The Journal of Experimental Medicine
The Rockefeller University Press $30.00
J. Exp. Med. Vol. 205 No. 8 1819-1828
Terms such as positive selection and deletion are
most commonly used to describe events occur-
ring during thymocyte development. Thymo-
cytes that have not correctly rearranged their
TCRs undergo a process referred to as death
by neglect and do not survive ( 1 ). Although
NK cells are not generally thought to undergo
such selective processes, recent evidence suggests
that developing NK cells may be “ licensed ” or
“ armed ” in the bone marrow to acquire a fully
functional status ( 2 – 6 ). Similar to positive selec-
tion of immature T cells that have correctly re-
arranged a functional TCR able to interact with
self-peptide – MHC, immature NK cells express-
ing inhibitory receptors may be given the
proper signals via interactions with their autolo-
gous MHC class I molecules to complete their
functional maturation ( 2, 3 ). In the absence
of these interactions, NK cells may either fail to
complete their maturation or, alternatively, may
be chronically stimulated, with each case resulting
in NK cells that are hyporesponsive ( 2, 3 ). There-
fore, selective pressures apparently are placed on
immature NK cells expressing inhibitory recep-
tors to ensure self-tolerance.
NK cells expressing activating receptors might
also undergo a selection process that would elim-
inate or inactivate developing NK cells bearing
an activating receptor recognizing a high affi n-
ity ligand expressed in the host ’ s bone marrow
in a manner similar to negative selection in thy-
mocytes. NK cells bearing the DAP12-associ-
ated activating Ly49H receptor interact with the
mouse CMV (MCMV) – encoded protein m157
on infected cells ( 7, 8 ), thus protecting mice
containing this NK cell subset against MCMV
infection ( 9 ). In a specifi c pathogen-free mouse
colony, immature NK cells will not encounter
this viral ligand during development in the bone
marrow. This provides a model system to deter-
mine how immature Ly49H + NK cells behave
when they encounter a ligand for an activating
receptor during development. Would immature
Ly49H + NK cells be deleted in a manner analo-
gous to thymocytes that have rearranged a self-
reactive TCR? Would these NK cells mature and
become chronically activated in the periphery?
Abbreviations used: ITAM,
activation motif; MCMV,
mouse CMV; MFI, mean fl uo-
The online version of this article contains supplemental material.
Tolerance of NK cells encountering
their viral ligand during development
Joseph C. Sun and Lewis L. Lanier
Department of Microbiology and Immunology and the Cancer Research Institute, University of California, San Francisco (UCSF),
San Francisco, CA 94143
During development, T and B cells encountering their cognate ligands via antigen-specifi c
receptors are deleted or rendered anergic. Like T and B cells, natural killer (NK) cells ex-
press certain receptors, such as Ly49H, associated with immunoreceptor tyrosine-based
activation motif – bearing adaptor proteins that transmit activating signals through Syk
family kinases. Ly49H binds with high affi nity to a mouse cytomegalovirus (MCMV) –
encoded glycoprotein, m157, but does not recognize self-antigens. For comparison with
the behavior of immature T and B cells exposed to foreign antigens, we addressed the fate
of Ly49H + NK cells that encountered their viral ligand during development by retroviral
transduction of bone marrow stem cells with m157. In chimeric mice expressing m157, we
observed a reduction in Ly49H + NK cells in multiple tissues and less Ly49H on the cell
surface. NK cells exposed to m157 during development appeared less mature, produced less
interferon ? when stimulated through Ly49H, and were unable to kill m157-bearing target
cells. After MCMV infection, these NK cells were severely impaired in their ability to
proliferate. Thus, if immature NK cells encounter ligands for their activating receptors,
regulatory mechanisms exist to keep these cells in an unresponsive state.
© 2008 Sun and Lanier. This article is distributed under the terms of an Attribu-
tion–Noncommercial–Share Alike–No Mirror Sites license for the fi rst six months
after the publication date (see http://www.jem.org/misc/terms.shtml). After six
months it is available under a Creative Commons License (Attribution–Noncom-
mercial–Share Alike 3.0 Unported license, as described at http://creativecommons.
NK CELL TOLERANCE | Sun and Lanier
of m157 by transient transfection of 293T cells and co-culturing
with an Ly49H-NFAT-GFP reporter cell line. m157-trans-
fected 293T cells activated > 70% of reporter cells, as measured
by GFP expression ( Fig. 1 A ). Using an antibody specifi c for
m157 ( 13 ), we demonstrated expression of this viral protein
on the surface of packaging cells transfected with the Mig-
m157 vector, on hematopoietic stem cells infected with the
Mig-m157 retrovirus, and on bone marrow cells from lethally
irradiated B6 mice reconstituted with m157 retrovirus – infected
hematopoietic stem cells 3 mo after transplantation ( Fig. 1 B ).
In contrast, m157 was not detected on cells from the Mig vec-
tor control group ( Fig. 1 B ).
Ly49H + NK cell numbers and Ly49H expression are reduced
in m157-expressing mice
NK cells in peripheral blood from the chimeric mice express-
ing m157 or the control vector were analyzed. Because stem
cells derived from congenic CD45.1 mice were injected into
irradiated CD45.2 recipient mice, donor cells were tracked
by staining with an antibody against CD45.1 ( Fig. 2 A ). Pe-
ripheral blood revealed a > 50% reduction in the percentage
of Ly49H + NK cells from m157-expressing mice compared
with Mig control mice, whereas the numbers of NK cells
bearing another DAP12-associated activating receptor, Ly49D,
were unchanged ( Fig. 2 A ). Ly49D and Ly49H are coexpressed
on a subset of double-positive cells ( 14 ); interestingly, the
percentage of these double-positive NK cells was diminished,
replaced by an increased frequency of single-positive Ly49D +
NK cells (Fig. S1, available at http://www.jem.org/cgi/content/
full/jem.20072448/DC1). Furthermore, surface Ly49H ex-
pression, as measured by the mean fl uorescence intensity (MFI)
of cells stained with anti-Ly49H mAb, was diminished on
NK cells from Mig-m157 mice compared with Mig control
mice ( Fig. 2 B ). As a control, the amount of Ly49D on NK
cells was similar between the two groups ( Fig. 2 B ), even on
the Ly49D + , Ly49H + double-positive NK cells, suggesting that
the lower expression of Ly49H observed in Mig-m157 mice
is specifi c to the expression of the Ly49H ligand. Similarly,
NKG2D was expressed at similar levels in the two groups
( Fig. 2 B ). NK cell subsets expressing inhibitory receptors
also showed no diff erences between m157-expressing and
control mice (Fig. S2). In addition, expression of inhibitory
Ly49 receptors on the Ly49H + NK cell subset from m157-
expressing mice showed no diff erences compared with control
mice (unpublished data). Reduced Ly49H + NK cell numbers
and Ly49H expression were observed in the peripheral blood
of m157-expressing mice at multiple time points after stem
cell reconstitution ( Fig. 2 C ).
