JOURNAL OF VIROLOGY, Apr. 2010, p. 3993–4001
Copyright © 2010, American Society for Microbiology. All Rights Reserved.
Vol. 84, No. 8
Killing of Avian and Swine Influenza Virus by Natural Killer Cells?
Hagit Achdout,1† Tal Meningher,2,3† Shira Hirsh,2,3Ariella Glasner,1Yotam Bar-On,1Chamutal Gur,1
Angel Porgador,4Michal Mendelson,4Michal Mandelboim,3† and Ofer Mandelboim1†*
The Lautenberg Center for General and Tumor Immunology, IMRIC, the Hebrew University Hadassah Medical School, Jerusalem,
Israel1; The Mina and Everard Goodman Faculty of Life Sciences, Bar-Ilan University, Ramat-Gan, 52900, Israel2;
Central Virology Laboratory, Ministry of Health, Public Health Services, Chaim, Sheba Medical Center,
Tel Hashomer, Ramat-Gan, Israel3; and Shraga Segal Department of Microbiology and Immunology,
Faculty of Health Sciences, and Cancer Research Center, Ben-Gurion University of
the Negev, Beer Sheva 84105, Israel4
Received 29 October 2009/Accepted 26 January 2010
Today, global attention is focused on two influenza virus strains: the current pandemic strain, swine origin
influenza virus (H1N1-2009), and the highly pathogenic avian influenza virus, H5N1. At present, the infection
caused by the H1N1-2009 is moderate, with mortality rates of less <1%. In contrast, infection with the H5N1
virus resulted in high mortality rates, and ca. 60% of the infected patients succumb to the infection. Thus, one
of the world greatest concerns is that the H5N1 virus will evolve to allow an efficient human infection and
human-to-human transmission. Natural killer (NK) cells are one of the innate immune components playing an
important role in fighting against influenza viruses. One of the major NK activating receptors involved in NK
cell cytotoxicity is NKp46. We previously demonstrated that NKp46 recognizes the hemagglutinin proteins of
B and A influenza virus strains. Whether NKp46 could also interact with H1N1-2009 virus or with the avian
influenza virus is still unknown. We analyzed the immunological properties of both the avian and the
H1N1-2009 influenza viruses. We show that NKp46 recognizes the hemagglutinins of H1N1-2009 and H5 and
that this recognition leads to virus killing both in vitro and in vivo. However, importantly, while the swine
H1-NKp46 interactions lead to the direct killing of the infected cells, the H5-NKp46 interactions were unable
to elicit direct killing, probably because the NKp46 binding sites for these two viruses are different.
Natural killer (NK) cells, which comprise 5 to 15% of pe-
ripheral blood lymphocytes, are a key frontline defense against
a number of pathogens, including intracellular bacteria, para-
sites, and most importantly with respect to the present study,
viruses (6, 40). The antiviral mechanisms by which NK cells
operate include both cytotoxic activity and cytokine/chemokine
secretion (21). The NK killing activity is executed by numerous
receptors, including NKG2D, NKp80, CD16, and the natural
cytotoxic receptors (NCRs): NKp30, NKp44, and NKp46 (7,
Although the cellular ligands for NKG2D were identified
(31, 38), the identity of several of the cellular ligands for the
human NCRs is still unknown, except for BAT3 and B7-H6,
which are ligands for NKp30 (8, 30). In contrast, viral ligands
were identified for the NCRs, and we demonstrated that pp65
of HCMV interacts with NKp30 (3) and that various influenza
virus hemagglutinins (HAs) are ligands for the NKp44 and
NKp46 receptors (5, 22). Supporting these observations, it was
recently shown that the HA-neuraminidase of Newcastle dis-
ease virus could also interact with NKp46 and NKp44 but not
with NKp30 (17). Furthermore, we have shown in vivo that in
the absence of NCR1 (the mouse homologue of NKp46),
A/PR8 influenza virus infection is lethal (14).
