-2b MHC class I molecules (1). Variegated Dd expression thus impacts specifically the NK cell subset expressing the
Dd-specific inhibitory Ly49A receptor and prevents selectively
missing Dd reactivity. However, the failure to acquire missing Dd
reactivity is not related to an inability of Ly49A to function as an
inhibitory receptor.
grafts judged by the low incorporation of [125I]UdR into the recipients’ spleens. This was not due to the inability of the grafted
cells to expand, as a high [125I]UdR incorporation was observed in
B6 (H-2b) recipient mice. In contrast to Dd single Tg mice, both
types of Dd ⫻ Cre double Tg mice were unable to reject the H-2b
graft (Fig. 5). The extent of H-2Dd loss in the two types of double
Tg mice was thus sufficient to prevent missing Dd recognition. The
nonreactive state was specific for H-2b grafts, as all mice rejected
MHC class I-deficient bone marrow grafts. While the data recapitulate those obtained in another model of H-2Dd mosaic mice
(17), they further establish that H-2Dd loss from some nonlymphoid cells was sufficient to abolish missing Dd recognition.
The Journal of Immunology
6261
Table I. Production of IFN-␥ by NK cell subsets
% intracellular IFN-␥⫹ Cells
Mouse Strain
B6
Dd Tg
Dd Mx-Cre
B6
Dd Tg
Dd Mx-Cre
B6
Dd Tg
Dd Mx-Cre
NK Cell
Subset
Ly49A⫹
Ly49 A⫺
Ly49C/I⫹
No.
C1498
C1498 Dd
2.9 ⫾ 1.8
4.6 ⫾ 2.7
4.4 ⫾ 1.8
3.4 ⫾ 2.4
4.4 ⫾ 2.1
4.4 ⫾ 1.7
4.1 ⫾ 2.3
3.9 ⫾ 2.5
4.2 ⫾ 1.8
25.6 ⫾ 6.0
36.7 ⫾ 7.3b
26.0 ⫾ 9.6
24.2 ⫾ 4.8
22.5 ⫾ 6.9
20.8 ⫾ 6.9
19.3 ⫾ 6.4
17.4 ⫾ 7.6
16.4 ⫾ 7.3
4.0 ⫾ 3.1c
13.3 ⫾ 8.9c
6.2 ⫾ 3.9c
19.5 ⫾ 6.7
18.6 ⫾ 8.3
15.5 ⫾ 6.8
16.7 ⫾ 6.4
14.6 ⫾ 8.2
12.3 ⫾ 6.0
a
Data are derived from six to eight independent experiments.
Significantly different from B6 ( p ⬍ 0.02) or B6 Dd Mx ( p ⬍ 0.05).
Significantly different ( p ⬍ 0.01) from corresponding value obtained with
C1498. Data evaluation is based on the two-tailed Student t test.
b
c
We describe in this report the construction of a molecular switch
to irreversibly extinguish a Tg MHC class I molecule. The Tg is
ubiquitously expressed and its down-regulation reflects present or
past expression of Cre recombinase. This mouse strain provides a
novel tool to establish various patterns of Dd expression using
available Cre Tgs that are under the control of different cis-acting
regulatory elements. Importantly, these patterns can be established
using a Tg founder line with a proven NK cell phenotype. Phenotypes in Dd ⫻ Cre Tg mice can thus be accurately compared with
those of founder mice. This is of importance in light of our observation that not all Dd-Tg founder mice were equally capable of
rejecting H-2Dd-negative bone marrow grafts (our unpublished observation). Differences in the capacity to reject grafts may be related to the site of Tg insertion. Indeed, several of the founder mice
initially displayed a mosaic Dd expression, raising the possibility
that mosaic Dd expression, which affects NK cell reactivity, persisted in some tissues of some Tg lines.
These novel mouse strains allowed us to study the function of
NK cells, which develop and exist in the presence of two types of
host cells. Some cells express and others lack the strong Ly49A
ligand H-2Dd, whereas all host cells express H-2b class I molecules. The use of two distinct Cre Tgs (Mx- and CMV-Cre) resulted in Dd loss in different compartments. The reactivity of NK
cells arising in these situations was compared with NK cells that
develop in an environment where all or no cells express H-2Dd,
respectively. Our data revealed functional differences among NK
cells from the different types of mice. Both Dd mosaic mouse
strains failed to reject H-2b bone marrow grafts, similar to mice
completely lacking Dd. In contrast to NK cells from Dd ⫻ CMVCreER mice, those from Mx-Cre mice were further unable to kill
H-2b lymphoblasts. This prompted us to specifically investigate the
NK cell subset expressing the relevant inhibitory MHC receptor
Ly49A. We found that Ly49A⫹ NK cells in mosaic mice were
indeed unable to perform missing Dd recognition; however, these
cells were not completely unreactive, as MHC class I-deficient
targets were readily killed.