3 mo after reconstitution, m157-expressing and control
Mig mice were killed, and the NK cells from multiple organs
were analyzed. The overall number of NK cells in all organs
showed no signifi cant diff erences between the two groups
(unpublished data). Similar to peripheral blood, Ly49H + NK
cell numbers and Ly49H surface expression were reduced in
the spleen, liver, and bone marrow of m157-expressing com-
pared with control mice ( Fig. 3 A ). Ly49D + NK cell numbers
Would these NK cells be desensitized to the ligand and
become hyporesponsive when challenged with infection dur-
ing adult life?
Studies to determine how T and B cells respond when
exposed to high affi nity foreign antigen during development
have typically used TCR or BCR transgenic mice against an-
tigens such as ovalbumin or hen egg lysozyme ( 10, 11 ). Be-
cause the Ly49H receptor has a high affi nity for the viral
glycoprotein m157 but does not recognize any self-antigen
( 12 ), this provides an analogous model system to evaluate the
fate of NK cells receiving strong immuno receptor tyrosine-
based activation motif (ITAM) – mediated signaling during
development. To address this question, we generated mice
that express m157 in the bone marrow and examined the
development of the Ly49H + NK cell subset. To our surprise,
we observed that Ly49H + NK cells were found in the pe-
riphery of m157-expressing mice; however, these NK cells
were present at lower numbers and were severely defective in
their ability to mediate cytotoxicity and proliferate in re-
sponse to MCMV infection.
Expression of m157 in cell lines and chimeric mice
The goal of this study was to determine the eff ects of early and
chronic stimulation of the Ly49H receptor with its viral ligand,
m157, on the development and function of NK cells. We cloned
m157 into the MigR1 retroviral vector and assessed expression
Figure 1. Expression of MCMV-encoded m157 protein in cell
lines and stem cell chimeric mice. (A) 293T cells transfected with Mig
control vector or Mig-m157 were co-cultured with the Ly49H-NFAT-GFP
reporter cell line. Percentages of reporter cells expressing GFP are shown.
(B) m157 expression in transfected Phoenix E cells, retrovirus-infected
bone marrow cells, and bone marrow cells from hematopoietic stem
cell – reconstituted irradiation chimeric mice was measured by using an
antibody against m157. Cells expressing m157 in the cells infected with
Mig control vector and the Mig-m157 construct are shown. The effi -
ciency of transduction with the control Mig and the Mig-m157 vectors,
as determined by GFP expression, was comparable (not depicted).
JEM VOL. 205, August 4, 2008
expressing animals. Therefore, the decrease in the percentage
of Ly49H + NK cells in the m157-expressing mice is not sim-
ply caused by an inability to discriminate between the Ly49H
positive and negative subsets.
Ly49H + NK cells from m157-expressing mice are
phenotypically less mature
NK1.1 is expressed early on immature NK cells, before ex-
pression of the Ly49 receptors ( 15 – 17 ). No diff erences were
observed in the amounts of NK1.1 expressed on Ly49H + NK
and Ly49D expression remained unchanged between the two
groups (unpublished data). NKG2D was also unaff ected on
NK cells in all tissues examined from m157-expressing or
control mice ( Fig. 3 A ). Although the percentages of Ly49H +
NK cells were diminished by > 60% in the spleen and liver of
m157-expressing mice, only a small decrease was observed in
the bone marrow ( Fig. 3 B ). It should be noted that although
the amount of Ly49H on the surface of the NK cells in the m157-
expressing mice was lower than in control mice, the Ly49H +
and Ly49H ? subsets were still clearly resolved in the m157-
Figure 2. Reduced Ly49H + NK cell numbers and Ly49H expression
in the peripheral blood of m157-expressing chimeric mice. (A) 28 d
after reconstitution of irradiated recipient mice with Mig control or Mig-
m157 – transduced stem cells (CD45.1 + ), peripheral blood was analyzed
for Ly49H + and Ly49D + NK cells in chimeric mice. Plots are gated on
CD3 ? , NK1.1 + cells, and percentages of Ly49H + (top) and Ly49D + (bottom)
NK cells (CD45.1 + ) are shown. MFIs for boxed regions are shown in pa-
rentheses. (B) Overlay of histogram plots comparing cell surface amounts
of Ly49H, Ly49D, and NKG2D on NK cells that express the activating re-
ceptor from Mig control and Mig-m157 chimeric mice. (C) Graphs show
the percentages of Ly49H + cells (left) and the MFI for Ly49H expression
(right) at 12, 28, and 75 d after reconstitution of irradiated recipient
mice. Data are presented as the mean ± SEM of three to fi ve mice at
each time point. Statistical differences in the percentage of Ly49H + NK
cells and the MFI of Ly49H expression between control and m157-ex-
pressing mice are indicated (*, P < 0.01; **, P < 0.001; ***, P < 0.0001).
Data are representative of three independent experiments.
Figure 3. Reduced Ly49H + NK cell numbers in the spleen, liver,
and bone marrow of m157-expressing mice. 3 mo after hematopoietic
stem cell reconstitution of irradiated recipient mice with either Mig con-
trol or Mig-m157 – transduced stem cells, mice were killed and NK cells
from the spleen, liver, and bone marrow were analyzed. (A) Plots show the
percentages of CD3 ? , NK1.1 + cells from the indicated organs that express
Ly49H and NKG2D. (B) Bar graph shows the percentages of Ly49H + NK
cells in tissues examined, presented as the mean ± SEM of three to fi ve
mice. Differences in percentages of Ly49H + NK cells between control and
m157-expressing mice were statistically signifi cant in the spleen and liver
(P < 0.01). Data are representative of three independent experiments.
NK CELL TOLERANCE | Sun and Lanier
remained higher on Ly49H + NK cells from m157-expressing
compared with control mice, with the sharpest contrast seen in
the bone marrow ( Fig. 4 A ). Other markers expressed at high
levels late during NK cell maturation, such as CD43 and
CD11b (Mac-1) ( 16 – 18 ), were present at lower amounts on
Ly49H + NK cells from m157-expressing mice compared with
controls ( Fig. 4 B ). In m157-expressing mice, Ly49H ? NK
cells did not appear phenotypically immature compared with
controls, indicating a preferential eff ect on the development of
the Ly49H + NK cells (unpublished data). Collectively, these
results suggest that NK cells encountering m157 during
development do not properly mature.
