products from Gram-positive bacteria,
MDP, or other agents recognized as
pathogen-associated molecular pat-
terns for 24 hr, during which time there
is a steady increase in the synthesis of
the IL-1b precursor. Most of the IL-1b
precursor remains in the cytosol, but
approximately 20% is slowly secreted
into the supernatant after cleavage by
caspase-1. Investigators studying the
inflammasome accelerate the cleav-
age of the IL-1b precursor by shocking
the cell with millimolar concentrations
of ATP. Within 15–20 min, processed
IL-1b is secreted. Whether it occurs
slowly overnight or in a matter of min-
utes, activation of caspase-1 by the
inflammasone is essential. Indeed, in
primary monocytes from patients with
the juvenile form of Still’s disease
(Pascual et al., 2005) or neonatal onset
Mansky et al., 2006) secrete signifi-
cantly more IL-1b whether in a resting
state or after endotoxin stimulation
during a 24 hr culture without the addi-
tion of ATP. Although from neutrophils
there are natural products that trigger
the P2X7 receptor and although ATP
can accumulate in the extracellular
space at sites of inflammation, the role
of pannexin-1 in the absence of the
ATP ‘‘quick fix’’ needs to be estab-
lished in primary human monocytes.
In contrast to the findings of pan-
nexin-1-mediated activation of the in-
flammasome, one must consider that
obligate intracellular micro-organisms
such as Mycobacteria and Leishmania
are not interested in inducing much of
ready to leave their host cell in much
larger numbers in order to spread
the infection. By isolating themselves
intracellularly, these microorganisms
avoid activating caspase-1. In support
of this, the Kanneganti et al. (2007)
study shows that upon entering the
cell, the stability of the endosome de-
termines the activation of caspase-1.
Chloroquine and other anti-malarials
are well-established therapeutics in
treating autoimmune diseases. Chlo-
roquine stabilized the endosome and
reduced caspase-1 activation by killed
bacteria. Therefore, the study by Kan-
neganti et al. (2007) also explains the
mechanism by which anti-malarials,
used to treat autoimmune diseases,
affects cytokine production.
Andrei, C., Margiocco, P., Poggi, A., Lotti, L.V.,
Torrisi, M.R., and Rubartelli, A. (2004). Proc.
Natl. Acad. Sci. USA 101, 9745–9750.
Boyden, E.D., and Dietrich, W.F. (2006). Nat.
Genet. 38, 240–244.
Brough, D., and Rothwell, N.J. (2007). J. Cell
Sci. 120, 772–781.
Dinarello, C.A. (1996). Blood 87, 2095–2147.
Goldbach-Mansky, R., Dailey, N.J., Canna,
S.W., Gelabert, A., Jones, J., Rubin, B.I.,
Kim, H.J., Brewer, C., Zalewski, C., Wiggs,
E., et al. (2006). N. Engl. J. Med. 355,
Kanneganti, T.-D., Lamkanfi, M., Kim, Y.-G.,
Chen, G., Park, J.-H., Franchi, L., Vandena-
beele, P., and Nunez, G. (2007). Immunity 26,
this issue, 433–443.
Larsen, C.M., Faulenbach, M., Vaag, A., Aage
Vølund, A., Ehses, J.A., Seifert, B., Mandrup-
Poulsen, T., and Donath, M.Y. (2007). N. Engl.
J. Med. 356, 1517–1526.
Martinon, F., and Tschopp, J. (2007). Cell
Death Differ. 14, 10–22.
Pascual, V., Allantaz, F., Arce, E., Punaro, M.,
and Banchereau, J. (2005). J. Exp. Med. 201,
Pelegrin, P., and Surprenant, A. (2007). J. Biol.
Chem. 282, 2386–2394.
Ready for Prime Time:
NK Cell Priming by Dendritic Cells
Eric O. Long1,*
1Laboratory of Immunogenetics, NIAID-NIH, 12441 Parklawn Drive, Rockville, MD 20852, USA
Natural killer (NK) cells were long thought to respond directly to infected cells and tumor cells. In this
issue of Immunity, Lucas et al. (2007) repeal this view by showing that NK cells acquire functionality
through priming by dendritic cells.
scribe natural killer (NK) cells as a ‘‘first
line of defense’’ against infections and
tumor cells and as potent effector cells
that secrete cytokines and kill target
cells to provide a rapid response
‘‘without the need for priming.’’ Recent
work has revealed a more complex
picture in which NK cells are an inte-
ing signals that regulate their activity
and, in turn, providing signals that
shape adaptive immune responses.