Two distinct models may account for our data as well as previous findings (17). In the first one, the absence of Dd from some
host cells may have a negative effect and silence NK cell clones
that would normally perform missing Dd recognition. The concept
of NK cell silencing is primarily supported by the observation that
missing Dd reactivity is restored when Dd-positive NK cells are
cultured separately from Dd-negative hematopoietic (17) or non-
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
Discussion
hematopoietic host cells (Fig. 3). However, because these Ly49A⫹
NK cells perform missing class I recognition and are thus not
silent, we propose an alternative interpretation of our observations.
The encounter with Dd-deficient host cells during NK cell development may not allow complete adaptation to self-MHC class I.
Whereas these NK cells acquire missing H-2b recognition, they do
not acquire proper missing Dd reactivity. This scenario is in line
with a number of recent observations. Mature NK cells are inhibited in a cumulative fashion when NK cells express multiple selfspecific Ly49 receptors (26, 27). This suggests that self-specific
inhibitory MHC receptors function independently, each contributing to NK cell inhibition. The acquisition of missing self-reactivity
during NK cell development may thus similarly occur independently, such that NK cells in a Dd mosaic mouse acquire missing
H-2b reactivity (using H-2b-specific inhibitory receptors) but fail to
acquire missing Dd reactivity (using Ly49A). However, the selective failure to acquire missing Dd reactivity was not due to a defective inhibitory function of Ly49A (Table I). We have recently
provided evidence that Ly49A-H-2d interaction during NK cell
development has a positive impact on NK cell development (28).
This positive effect was only observed when H-2d was present on
both the radioresistant and radiosensitive compartments. In addition, the expression of a human Ig-like killer inhibitory receptor
may enhance the survival of memory T cells in Tg mice (29).
These findings are consistent with the idea that MHC-specific receptors, in addition to the well-established inhibitory role, may
perform additional functions. Indeed, besides SHP-1, the inhibitory motif in the Ly49A cytoplasmic tail recruited also SHP-2 (30).
The role of SHP-2 is unclear; however, it seems to be involved
preferentially in activating signaling cascades (31). Finally, killer
inhibitory receptors were shown to recruit phosphatidylinositol
3-kinase and thus potentially couple to signaling pathways that
provide growth and/or survival signals (32, 33). It is thus possible
that negative and putative positive functions of Ly49A are dissociated in Dd mosaic mice. Finally, the fact that NK cells react to
missing Dd once they are cultured separately from Dd-negative
host cells could also be interpreted as a rapid acquisition of missing Dd reactivity as soon as Ly49A is continuously engaged by
neighboring cells. Although we currently favor the second hypothesis, additional experiments will be required to discriminate between the two models.
The acquisition of missing self-reactivity seems to be completely prevented by ligand deficiency in some cells of the radioresistant or the radiosensitive (hematopoietic) compartment (Refs.
7, 17, and 34 and this study). NK cells in all these mice encountered ligand-deficient host cells throughout the entire body, including the bone marrow, where NK cell adaptation to the MHC class
I environment may take place. Therefore, it will be of interest to
refine the analysis and to assess the function of NK cells that
develop in a homogeneously Dd⫹ bone marrow yet encounter Dddeficient host cells in the periphery. That can now be achieved
using Tg mice, which express Cre in an organ-specific fashion. It
will also be important to determine the minimal number of H-2Dddeficient host cells required to prevent missing Dd recognition. In
this context it has been known for a long time that NK cells have
a limited capacity to eliminate grafted tumor cells (104–105 cells)
even if these cells lack MHC class I partially or completely (35,
36). The relative inefficiency of NK cells has primarily been attributed to their limited capacity for clonal expansion. The data
obtained in mosaic mice thus raise the possibility that the presence
of sufficient numbers of class I-deficient tumor cells may at some
stage disable NK cell reactions.
6262
Acknowledgments
We are grateful to I. Miconnet for assistance with i.v. injections, M. Aguet
(Swiss Institute for Experimental Cancer Research) and P. Chambon (Institut de Génétique et de Biologie Moléculaire et Cellulaire) for providing
Cre-Tg mice, W. Seaman (University of California) for providing C1498
Dd cells, and J.-C. Cerottini for critical reading of the manuscript.
References
16. Kubota, A., S. Kubota, S. Lohwasser, D. L. Mager, and F. Takei. 1999. Diversity
of NK cell receptor repertoire in adult and neonatal mice. J. Immunol. 163:212.