Ly49H + NK cells from m157-expressing mice produce less
IFN- ? and exhibit diminished cytotoxicity ex vivo
Because of the altered phenotype of the Ly49H + NK cells from
m157-expressing mice, we tested the functionality of these
cells ex vivo. Splenocytes were incubated with plate-bound
antibodies against NK1.1, Ly49D, or Ly49H for 4 h in the
presence of Brefeldin A, and intracellular IFN- ? production
was measured. Although anti-NK1.1 and anti-Ly49D stimula-
tion gave comparable numbers of IFN- ? – producing NK cells
between m157-expressing and control mice, anti-Ly49H stim-
ulation resulted in robust IFN- ? production only by NK cells
from Mig control mice and not NK cells from m157-express-
ing mice ( Fig. 5 ). However, this selective hyporesponsiveness
to anti-Ly49H mAb stimulation was partially rescued if the
NK cells from m157-expressing mice were cultured in the
presence of high concentrations of IL-2 (Fig. S3, available at
We tested the ability of Ly49H + NK cells from 28-d bone
marrow – reconstituted control and m157-expressing mice to
kill m157-bearing target cells. Splenocytes from both groups
of mice were enriched for NK cells by using magnetic bead
cell sorting. Comparable numbers of Ly49H + NK cells (as
determined by fl ow cytometry) were coincubated with m157-
transduced Ba/F3 targets or untransduced Ba/F3. Although
Ly49H + NK cells from Mig control mice were able to effi -
ciently lyse m157-bearing target cells, Ly49H + NK cells from
m157-expressing mice showed a severely diminished cytotoxic
response ( Fig. 6 A ). At 3 mo after reconstitution, Ly49H +
NK cells from m157-expressing mice continued to show a
decreased ability to kill BaF3-m157 target cells compared
with Mig control mice ( Fig. 6 B ). Overall, NK cell – mediated
cytotoxicity in m157-expressing mice remained intact because
lysis of Ba/F3 targets expressing Rae-1 ? (an NKG2D ligand)
or YAC-1 targets was comparable to control mice ( Fig. 6, C
and D ). Together with the cytokine data, these experiments
demonstrate that NK cells that develop in the presence of
m157 show severely diminished eff ector functions when
stimulated through the Ly49H receptor.
Reduced expansion of activated Ly49H + NK cells
from m157-expressing mice
Because the Ly49H + NK cells from m157-expressing mice
were less responsive in vitro when stimulated with anti-Ly49H
cells in the spleen, liver, and bone marrow from m157-express-
ing mice and control mice ( Fig. 4 A ). Other receptors expressed
before the Ly49 receptors, such as CD122 ( 15 – 17 ), were unal-
tered on Ly49H + NK cells from m157-expressing compared
with control Mig mice (unpublished data). Expression of CD117
(c-Kit), a developmental marker that is turned on after Ly49H
expression but then down-regulated as NK cells mature ( 15 – 17 ),
Figure 4. Ly49H + NK cells in the spleen, liver, and bone marrow of
m157-expressing mice are phenotypically less mature. (A) Histo-
grams show the expression of the developmental markers NK1.1 (top) and
c-Kit (bottom) on Ly49H + NK cells in the spleen, liver, and bone marrow of
Mig control or Mig-m157 chimeric mice. (B) Plots show expression of the
maturation markers CD43 and CD11b (Mac-1) on Ly49H + NK cells in the
spleen, liver, and bone marrow of Mig control or Mig-m157 chimeric
mice. Quadrant markers were arbitrarily set to allow visual comparison of
the most mature NK cell populations that coexpressed high amounts of
both CD43 and Mac-1 in the different tissues. Percentages are shown for
each quadrant in the dot plots. Data are representative of three indepen-
dent experiments, with three to fi ve mice per group.
JEM VOL. 205, August 4, 2008
however, the transferred NK cells (CD45.1 + ) analyzed 36 h
after MCMV infection showed very little expansion in the
spleen at this early time point (not depicted). Interestingly,
7 d after infection, the Ly49H + NK cells from m157-ex-
pressing mice were unable to expand as well as their coun-
terpart Ly49H + NK cells from control Mig mice ( Fig. 7 C ).
The endogenous Ly49H dim population (CD45.1 ? ) found in
DAP12-defi cient recipients demonstrate much less expan-
sion in response to MCMV and remain at < 50% of the entire
NK cell pool ( Fig. 7 C and not depicted). After infection,
Ly49H + NK cells from both m157-expressing and control
mice expressed high amounts of KLRG1 on the cell surface
( Fig. 7 C ), signifying activation of these proliferating cells
( 20 ). The adoptively transferred Ly49H + NK cells from con-
trol Mig mice expanded 5 – 10-fold (a statistically signifi cant
amount; P = 0.004) more than the Ly49H + NK cells from
m157-expressing mice in response to MCMV infection ( Fig.
7 D ). These experiments demonstrate that Ly49H + NK cells
exposed to their cognate viral ligand during development
are defective in their ability to “ sense ” m157 during MCMV
infection and respond in a rapid manner.
In these studies, we examined the in vivo response of
the NK cells from control Mig or Mig-m157 mice to MCMV
by adoptive transfer, rather than by simply challenging the
chimeric mice with MCMV, for several reasons. The adop-
tive transfer experiments allowed us to match the number
of transferred control and m157-exposed NK cells in the
or co-cultured with m157-bearing target cells, we tested the
in vivo functionality of these cells. Splenic NK cells from
irradiation bone marrow chimeric mice expressing m157 or
vector control were purifi ed and adoptively transferred into
DAP12-defi cient recipient mice ( Fig. 7 A ). We have previ-
ously shown that in DAP12-defi cient mice, Ly49H is ex-
pressed at much lower levels than in wild-type mice ( 19 ).