Important functional interactions be-
tween dendritic cells (DCs) and NK
cells result in their mutual regulation
(Degli-Esposti and Smyth, 2005). Ex-
amples include the following: trans-
presentation of interleukin-15 (IL-15)
by IL-15Ra on DCs in vitro stimulates
the cytotoxic activity of NK cells and
their ability to produce interferon-g
(IFN-g) (Koka et al., 2004); NK cells ac-
tivated by tumor cells or virus-infected
Immunity 26, April 2007 ª2007 Elsevier Inc.
cells in vivo prime DCs to produce
IL-12, thereby promoting T helper 1
(Th1) responses (Mailliard et al., 2003;
Mocikat et al., 2003); and activated
binding of the NKp30 receptor to a
ligand on DCs (Degli-Esposti and
By using an elegant system of in-
CD11c+DCs in mice, Lucas et al.,
(2007) have now shown that naive NK
cells do not acquire effector function
unless a priming step has occurred
by contact with DCs in draining lymph
nodes. To trigger an immune re-
sponse, the authors injected mice
with Toll-like receptor (TLR) ligands or
infected mice with viruses or the bac-
terium Listeria monocytogenes. They
then monitored the priming of NK cells
a day later by the ability of these cells
to kill or secrete IFN-g in response to
target cells in vitro. By using a combi-
nation of in vivo and in vitro exper-
iments, including cell transfers be-
tween normal and genetically altered
mice, Lucas et al. (2007) defined spe-
cific requirements for the priming of
NK cells and delineated a cascade of
secretion of type I IFN by TLR+cells,
expression of type I IFN receptor by
DCs, stimulation of IL-15Ra and IL-15
expression by DCs, and trans-presen-
tationof IL-15byDCsto NKcells inthe
lymph node (Figure 1). A requirement
for NK cells was L-selectin-dependent
migration to draining lymph nodes. NK
cell priming was independent of IL-12
and of either T or B cells. Priming ex-
periments with purified naive NK cells
in vitro demonstrated a requirement
tor+and IL-15+DCs, which had been
stimulated in vivo.
Further, Lucas et al. (2007) show
NK cells isolated from mice infected
(LCMV), with vaccinia virus, and with
Listeria had a primed phenotype,
which was dependent on CD11c+DC.
Ablation of CD11c+DCs resulted in
elevated bacterial load, suggesting a
contribution of NK cells to protection
from Listeria infection. Unfortunately, it
is very difficult to selectively monitor
the contribution of NK cell priming to
immune defense entirely in vivo. Abla-
tion of DCs dramatically impairs adap-
tive immune responses, in addition to
the loss of NK cell priming. Interfering
with IL-15 trans-presentation is not
an experimental option either, be-
cause NK cells depend on IL-15 for
survival. However, the work by Lucas
et al. (2007) does establish that prim-
ing of NK cells in vivo is mediated
specifically by CD11c+DCs, despite
expression of IL-15Ra and IL-15 on
other cells. It is possible that other
cells trans-present IL-15 to provide
the required signal for NK cell survival,
but not a complete signal for priming.
NK cells are not fully activated dur-
ing priming by DCs. Primed NK cells
express CD69 at the cell surface and
accumulate intracellular granzyme B,
but do not spontaneously produce
vation of NK cells requires additional
signals. In this respect, NK cell prim-
ing by DC in lymph nodes during
infections may differ from the NK cell
activation induced by DC–NK cell
interactions in lymph nodes during
immunization. Rapid CXCR3-depen-
dent recruitment of NK cells to lymph
nodes, resulting in secretion of IFN-g
by NK cells, has been observed after
gen (Martin-Fontecha et al., 2004).
Conceivably, sequential and qualita-
tively different DC–NK cell contacts
occur, a first one for priming of naive
NK cells and a later one resulting in
mutual activation of DC and previously
primed NK cells.
Lucas et al. (2007) show that IL-15 is
sufficient in vitro and necessary in vivo
for priming of NK cells. These results
suggest that additional steps are re-
quired to control priming of NK cells
in vivo. Naive NK cells do not emerge
from bone marrow already primed,
and yet NK cell development is cru-
presentation by stromal cells in the
bone marrow. Several hypotheses
could explain the different outcomes
of IL-15 trans-presentation in bone
marrow and in lymph nodes. First, NK
cells migrating to lymph nodes may
Figure 1. Priming of NK Cells In Vivo
Infected cells release TLR ligands (e.g., double-stranded RNA), which activate TLR+cells to pro-
duce type I IFN. The source of type I IFN is unknown, but could include plasmacytoid DC. Type I
IFN upregulates expression of IL-15Ra and IL-15 by DC. In parallel, naive NK cells migrate from
peripheral blood to draining lymph nodes. L-selectin (CD62L) on NK cells is required for homing
to lymph nodes. Direct contact of DC and NK cells in lymph nodes primes NK cells. Priming re-
quires trans-presentation of IL-15 by DC, and possibly other receptor-ligand interactions. Primed
NK cells emerge in peripheral blood and tissues within 8 hr. Additional activation signals convert
primed NK cells into effector cells.
Immunity 26, April 2007 ª2007 Elsevier Inc.
represent a more mature, priming-
competent state, whereas naive NK
cells in the bone marrow may still
be priming incompetent. Alternatively,
additional signals, besides IL-15 trans-
ing in lymph nodes. DC–NK cell con-
tacts involve many receptor-ligand
interactions, which could provide the
required signal (Figure 1).