Human influenza virus (H1 and H3 subtype) infections pose
a major threat to the entire population, as exemplified by the
three major influenza pandemics that occurred during the 20th
century. The Asian (A/H2N2) in 1957 to 1958 and the Hong
Kong (A/H3N2) pandemics in 1968 to 1969 resulted in the
deaths of 1 to 2 million people and the 1918 “Spanish flu”
(A/H1N1) pandemic killed around 50 million people (18). At
present, the worldwide concern regarding influenza pandemics
concentrates mainly on two viruses: the A/H1N1 swine origin
influenza virus (H1N1-2009), which currently causes only a
moderate pandemic (the mortality rates are ca. 1%) but is
more pathogenic than a regular seasonal influenza virus (19,
26, 27), and the avian influenza virus carrying the unique H5
HA (20). The avian influenza virus is quite deadly and, al-
though it remains a zoonotic infection, ca. 60% of infected
humans died due to the infection (28).
The unique properties of the H5 protein of the avian influ-
enza virus are one of the main reasons for the virulence of the
virus. The H5 of the avian influenza virus binds to cell surface
glycoproteins or glycolipids containing terminal sialyl-galacto-
syl residues linked by 2-3-linkage [Neu5Ac(?2-3)Gal] that are
found in the human conjunctiva and ciliated portion of the
respiratory columnar epithelium (33). In contrast, human vi-
ruses (including all three strains that caused the pandemics
described above and the H1N1-2009) bind to receptors that
mostly contain terminal 2-6-linked sialyl-galactosyl moieties
[Neu5Ac(?2-6)Gal]. Such glycosylations are predominant on
epithelial cells in the nasal mucosa, paranasal sinuses, pharynx,
trachea, and bronchi (33, 37). It has been suggested that the
lack of human-to-human transmission of avian influenza vi-
ruses is due to their ?2,3-SA receptor binding preference, and
* Corresponding author. Mailing address: Lautenberg Center for
General and Tumor Immunology, IMRIC, The Hebrew University
Hadassah Medical School, Jerusalem, Israel. Phone: 972-2-6757515.
Fax: 972-2-6424653. E-mail: firstname.lastname@example.org.
† H.A., T.M., M.M., and O.M. contributed equally to this study.
?Published ahead of print on 3 February 2010.
the concern is that genetic changes in H5 might alter its pref-
erence from ?2,3-SA to ?2,6-SA, allowing human-to-human
In our previous studies (4, 22) we showed that the interac-
tion between NKp46 and influenza virus HAs depends on the
sialylation of the NKp46 receptor. We further demonstrated
that the sialic acid residues, which are linked via ?2,6 to the
threonine 225 residue of NKp46, are crucial for the NKp46
interactions with the various influenza virus HAs (4).
We show that, both in vitro and in vivo, the killing of H1N1-
2009-infected cells is correlated with the degree of NKp46
binding. Surprisingly, we observed that although NKp46 effi-
ciently recognized the avian H5 HA, such interactions were
unable to elicit the direct killing of the infected cells. By using
mutagenesis analysis experiments and killing assays we dem-
onstrate that NKp46 interacts with H1 and H5 at distinct sites,
since we show that the sugar carrying residue at position 225 is
crucial for the NKp46-H1N1-2009 interactions, whereas the
interaction of H5 with NKp46 depends on both residues 216
MATERIALS AND METHODS
Cells. The cell lines used in the present study were the human hepatoma cell
line Hep3b, the human choriocarcinoma cell line Jeg3, and the mouse lympho-
blastlike mastocytoma cell line P815. For the generation of Jeg3 cells expressing
MICB (Jeg3/MICB), we inserted the cDNA sequence of human MICB instead
of the green fluorescent protein (GFP) of the lentiviral vector SIN18-pRLL-
hEFIap-EGFP-WRPE (35) and stably infected Jeg3 cells. As a control, we
infected Jeg3 cells with the intact vector SIN18-pRLL-hEFIap-EGFP-WRPE,
generating Jeg3/GFP cells. Primary NK cells were isolated from PBLs using a
human NK cell isolation kit and the autoMACS instrument (Miltenyi Biotec,
Auburn, CA). NK cells were kept in culture as described previously (24).
MAbs and fusion proteins. The monoclonal antibodies (MAbs) used in the
present study included anti-NKp46 MAbs, 461-G1 (IgG1) (4), and clone 9E2
(IgG1) (11), as well as anti-influenza virus type A (H1), H17-L2, and H28-E23
anti-H1 (kindly provided by Jonathan W. Yewdell). As negative controls, we
used anti-CD3 MAb T3D and the anti-CD99 antibody, 12E7 (IgG1). The anti-
MICA, MICB, ULBP1-3, and NKp30 antibodies were all purchased from R&D
Systems. Anti-NKG2D hybridoma C7 used for the in vivo NKG2D blocking was
kindly provided by W. M. Yokoyama.