We then measured the activation and proliferation of the
transferred wild-type Ly49H + NK cells from m157-express-
ing or control mice after MCMV infection ( Fig. 7 A ). Donor
Ly49H + NK cell numbers were normalized before injection
so that the same percentage of Ly49H + cells was present in
the NK cell population of the recipient mice before infection
( Fig. 7 B ). The donor Ly49H + NK cells from m157-express-
ing mice remained Ly49H lo even after adoptive transfer into
recipient mice ( Fig. 7 B ). NK cells from both control and
m157-expressing mice displayed the typical early activation
profi le attributable to the massive infl ammatory response
induced immediately after viral infection (Fig. S4, available at
Figure 5. NK cells from m157-expressing mice produce less IFN- ?
when stimulated with anti-Ly49H. Freshly isolated splenocytes
from either Mig control or m157-expressing mice were stimulated with
plate-bound antibodies against the NK cell surface markers NK1.1,
Ly49D, Ly49H, or control (no antibody). Plots show the percentages of
CD3 ? , CD11b + , DX5 + cells that produce IFN- ? . Data are representative
of three independent experiments.
Figure 6. Ly49H + NK cells from m157-expressing mice exhibit
diminished cytotoxicity against m157-bearing targets. NK cells from
Mig control or Mig-m157 chimeric mice at 28 d (A) or 3 mo (B) after re-
constitution were MACS purifi ed. Equivalent numbers of Ly49H + NK cells,
as determined by fl ow cytometry, were incubated with Ba/F3 or m157-
transfected Ba/F3. (C) Total NK cells from Mig or Mig-m157 mice were
incubated with Ba/F3 or Rae-1 – transfected Ba/F3 target cells. (D) Total
NK cells from Mig or Mig-m157 mice were incubated with YAC-1 target
cells. Data are representative of three independent experiments.
NK CELL TOLERANCE | Sun and Lanier
vector control mice ( Fig. 8 A ). In contrast, no diff erences in
Ly49D + NK cell numbers or Ly49D surface expression were
observed between the three groups ( Fig. 8 A ). When gating on
Mig-control (CD45.2 + ) or Mig-m157 (CD45.1 + ) cells within
NK cell populations from 1:1 mixed chimeras, similarly dimin-
ished Ly49H expression was observed at days 12 and 28 after
reconstitution compared with NK cells from control Mig mice
( Fig. 8 B ). These fi ndings suggest that the regulatory eff ects of
m157 on the Ly49H + NK cell population can occur in trans
and that expression of m157 does not have to be intrinsic to
the NK cell being rendered less responsive.
The innate immune system is evolutionarily older than the
adaptive immune system. Although NK cells are considered
members of the innate immune system, they bear many strik-
ing phenotypic and functional similarities to T cells ( 21 – 23 ).
NK cells might constitute a lineage of cells that provides an
evolutionary link between the innate immune system and the
antigen-specifi c receptor-bearing cells of adaptive immunity,
possessing certain attributes of both. If NK cells are indeed a
bridge between innate and adaptive immunity, then hallmarks
of the adaptive immune system may have evolved fi rst in
this cell type. In fact, the paradigm is beginning to shift, given
recipient mice. Moreover, if we had simply infected the Mig-
m157 chimeric mice with MCMV, the host hematopoietic
cells in the mice would express m157 so that the Ly49H + NK
cells would be unable to distinguish between MCMV-infected
and normal host cells expressing the transduced m157.
Inhibition of Ly49H + NK cell maturation in trans
by cells expressing m157
To determine whether the immature phenotype and hypore-
sponsiveness of Ly49H + NK cells from m157-expressing mice
could be caused by expression of m157 on an adjacent cell dur-
ing development, we made 1:1 Mig/Mig-m157 mixed bone
marrow chimeric mice and compared NK cell populations
with m157-expressing and control Mig mice. To track indi-
vidual cell populations in 1:1 mixed chimeras, Mig-infected
control bone marrow was derived from B6 (CD45.2 + ) mice,
and m157-expressing bone marrow was derived from con-
genic CD45.1 mice. Peripheral blood on day 28 after reconsti-
tution revealed a > 50% reduction in the percentage of Ly49H +
NK cells from 1:1 mixed chimeras compared with Mig control
mice, with percentages similar to Ly49H + NK cell numbers
from m157-expressing mice ( Fig. 8 A ). Similar to m157-ex-
pressing mice, surface Ly49H expression was diminished on all
NK cells from 1:1 mixed chimeric mice compared with Mig
Figure 7. Reduced expansion of Ly49H + NK cells from m157-expressing mice. (A) Schematic diagram of experiment. NK cells from Mig control or
m157-expressing mice were MACS purifi ed, and equivalent numbers of Ly49H + NK cells were adoptively transferred to DAP12-defi cient recipient mice.
Recipient mice were infected with MCMV 1 d later. Splenic NK cells were analyzed 36 h and 7 d after infection. (B) Plots show transferred NK cells
(CD45.1 + ) in the spleen of recipient DAP12-defi cient (CD45.2 + ) mice before infection. Percentages of Ly49H + cells within the total NK cell population are
found in the gated region of each plot. (C) Plots show transferred NK cells (CD45.1 + ) in recipient DAP12-defi cient mice 7 d after MCMV infection. Percent-
ages of Ly49H + cells within the total NK cell population are gated in each plot. Activation marker KLRG1 was measured on total transferred NK cell popu-
lations. (D) Expansion of the transferred Ly49H + NK cells is shown in the graphs. The percentages of transferred Ly49H + cells in the entire NK cell
population (left) and the absolute number of transferred Ly49H + NK cells (right) are plotted for 0, 1.5, and 7 d after infection. Data are presented as the
mean ± SEM of two or three mice at each time point. Differences in the percentage and absolute number of Ly49H + NK cells between control and m157-
expressing mice were statistically signifi cant at day 7 after infection (P = 0.004). Data are representative of three independent experiments.
JEM VOL. 205, August 4, 2008
that immature NK cells undergo selective events in the bone
marrow that determine their fate in the periphery. The ques-
tion of whether developing NK cells bearing activating recep-
tors experience similar selective pressures was unaddressed.
Therefore, we sought to determine whether self-tolerance oc-
curs in developing NK cells bearing ITAM-signaling activat-
ing receptors in a manner similar to developing T or B cells.