Naive mouse NK cells do not survive
in vitro in the absence of IL-2 or IL-15.
do not bind to integrin LFA-1 ligands
and are unresponsive to target cells.
This suggests the existence of tonic
signals for survival of NK cells, which
are distinct from signals for priming.
The results of Lucas et al. (2007) imply
that DC are not required for NK cell
survival. In contrast to mouse NK cells,
human resting NK cells, freshly iso-
lated from peripheral blood, survive
several days without cytokines, bind
to LFA-1 ligands such as ICAM, re-
spond directly to ligands on target
cells, and mediate antibody-depen-
dent cellular cytotoxicity (ADCC) and
natural cytotoxicity in the absence of
cytokines (Bryceson et al., 2006). The
difference with mouse NK cells is likely
due to the constant exposure of the
human population to foreign organ-
isms, including commensal bacteria.
A normal environmental
without pathology may be sufficient
to maintain a pool of primed NK cells.
Activation of human resting NK cells
by ligands on target cells results in
degranulation by only a fraction of NK
cells (Bryceson et al., 2006), which
would be consistent with subsets of
both primed and naive NK cells in
Mice pretreated with poly(I:C), a
ligand for TLR3, are often used as
a source of ‘‘freshly isolated’’ NK cells.
Such preactivation of mouse NK cells
in vivo may result in a priming state
moresimilar to thatof ‘‘resting’’human
NK cells. Maintenance of a pool of
primed NK cells during normal envi-
ronmental exposure could be advan-
tageous. Because priming is not anti-
gen specific, primed NK cells would
be available to respond rapidly to tu-
mor cells or infected cells, even when
these cells do not elicit type I IFN
production or recruitment of NK cells
to lymph nodes.
Because individual NK cells ex-
press different combinations of acti-
vating and inhibitory receptors for
MHC class I, a selective process is
in place to ensure that each NK cell
receives sufficient inhibitory signals
to counterbalance activation signals.
This ‘‘tuning’’ process has been re-
ferred to as self-tolerance, licensing,
or education (Anfossi et al., 2006;
Raulet and Vance, 2006; Yokoyama
and Kim, 2006). As a result of this
process, the intrinsic responsiveness
of each NK cell is proportional to
the degree of inhibition received by
receptors for self MHC. Most likely,
this tuning occurs early, when NK
cells acquire their repertoire of MHC
class I-specific receptors. It would be
interesting to know whether tuning of
NK cell responsiveness impacts on
their ability to receive priming signals
from DC in lymph nodes.
Several interesting questions, which
will stimulate further work, are raised
by these new findings. What are the
exact requirements for priming? What
chemokines recruit naive NK cells to
lymph nodes? At which stage of NK
cell maturation does priming occur?
And for how long do NK cells maintain
a primed state? The need for NK cell
priming in vivo should not be seen as
an impediment to NK cell responses.
NK cells are still capable of a much
more rapid and massive response
than antigen-specific T cells and B
cells. The effector phase of NK cell
responses to virus infections peaks at
day 2 or 3, long before clonal ex-
pansion of antigen-specific cells can
provide additional and more specific
resistance to pathogens. The new
work by Lucas et al. (2007) provides
more evidence that NK cells are not
operating like a trigger-happy militia,
but rather as a disciplined regiment
that abides by certain rules of engage-
ment in order to properly target and
control their effector functions.
tynck, S., Stewart, C.A., Breso, V., Frassati, C.,
Reviron, D., Middleton, D., et al. (2006). Immu-
nity 25, 331–342.
Bryceson, Y.T., March, M.E., Ljunggren, H.G.,
and Long, E.O. (2006). Immunol. Rev. 214,
Degli-Esposti, M.A., and Smyth, M.J. (2005).
Nat. Rev. Immunol. 5, 112–124.
Koka, R., Burkett, P., Chien, M., Chai, S.,
Boone, D.L., and Ma, A. (2004). J. Immunol.
Aichele, P., and Diefenbach, A. (2007). Immu-
nity 26, this issue, 503–517.
Mailliard, R.B., Son, Y.I., Redlinger, R., Coates,
P.T., Giermasz, A., Morel, P.A., Storkus, W.J.,
and Kalinski, P. (2003). J. Immunol. 171,
Martin-Fontecha, A., Thomsen, L.L., Brett, S.,
Gerard, C., Lipp, M., Lanzavecchia, A., and
Sallusto, F. (2004). Nat. Immunol. 5, 1260–
Mocikat, R., Braumu ¨ller, H., Gumy, A., Egeter,
O., Ziegler, H., Reusch, U., Bubeck, A., Louis,
J., Mailhammer, R., Riethmu ¨ller, G., et al.
(2003). Immunity 19, 561–569.
Raulet, D.H., and Vance, R.E. (2006). Nat. Rev.
Immunol. 6, 520–531.
Immunity 26, April 2007 ª2007 Elsevier Inc.