17. Johansson, M. H., C. Bieberich, G. Jay, K. Kärre, and P. Höglund. 1997. Natural
killer cell tolerance in mice with mosaic expression of major histocompatibility
complex class I molecules. J. Exp. Med. 186:353.
18. Höglund, P., J. Sundbäck, M. Y. Olsson-Alheim, M. Johanson, C. Ohlen, H.-G.
Ljunggren, C. L. Sentman, and K. Kärre. 1997. Host MHC class I gene control
of NK cell specificity in the mouse. Immunol. Rev. 155:11.
19. Evans, G. A., D. H. Margulies, B. Shykind, J. G. Seidman, and K. Ozato. 1982.
Exon shuffling: mapping polymorphic determinants on hybrid mouse transplantation antigens. Nature 300:755.
20. Kuhn, R., F. Schwenk, M. Aguet, and K. Rajewsky. 1995. Inducible gene targeting in mice. Science 269:1427.
21. Feil, R., J. Brocard, B. Mascrez, M. LeMeur, D. Metzger, and P. Chambon. 1996.
Ligand-activated site-specific recombination in mice. Proc. Natl. Acad. Sci. USA
93:10887.
22. Chadwick, B. S., and R. G. Miller. 1992. Hybrid resistance in vitro: possible role
of both class I MHC and self peptides in determining the level of target cell
sensitivity. J. Immunol. 148:2307.
23. Salcedo, M., M. Andersson, S. Lemieux, L. Van Kaer, B. J. Chambers, and H.-G.
Ljunggren. 1998. Fine tuning of natural killer cell specificity and maintenance of
self tolerance in MHC class I-deficient mice. Eur. J. Immunol. 28:1315.
24. Kubota, A., R. H. Lian, S. Lohwasser, M. Salcedo, and F. Takei. 1999. IFN-␥
production and cytotoxicity of IL-2-activated murine NK cells are differentially
regulated by MHC class I molecules. J. Immunol. 163:6488.
25. Mason, L. H., J. Willette-Brown, A. T. Mason, D. McVicar, and J. Ortaldo. 2000.
Interaction of Ly49D⫹ NK cells with H-2Dd target cells leads to Dap-12 phosphorylation and IFN-␥ secretion. J. Immunol. 164:603.
26. Mason, L. H., J. R. Ortaldo, H. A. Young, V. Kumar, M. Bennett, and
S. K. Anderson. 1995. Cloning and functional characteristics of murine LGL-1:
a member of the Ly-49 gene family (Ly-49G2). J. Exp. Med. 182:293.
27. Hanke, T., and D. H. Raulet. 2001. Cumulative inhibition of NK cells and T cells
resulting from engagement of multiple inhibitory Ly49 receptors. J. Immunol.
166:3002.
28. Lowin-Kropf, B., and W. Held. 2000. Positive impact of inhibitory Ly49 receptor-MHC class I interaction on NK cell development. J. Immunol. 165:91.
29. Ugolini, S., C. Arpin, N. Anfossi, T. Walzer, A. Cambiaggi, R. Förster, M. Lipp,
R. E. Toes, C. J. Melief, J. Marvel, and E. Vivier. 2001. Involvement of inhibitory
NKRs in the survival of a subset of memory-phenotype CD8⫹ T cells. Nat.
Immunol. 2:430.
30. Olcese, L., P. Lang, F. Vely, A. Cambiaggi, D. Marguet, M. Blery, K. L. Hippen,
R. Biassono, A. Moretta, L. Moretta, et al. 1996. Human and mouse killer-cell
inhibitory receptors recruit PTP1C and PTP1D protein tyrosin phosphatases.
J. Immunol. 156:4531.
31. Huyer, G., and D. R. Alexander. 1999. Immune signaling: SHP-2 docks at multiple ports. Curr. Biol. 9:R129.
32. Marti, F., A. Xu, R. Selvakumar, R. Brent, B. Dupont, and P. D. King. 1998.
LCK-phosphorylated human killer cell-inhibitory receptors recruit and activate
phosphatidylinositol 3-kinase. Proc. Natl. Acad. Sci. USA 95:11810.
33. Vanhaesebroeck, B., and M. D. Waterfield. 1999. Signaling by distinct classes of
phosphoinositide 3-kinases. Exp. Cell Res. 26:239.
34. Johansson, M. H., C. Bieberich, A. Kase-Sjöström, T. Yoshioka, E. Höglund,
B. C. Christy, G. Scangos, K. Kärre, G. Jay, and P. Höglund. 2000. Differential
effects on T cell and NK cell development by tissue-specific expression of H-2Dd
transgene. Eur. J. Immunol. 30:525.
35. Hoglund, P., H. G. Ljunggren, C. Ohlen, R. L. Ahrlund, G. Scangos,
C. Bieberich, G. Jay, G. Klein, and K. Karre. 1988. Natural resistance against
lymphoma grafts conveyed by H-2Dd transgene to C57BL mice. J. Exp. Med.