For comparison with previous studies using TCR or BCR
transgenic mice exposed to high affi nity antigens during
development, we established a model system whereby the
ITAM-signaling Ly49H receptor expressed on NK cells would
encounter the foreign m157 glycoprotein during their develop-
ment in the bone marrow. Using this model, we demon-
strate that when developing Ly49H + NK cells encounter their
cognate viral ligand, certain regulatory mechanisms render
the cells in the periphery hyporesponsive. Rather than a com-
plete deletion, as seen in thymocytes possessing TCRs that
undergo high affi nity interactions with self-peptide – MHC
complexes ( Fig. 9 A ), Ly49H + NK cells that interact with
m157 survive but populate the periphery in lower numbers
and express diminished amounts of the Ly49H receptor on
the cell surface. Furthermore, these Ly49H-bearing NK cells
from mice that express m157 were defective in their ability
produce eff ector cytokines and mediate killing via Ly49H-
specifi c stimulation and to specifi cally expand after MCMV
infection. Perhaps the term “ disarming ” is best applied to
these activating receptor – bearing NK cells that are actively
rendered hyporesponsive by bone marrow – derived cells ex-
pressing cognate ligand for Ly49H ( Fig. 9 B ). Just as NK cells
lacking inhibitory receptors for self – MHC class I can exit the
bone marrow and populate the periphery but are refractory
to stimulation through their activating receptors, disarmed
NK cells bearing activating receptors that have encountered
their cognate ligand during development undergo a similar
fate ( Fig. 9 B ). Interestingly, there is a small population of
Ly49H + NK cells that eludes tolerance in our system, and
they can produce cytokines ( Fig. 5 ) and undergo proliferative
expansion ( Fig. 7 ) via specifi c Ly49H receptor ligation. Fur-
ther studies are required to determine the mechanism of this
escape and establish whether or not these NK cells behave
similarly to self-reactive T cells that have evaded central toler-
ance mechanisms and are either rendered anergic or mediate
autoimmunity in the periphery.
We have used m157 as a model system to evaluate the
consequences of early exposure of NK cells during their de-
velopment to a high affi nity non – self-ligand for an ITAM-
bearing receptor. Although MCMV persists in the host after
resolution of the primary infection, MCMV is preferentially
localized in the salivary gland. Therefore, it is unlikely that
NK cells, which develop predominantly in the bone mar-
row, would encounter cells expressing m157 during their early
development. Our present studies do not address the conse-
quences of prolonged exposure of mature NK cells, as com-
pared with developing NK cells, to m157. In adult mice infected
with MCMV, there is expansion of mature Ly49H + NK cells,
followed by a contraction phase in which most of these NK
recent data documenting selective pressures on developing
NK cells ( 2, 3 ), clonal expansion of NK cell subsets ( 24 – 26 ),
and possibly even immune memory in NK cells ( 27 ).
Although T cell development is relatively well under-
stood, little is known about how NK cells are regulated during
their development ( 21 – 23 ). Recently, two groups have pro-
posed comparable, yet distinct, models for mechanisms that
regulate the development of NK cells bearing inhibitory re-
ceptors for self – MHC class I ( 2, 3 ). Although the precise
mechanism for self-tolerance is yet to be determined, it is clear
Figure 8. Expression of m157 results in trans down-regulation of
Ly49H expression and Ly49H + NK cell numbers. (A) 28 d after hema-
topoietic stem cell reconstitution of irradiated recipient mice with Mig
control, Mig-m157, or a 1:1 mixture of Mig control/Mig-m157 stem cells,
peripheral blood was analyzed for Ly49H + and Ly49D + NK cells in mixed
chimeric mice. Plots are gated on CD3 ? , NK1.1 + cells, and percentages of
NKG2D + Ly49H + (top) and Ly49D + (bottom) NK cells are shown. MFIs for
boxed regions are shown in parentheses. (B) Overlay histogram plots ex-
amine the amounts of Ly49H on the cell surface of NK cells derived from
Mig control (CD45.2 + ) and Mig-m157 (CD45.1 + ) stem cell populations
within the 1:1 mixed chimeric mice. Peripheral blood from 12 and 28 d
after reconstitution are shown. Data are representative of three indepen-
dent experiments, with three to fi ve mice per group at each time point.
NK CELL TOLERANCE | Sun and Lanier
expressed by the strong ? -actin promoter in the transgenic
mice compared with the lower amounts of m157 produced
by the retroviral vector. Nonetheless, the major conclusions
from both of these complementary studies are in agreement.
Previous studies have examined the consequences of
transgenic expression of ligands for certain other NK cell re-
ceptors. In transgenic B6 mice constitutively expressing Rae-
1, NKG2D (a receptor for Rae-1) was down-regulated on
developing and mature NK cells ( 30, 31 ). These transgenic
mice were unable to reject Rae-1 – bearing tumors that are re-
jected by syngeneic, nontransgenic mice and were unable to
reject allogeneic bone marrow grafts expressing Rae-1. Our
current study fundamentally diff ers from these experiments
involving NKG2D ligand expression in several ways. Rae-1 is
a self-protein, whereas m157 is foreign. Moreover, although
Rae-1 is not expressed in most healthy adult tissues, we have
detected constitutive expression of Rae-1 in the liver, a site
where NK cells are known to traffi c ( 32 ). In addition, the sig-
naling pathways used by Ly49H and NKG2D are distinct.
Previous studies have also investigated the DAP12-associ-
ated Ly49D receptor in mice bearing H-2D d , a ligand of
Ly49D. In contrast to our fi ndings with m157 and Ly49H,
Ly49D + NK cells developing in a host expressing H-2D d
were not diminished in frequency or number ( 33 ). More-
over, in mixed bone marrow chimeras in which B6 Ly49D +
cells presumably die by apoptosis ( 24 – 26 ). Whether the re-
sponsiveness of the remaining MCMV-experienced NK cells
is aff ected has not yet been addressed.
Concurrent with our studies using retroviruses to express
m157 in mice, Tripathy et al. generated m157-expressing
transgenic mice (see Tripathy et al. [ 28 ] on p. 1829 of this
issue). In accordance with our fi ndings, these investigators
also observed a diminished number and functional impairment
of Ly49H + NK cells in these m157-transgenic mice. In our
adoptive transfer studies, we demonstrated a profound defect
in the proliferation of Ly49H + NK cells from m157-express-
ing mice after recipient mice were infected with MCMV.
This severe impairment in the expansion of these m157-
exposed Ly49H + NK cells would likely result in ineffi cient
clearance of MCMV in these hosts because the Ly49H + NK
cell subset is known to be responsible for the NK cell – medi-
ated protection of B6 mice ( 9, 25, 29 ). Indeed, Tripathy et al.
( 28 ) have observed that m157-transgenic mice are more sus-
ceptible to MCMV infection than wild-type mice, with the
lack of resistance attributed in part to the diminished cyto-
toxic function of Ly49H + NK cells in the transgenic mice.