The fusion proteins used in the present study were generated by fusion of the
extracellular of the various human receptors to a human IgG1, as described
previously (23). The point mutations in the NKp46 protein—T125V, N216V, and
T225V—were generated by using a PCR-based, site-directed mutagenesis ap-
proach, as described previously (4).
Viruses and viral infection. vnh5n1-pr8/cdc-rg H5N1 (abbreviated as H5
[avian]) (16) and the human flu viruses A/Puerto Rico/8/34 H1N1 (abbreviated
as A/PR8), A/Texas H3N2 (abbreviated as H3N2), and A/Swine/Israel/2009
H1N1 (abbreviated as H1N1-2009) were generated, and the cells were infected
as described previously (1).
Cytotoxicity assay. The cytotoxic activity of primary bulk NK cells against the
various target cells was assessed in 5-h35S release assays, as previously described
(24). In experiments in which MAbs were included, the final MAb concentration
was 5 ?g/ml. In all assays, the spontaneous release was ?25% of the maximal
Mice experiments. All experiments were performed using 12- to 16-week-old
C57BL/6 mice. The generation of NKp46/NCR1 knockout mice was previously
described (14). For influenza virus infection, mice were anesthetized and inoc-
ulated intranasally with 1.6 50% tissue culture infective dose(s) (TCID50) of the
H5N1 virus/mouse or with 16 TCID50of the H1N1-2009 virus/mouse. For
NKG2D blocking, 6 h prior to infection 300 ?g of C7 purified MAb/mouse was
injected intraperitoneally. To measure the infection efficiency, mice were sacri-
ficed 5 or 6 days after virus inoculation; the lungs were then removed, homog-
enized in Dulbecco modified Eagle medium, and stored at ?70°c. To determine
the virus titer, the lungs were thawed and homogenized with an OMNI homog-
enizer, RNA was extracted from the lungs homogenate, and the virus titer was
measured by a real-time reverse transcription-PCR (RT-PCR) assay. The virus
titer was determined based on the viral genome copy numbers (15).
NKp46 recognizes avian influenza virus-infected cells. We
have previously shown that NKp46, but not NKp30, recognizes
cells infected with human influenza viruses (5, 22) and that this
interaction is dependent on the sialic acid residues of NKp46
(4). To test whether NKp46 will also recognize avian virus-
infected cells, we infected the choriocarcinoma cell line with
A/PR8 H1N1 and with another virus containing the core of the
A/PR8 virus and the envelope of the H5N1 virus (H5 avian
virus) (16). We used this particular H5N1 virus because it
differs from A/PR8 H1N1 only in its HA and NA proteins. As
shown in Fig. 1A, the H1 and H5 proteins are expressed on the
cell surface after the infection. Importantly, as we previously
reported earlier (4), increased binding of NKp46-Ig was ob-
served both to the H1 and, as seen here, also to the H5-
infected cells (Fig. 1B). In agreement with our previous results,
no binding of NKp30-Ig was observed to any of the cells irre-
spective of whether they were infected or not (Fig. 1B).
We next tested the killing of the Jeg3-infected cells by NK
cells. Surprisingly, although NKp46-Ig efficiently recognized
the H5-infected Jeg3 cells (Fig. 1), the killing of NK cells was
not enhanced (Fig. 2A), even when higher effector-to-target
ratios were used (data not shown). In marked contrast, and as
we previously reported (22), an increased killing was observed
when the Jeg3 cells were infected with the A/PR8 influenza
virus (Fig. 2A) or with other human influenza viruses (data not
shown). This increased killing of A/PR8 influenza virus-in-
FIG. 1. NKp46 recognition of influenza virus-infected Jeg3 cells.
fluorescence-activated cell sorting (FACS) staining of uninfected Jeg3
cells (gray filled histogram) or Jeg3 cells infected with A/PR8 or H5
(avian) viruses (black line). Staining was performed with anti-influenza
virus A (Anti-flu A), NKp30-Ig, and NKp46-Ig fusion proteins (B) (in-
dicated in the x axis of the figure). The figure shows the results for one
representative experiment out of four performed.