168:1469.
36. Glas, R., K. Sturmhöfel, G. J. Hämmerling, K. Kärre, and H.-G. Ljunggren. 1992.
Restoration of a tumorigenic phenotype by ␤2-microglobulin transfection to EL-4
mutant cells. J. Exp. Med. 175:843.
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
1. Hanke, T., H. Takizawa, C. W. McMahon, D. H. Busch, E. G. Pamer,
J. D. Miller, J. D. Altman, Y. Liu, D. Cado, F. A. Lemonnier, et al. 1999. Direct
assessment of MHC class I binding by seven Ly49 inhibitory NK cell receptors.
Immunity 11:67.
2. Karlhofer, F. M., R. K. Ribaudo, and W. M. Yokoyama. 1992. MHC class I
alloantigen specificity of Ly-49⫹ IL-2 activated natural killer cells. Nature 358:
66.
3. Raulet, D. H., R. E. Vance, and C. W. McMahon. 2001. Regulation of the natural
killer cell receptor repertoire. Annu. Rev. Immunol. 19:291.
4. Liao, N., M. Bix, M. Zijlstra, R. Jaenisch, and D. Raulet. 1991. MHC class I
deficiency: susceptibility to natural killer (NK) cells and impaired NK activity.
Science 253:199.
5. Hoglund, P., C. Ohlen, E. Carbone, L. Franksson, H. Ljunggren, A. Latour,
B. Koller, and K. Karre. 1991. Recognition of ␤2-microglobulin-negative (␤2m⫺)
T-cell blasts by natural killer cells from normal but not from ␤2m⫺ mice: nonresponsiveness controlled by ␤2m⫺ bone marrow in chimeric mice. Proc. Natl.
Acad. Sci. USA 88:10332.
6. Raulet, D. H. 1999. Development and tolerance of natural killer cells. Curr. Opin.
Immunol. 11:129.
7. Wu, M.-F., and D. H. Raulet. 1997. Class I-deficient hemopoietic cells and nonhemopoietic cells dominantly induce unresponsiveness of natural killer cells to
class I-deficient bone marrow cell grafts. J. Immunol. 158:1628.
8. Ohlen, C., G. Kling, P. Höglund, M. Hansson, G. Scangos, C. Bieberich, G. Jay,
and K. Karre. 1989. Prevention of allogeneic bone marrow graft rejection by H-2
transgene in donor mice. Science 246:666.
9. Held, W., D. Cado, and D. H. Raulet. 1996. Transgenic expression of the Ly49A
natural killer cell receptor confers class I major histocompatibility complex
(MHC)-specific inhibition and prevents bone marrow allograft rejection. J. Exp.
Med. 184:2037.
10. Cambiaggi, A., C. Verthuy, P. Naquet, F. Romagne, P. Ferrier, R. Biassoni,
A. Moretta, L. Moretta, and E. Vivier. 1997. Natural killer cell acceptance of H-2
mismatch bone marrow grafts in transgenic mice expressing HLA-Cw3 specific
killer cell inhibitory receptor. Proc. Natl. Acad. Sci. USA 94:8088.
11. Liu, J., M. A. Morris, P. Nguyen, T. C. George, E. Koulich, W. C. Lai,
J. D. Schatzle, V. Kumar, and M. Bennett. 2000. Ly49I NK cell receptor transgene inhibition of rejection of H2b mouse bone marrow transplants. J. Immunol.
164:1793.
12. Olsson, M. Y., K. Karre, and C. L. Sentman. 1995. Altered phenotype and function of natural killer cells expressing the major histocompatibility complex receptor Ly-49 in mice transgenic for its ligand. Proc. Natl. Acad. Sci. USA 92:
1649.
13. Dorfman, J. R., and D. H. Raulet. 1996. Major histocompatibility complex genes
determine natural killer cell tolerance. Eur. J. Immunol. 26:151.
14. Yu, Y. Y., T. George, J. Dorfman, J. Roland, V. Kumar, and M. Bennett. 1996.
The role of Ly49A and 5E6 (Ly49C) molecules in hybrid resistance mediated by
murine natural killer cells against normal T cell blasts. Immunity 4:67.
15. Valiante, N. M., M. Uhrberg, H. G. Shilling, K. Lienert-Weidenbach,
K. L. Arnett, A. D’Andrea, J. H. Phillips, L. L. Lanier, and P. Parham. 1998.
Functionally and structurally distinct NK cell receptor repertoires in the peripheral blood of two human donors. Immunity 7:739.
ACQUISITION OF MISSING SELF-REACTIVITY