Although there were some diff erences observed in these two
experimental models (e.g., apparently a more profound im-
pairment of non-Ly49H receptors in the m157-transgenic
mice), this might be accounted for by the amount of m157
Figure 9. Regulatory events during T and NK cell development. (A) Possible events and outcomes during T cell development in the thymus. Thymo-
cytes unable to interact with peptide-MHC do not survive and undergo death by neglect. Thymocytes interacting with peptide-MHC at low affi nity
are positively selected and survive to populate the periphery. Thymocytes interacting with peptide-MHC at high affi nity will die by negative selection.
(B) Possible events and outcomes during NK cell development in the bone marrow. Immature NK cells expressing inhibitory receptors that cannot interact
with cognate MHC will survive but are hyporesponsive. Immature NK cells expressing inhibitory receptors are able to interact with cognate MHC and
populate the periphery as functional effector cells. Immature NK cells expressing activating receptors do not interact with cognate ligand and will survive
to become functional effector cells. Immature NK cells expressing activating receptors encounter cognate ligand and survive but are hyporesponsive.
JEM VOL. 205, August 4, 2008
provided by W. Yokoyama (Washington University, St. Louis, MO). Intra-
cellular IFN- ? staining (BD Biosciences) was performed after in vitro culture
in media containing Brefeldin A for 4 h at 37 ° C ( 41 ). Flow cytometry was
performed on an LSRII and data were analyzed with CellQuest software
(both from Becton Dickinson).
Ex vivo NK cell stimulation assay. 2 × 10 6 splenocytes were cultured
in 24-well plates coated with the antibodies indicated in the fi gures at
10 μ g/ml. Intracellular staining was performed after a 4-h incubation at 37 ° C
in the presence of Brefeldin A. To avoid NK cell activation via CD16, cells
were incubated with the neutralizing anti-CD16/32 mAb, 2.4G2, before
performing stimulation assays. Because antibodies against NK1.1 were used
in some experiments for stimulation, the percentages of IFN- ? – producing
NK cells were determined by gating on CD3 ? , CD11b + , DX5 + cells.
Ex vivo NK cell – mediated cytotoxicity assays. Splenocytes were en-
riched for NK cells by using an NK cell isolation kit (Miltenyi Biotec) fol-
lowed by autoMACS magnetic bead separation. These ex vivo NK cells were
used as eff ector cells in a 6-h 51 Cr release assay ( 42 ) against Ba/F3, m157-
transfected Ba/F3 ( 7 ), YAC-1, and Rae-1 ? – transfected Ba/F3 ( 43 ) targets.
Adoptive transfer and MCMV infections. NK cells were purifi ed from
the spleens of m157-expressing or control mice by using an NK cell isolation
kit followed by autoMACS magnetic bead separation. 5 × 10 5 Ly49H + NK
cells from each group were adoptively transferred into DAP12-defi cient re-
cipient mice. The following day, 5 × 10 4 PFU of a salivary gland stock of
MCMV (Smith strain) was injected i.p.
Statistical analysis. Statistical diff erences in percentage and absolute num-
ber of Ly49H + NK cells and MFI of Ly49H expression between control and
m157-expressing mice were determined by using the two-tailed unpaired
Student ’ s t test.
Online supplemental material. Fig. S1 shows specifi c Ly49H + and
Ly49D + NK cell subsets in the peripheral blood and spleens of control and
m157-expressing mice. Fig. S2 shows the expression of activating and inhib-
itory NK cell receptors in control and m157-expressing mice. Fig. S3 shows
IFN- ? production by NK cells from control and m157-expressing mice in
the presence of anti-Ly49H mAbs and IL-2. Fig. S4 shows the early activa-
tion phenotype of Ly49H + NK cells from control and m157-expressing
mice after MCMV infection. Online supplemental material is available at
We thank the Lanier laboratory for insightful comments and helpful discussions,
and Joshua Beilke for critical review of this manuscript. We thank Wayne Yokoyama
and Sandeep Tripathy for sharing their manuscript with us before publication.
National Institutes of Health grant AI068129 supported this work. J.C. Sun is
supported by the Irvington Institute for Immunological Research. L.L. Lanier is an
American Cancer Society Research Professor.
The authors have no confl icting fi nancial interests.
Submitted: 16 November 2007
Accepted: 21 March 2008
1 . Goldrath , A.W. , and M.J. Bevan . 1999 . Selecting and maintaining a
diverse T-cell repertoire. Nature . 402 : 255 – 262 .
2 . Kim , S. , J. Poursine-Laurent , S.M. Truscott , L. Lybarger , Y.J. Song ,
L. Yang , A.R. French , J.B. Sunwoo , S. Lemieux , T.H. Hansen , and
W.M. Yokoyama . 2005 . Licensing of natural killer cells by host major
histocompatibility complex class I molecules. Nature . 436 : 709 – 713 .
3 . Fernandez , N.C. , E. Treiner , R.E. Vance , A.M. Jamieson , S. Lemieux ,
and D.H. Raulet . 2005 . A subset of natural killer cells achieves self-
tolerance without expressing inhibitory receptors specifi c for self-MHC
molecules. Blood . 105 : 4416 – 4423 .
NK cells were exposed to H-2D d cells, there was no down-
modulation of the Ly49D receptor on these cells ( 34 ). Studies
of the functional responses of these Ly49D + NK cells devel-
oping in H-2D d – bearing hosts are complicated by the fact
that a majority of Ly49D + NK cells coexpress inhibitory Ly49
receptors, such as Ly49A, Ly49C, Ly49I, and Ly49G2, that
also recognize H-2D d as a ligand. Moreover, although func-
tional studies indicate that H-2D d is a ligand for Ly49D ( 35,
36 ), there has as yet been no evidence that Ly49D can directly
bind to H-2D d ( 37 ). Indeed, we have failed to observe acti-
vation of NFAT-GFP reporter cells expressing Ly49D and
DAP12 co-cultured with H-2D d – bearing cells (unpublished
data). These results suggest that Ly49D is likely a very low af-
fi nity receptor for H-2D d , much less than the affi nity of the
inhibitory Ly49 receptors binding to H-2D d . In contrast,
there is no evidence that Ly49H binds to any known H-2
ligand, but it does bind with high affi nity to m157 ( 12 ).
Therefore, in establishing a model system to explore the con-
sequences of engaging an ITAM-signaling Ly49 receptor with
a high affi nity ligand on NK cell development, Ly49H and
m157 provide a much simpler and better-defi ned receptor –
ligand pair to address these questions.
As we are beginning to uncover the events that dictate
survival and function in NK cell development, further studies
are required to elucidate the precise molecular mechanisms
and signals by which NK cells mature and become eff ector
cells in peripheral organs. Understanding the regulatory mech-
anisms at work during NK cell development that safeguard
against peripheral autoimmunity will have important clin-
MATERIALS AND METHODS
Mice. C57BL/6 and congenic (CD45.1) mice were purchased from the
National Cancer Institute. Rag2 ? / ? × IL-2R common – ? chain ? / ? B6 mice
were purchased from Taconic. B6 DAP12-defi cient mice were bred at
UCSF. Experiments were performed according to the UCSF Institutional
Animal Care and Use Committee guidelines.
Retroviral vectors. m157 was cloned into the MigR1 vector, and virus
was generated using the Phoenix ecotropic packaging cell line ( 38 ). Cells
were incubated for 24 – 48 h, and analyzed for expression of m157 surface
protein by fl ow cytometry and by culturing overnight with Ly49H-NFAT-
GFP reporter cells, as previously described ( 7 ). An empty MigR1 vector
containing only GFP was used as a control.
Bone marrow chimeric mice. Retrovirus supernatant was added to B6
bone marrow cells ( 38 ) with 4 μ g/ml of polybrene in 6-well culture dishes
and spun at 1,300 g for 2 h at 37 ° C. After centrifugation, cells were incu-
bated overnight and the transduction was repeated the following day. On the
fourth day, 10 6 cells were injected into lethally irradiated (1,000 rad) mice
maintained on antibiotic water. Rag2 ? / ? × ? c ? / ? or DAP12-defi cient B6
mice were used as recipients because they contain no NK cells ( 39, 40 ) or
express Ly49H + NK cells that respond ineffi ciently to MCMV, respectively.
Note that a subset of DAP12-defi cient NK cells express very low levels of
Ly49H on the cell surface ( 19 ).
Flow cytometry and intracellular staining. Cells were directly stained
with antibodies against m157, NK1.1, DX5, CD3, Ly5.1, Ly49H, Ly49D,
NKG2D, CD69, KLRG1, c-Kit, CD43, and Mac-1 (CD11b; all obtained
from eBioscience or BD Biosciences). Anti-m157 and anti-Ly49H were
NK CELL TOLERANCE | Sun and Lanier
crete subset of natural killer cells reactive with monoclonal antibody to
Ly49H. J. Exp. Med. 194 : 29 – 44 .
26 . Andrews , D.M. , A.A. Scalzo , W.M. Yokoyama , M.J. Smyth , and M.A.
Degli-Esposti . 2003 . Functional interactions between dendritic cells and
NK cells during viral infection. Nat. Immunol. 4 : 175 – 181 .
27 . O ’ Leary , J.G. , M. Goodarzi , D.L. Drayton , and U.H. von Andrian .
2006 . T cell- and B cell-independent adaptive immunity mediated by
natural killer cells. Nat. Immunol. 7 : 507 – 516 .
28 . Tripathy , S.K. , P.A. Keyel , L. Yang , J.T. Pingel , T.P. Cheng , A.
Schneeberger , and W.M. Yokoyama . 2008 . Continuous engagement
of a self-specifi c activation receptor induces NK cell tolerance. J. Exp.
Med. 205 : 1829 – 1841 .
29 . Lee , S.-H. , S. Gerard , D. Macina , M. Busa , A. Zafer , A. Belouchi , P.
Gros , and S.M. Vidal . 2001 . Susceptibility to mouse cytomegalovirus is
associated with depletion of an activating natural killer cell receptor of
the C-type lectin superfamily. Nat. Genet. 28 : 42 – 45 .
30 . Ogasawara , K. , J. Benjamin , R. Takaki , J.H. Phillips , and L.L. Lanier .
2005 . Function of NKG2D in natural killer cell-mediated rejection of
mouse bone marrow grafts. Nat. Immunol. 6 : 938 – 945 .
31 . Oppenheim , D.E. , S.J. Roberts , S.L. Clarke , R. Filler , J.M. Lewis , R.E.
Tigelaar , M. Girardi , and A.C. Hayday . 2005 . Sustained localized expression
of ligand for the activating NKG2D receptor impairs natural cytotoxicity in
vivo and reduces tumor immunosurveillance. Nat. Immunol. 6 : 928 – 937 .
32 . Vilarinho , S. , K. Ogasawara , S. Nishimura , L.L. Lanier , and J.L. Baron .
2007 . Blockade of NKG2D on NKT cells prevents hepatitis and the
acute immune response to hepatitis B virus. Proc. Natl. Acad. Sci. USA .
104 : 18187 – 18192 .
33 . George , T.C. , J.R. Ortaldo , S. Lemieux , V. Kumar , and M. Bennett .
1999 . Tolerance and alloreactivity of the Ly49D subset of murine NK
cells. J. Immunol. 163 : 1859 – 1867 .
34 . Zhao , Y. , H. Ohdan , J.O. Manilay , and M. Sykes . 2003 . NK cell toler-
ance in mixed allogeneic chimeras. J. Immunol. 170 : 5398 – 5405 .
35 . George , T.C. , L.H. Mason , J.R. Ortaldo , V. Kumar , and M. Bennett .
1999 . Positive recognition of MHC class I molecules by the Ly49D
receptor of murine NK cells. J. Immunol. 162 : 2035 – 2043 .
36 . Nakamura , M.C. , P.A. Linnemeyer , E.C. Niemi , L.H. Mason , J.R.
Ortaldo , J.C. Ryan , and W.E. Seaman . 1999 . Mouse Ly-49D recog-
nizes H-2Dd and activates natural killer cell cytotoxicity. J. Exp. Med.
189 : 493 – 500 .
37 . 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 – 77 .
38 . Sun , J.C. , S.M. Lehar , and M.J. Bevan . 2006 . Augmented IL-7 signal-
ing during viral infection drives greater expansion of eff ector T cells but
does not enhance memory. J. Immunol. 177 : 4458 – 4463 .
39 . Cao , X. , E.W. Shores , J. Hu-Li , M.R. Anver , B.L. Kelsall , S.M. Russell ,
J. Drago , M. Noguchi , A. Grinberg , E.T. Bloom , et al . 1995 . Defective
lymphoid development in mice lacking expression of the common cy-
tokine receptor gamma chain. Immunity . 2 : 223 – 238 .
40 . DiSanto , J.P. , W. Muller , D. Guy-Grand , A. Fischer , and K. Rajewsky .
1995 . Lymphoid development in mice with a targeted deletion of the inter-
leukin 2 receptor gamma chain. Proc. Natl. Acad. Sci. USA . 92 : 377 – 381 .
41 . Sun , J.C. , M.A. Williams , and M.J. Bevan . 2004 . CD4+ T cells are
required for the maintenance, not programming, of memory CD8+ T
cells after acute infection. Nat. Immunol. 5 : 927 – 933 .
42 . Hesslein , D.G. , R. Takaki , M.L. Hermiston , A. Weiss , and L.L. Lanier .
2006 . Dysregulation of signaling pathways in CD45-defi cient NK cells
leads to diff erentially regulated cytotoxicity and cytokine production.
Proc. Natl. Acad. Sci. USA . 103 : 7012 – 7017 .
43 . Cerwenka , A. , A.B. Bakker , T. McClanahan , J. Wagner , J. Wu , J.H.
Phillips , and L.L. Lanier . 2000 . Retinoic acid early inducible genes
defi ne a ligand family for the activating NKG2D receptor in mice.
Immunity . 12 : 721 – 727 .
4 . Yokoyama , W.M. , and S. Kim . 2006 . How do natural killer cells fi nd
self to achieve tolerance? Immunity . 24 : 249 – 257 .
5 . Raulet , D.H. 2006 . Missing self recognition and self tolerance of natural
killer (NK) cells. Semin. Immunol. 18 : 145 – 150 .
6 . Raulet , D.H. , and R.E. Vance . 2006 . Self-tolerance of natural killer
cells. Nat. Rev. Immunol. 6 : 520 – 531 .
7 . Arase , H. , E.S. Mocarski , A.E. Campbell , A.B. Hill , and L.L. Lanier .
2002 . Direct recognition of cytomegalovirus by activating and inhibi-
tory NK cell receptors. Science . 296 : 1323 – 1326 .
8 . Smith , H.R. , J.W. Heusel , I.K. Mehta , S. Kim , B.G. Dorner , O.V.
Naidenko , K. Iizuka , H. Furukawa , D.L. Beckman , J.T. Pingel , et al .
2002 . Recognition of a virus-encoded ligand by a natural killer cell
activation receptor. Proc. Natl. Acad. Sci. USA . 99 : 8826 – 8831 .
9 . Brown , M.G. , A.O. Dokun , J.W. Heusel , H.R. Smith , D.L. Beckman ,
E.A. Blattenberger , C.E. Dubbelde , L.R. Stone , A.A. Scalzo , and W.M.
Yokoyama . 2001 . Vital involvement of a natural killer cell activation
receptor in resistance to viral infection. Science . 292 : 934 – 937 .
10 . Starr , T.K. , S.C. Jameson , and K.A. Hogquist . 2003 . Positive and neg-
ative selection of T cells. Annu. Rev. Immunol. 21 : 139 – 176 .
11 . Cambier , J.C. , S.B. Gauld , K.T. Merrell , and B.J. Vilen . 2007 . B-cell
anergy: from transgenic models to naturally occurring anergic B cells?
Nat. Rev. Immunol. 7 : 633 – 643 .
12 . Adams , E.J. , Z.S. Juo , R.T. Venook , M.J. Boulanger , H. Arase , L.L.
Lanier , and K.C. Garcia . 2007 . Structural elucidation of the m157
mouse cytomegalovirus ligand for Ly49 natural killer cell receptors. Proc.
Natl. Acad. Sci. USA . 104 : 10128 – 10133 .
13 . Tripathy , S.K. , H.R. Smith , E.A. Holroyd , J.T. Pingel , and W.M.
Yokoyama . 2006 . Expression of m157, a murine cytomegalovirus-
encoded putative major histocompatibility class I (MHC-I)-like pro-
tein, is independent of viral regulation of host MHC-I. J. Virol. 80 :
545 – 550 .
14 . Smith , H.R. , H.H. Chuang , L.L. Wang , M. Salcedo , J.W. Heusel , and
W.M. Yokoyama . 2000 . Nonstochastic coexpression of activation re-
ceptors on murine natural killer cells. J. Exp. Med. 191 : 1341 – 1354 .
15 . Rosmaraki , E.E. , I. Douagi , C. Roth , F. Colucci , A. Cumano , and J.P.
Di Santo . 2001 . Identifi cation of committed NK cell progenitors in
adult murine bone marrow. Eur. J. Immunol. 31 : 1900 – 1909 .
16 . Colucci , F. , M.A. Caligiuri , and J.P. Di Santo . 2003 . What does it take
to make a natural killer? Nat. Rev. Immunol. 3 : 413 – 425 .
17 . Vosshenrich , C.A. , S.I. Samson-Villeger , and J.P. Di Santo . 2005 .
Distinguishing features of developing natural killer cells. Curr. Opin.
Immunol. 17 : 151 – 158 .
18 . Kim , S. , K. Iizuka , H.S. Kang , A. Dokun , A.R. French , S. Greco , and
W.M. Yokoyama . 2002 . In vivo developmental stages in murine natural
killer cell maturation. Nat. Immunol. 3 : 523 – 528 .
19 . Bakker , A.B. , R.M. Hoek , A. Cerwenka , B. Blom , L. Lucian , T.
McNeil , R. Murray , L.H. Phillips , J.D. Sedgwick , and L.L. Lanier .
2000 . DAP12-defi cient mice fail to develop autoimmunity due to im-
paired antigen priming. Immunity . 13 : 345 – 353 .
20 . Huntington , N.D. , H. Tabarias , K. Fairfax , J. Brady , Y. Hayakawa ,
M.A. Degli-Esposti , M.J. Smyth , D.M. Tarlinton , and S.L. Nutt .
2007 . NK cell maturation and peripheral homeostasis is associated with
KLRG1 up-regulation. J. Immunol. 178 : 4764 – 4770 .
21 . Lanier , L.L. 2005 . NK cell recognition. Annu. Rev. Immunol.
23 : 225 – 274 .
22 . Yokoyama , W.M. , S. Kim , and A.R. French . 2004 . The dynamic life of
natural killer cells. Annu. Rev. Immunol. 22 : 405 – 429 .
23 . Di Santo , J.P. 2006 . Natural killer cell developmental pathways: a ques-
tion of balance. Annu. Rev. Immunol. 24 : 257 – 286 .
24 . Dokun , A.O. , S. Kim , H.R. Smith , H.S. Kang , D.T. Chu , and W.M.
Yokoyama . 2001 . Specifi c and nonspecifi c NK cell activation during
virus infection. Nat. Immunol. 2 : 951 – 956 .
25 . Daniels , K.A. , G. Devora , W.C. Lai , C.L. O ’ Donnell , M. Bennett ,
and R.M. Welsh . 2001 . Murine cytomegalovirus is regulated by a dis-