3994ACHDOUT ET AL.J. VIROL.
fected cells was due to the interaction between NKp46 and the
H1 of A/PR8 virus, as it was blocked by an anti-H1 MAb (Fig.
2B). We therefore concluded that NKp46 could interact with
H5, but this recognition was insufficient to activate killing by
Killing of the avian virus-infected cells is possible when
several NK activating receptors are engaged. Our hypothesis
to explain these results was that maybe the NKp46-H5 inter-
actions are insufficient by themselves to induce direct killing
and that maybe for this particular interaction the help of other
killer receptors such as NKG2D, which was shown to partici-
pate in influenza virus eradication (12, 34), is needed. This
might explain why, when we infected Jeg3 cells that do not
express any of the known killer ligands (Fig. 3A) and, indeed,
are almost resistant to killing (Fig. 2), the interaction of H5
with NKp46 was insufficient to induce direct killing. To test this
option, we used the human hepatoma cell line, Hep3b, which
expresses killer ligands such as MICA, MICB, and unknown
cellular ligands for the NCRs (Fig. 3B); infected these cells
with the A/PR8 or with the H5 viruses; and tested their killing
by NK cells. Interestingly, a similar increase in the killing of the
infected cells was observed when Hep3b cells were infected
with A/PR8 or with H5N1 influenza viruses (Fig. 3C).
To directly demonstrate that H5 is able to trigger the
NKp46-mediated killing only in the presence of additional
ligands, we expressed the MICB protein (a ligand of the acti-
vating NKG2D receptor ) in Jeg3 cells (Jeg3/MICB) and
also expressed GFP (Jeg3/GFP), which was used as a control
(Fig. 3D). The various Jeg3 cells were infected with the H5
influenza virus and then used in killing assays (Fig. 3E). As
expected, the presence of MICB in Jeg3 cells slightly enhanced
the killing of the uninfected cells, probably because it is rec-
ognized by the activating receptor NKG2D. Importantly, and
as we predicted, the killing of Jeg3/MICB cells infected with
the H5 virus was substantially increased compared to the kill-
ing of uninfected Jeg3/MICB cells (Fig. 3E).
NKp46 interacts with H5 and H1 in distinct binding sites.
We have previously shown that the binding of NKp46 to the
various HAs (excluding H5) is restricted to the membrane-
proximal domain of the receptor, requires the sialylation of
NKp46, and preferentially requires the Neu5Ac ?(2,6)-Gal
linkage (4, 22). In contrast, the avian H5 protein preferentially
binds to ?2,3-sialosides and only weakly binds to certain ?2,6-
NKp46 contains three potential glycosylation sites located at
Thr125, Asn216, and Thr225 (Fig. 4A) (29), of which Thr225 is
the main amino acid residue, critically involved in the HA
interaction with NKp46 (4). To test whether the lack of direct
killing of H5-expressing cells could be explained by a differen-
tial binding of NKp46 to H1 versus H5, we mutated the three
potential glycosylation residues of NKp46 by replacing them
with valine (T125V, N216V, and T225V). The mutated NKp46
proteins were appended to human Fc, and the binding of the
mutated proteins was compared to the binding of the wild-type
NKp46 protein. Importantly, whereas only a slight reduction in
the binding of the mutated protein T125V-Ig to the H5-in-
fected Jeg3 was observed, both N216V-Ig and T225V-Ig
showed markedly reduced binding to H5-infected Jeg3 cells
(Fig. 4B). Thus, whereas T225 plays a crucial role in the inter-
action between H1 and NKp46 (4) (see also Fig. 6 below), with
regard to H5, two glycosylation sites on NKp46, located at
positions N216 and T225, are involved in the binding.
To further establish that the binding sites of H5 and H1 in
NKp46 are distinct, we tested the ability of different anti-
NKp46 MAbs to inhibit the NKp46-mediated killing of Jeg3/
MICB cells infected with the A/PR8 or with the H5 viruses. As
shown in Fig. 5, the 461-G1 MAb, which is directed against the
D1 domain of NKp46 (a domain that does not contain glyco-
sylation and is not involved in the killing of influenza virus-
infected cells ), did not block the killing. Importantly, block-
ing of NKp46 with the 9E2 MAb significantly reduced the
killing of the H5-infected cells but did not affect the killing of
the A/PR8-infected cells (Fig. 5), suggesting that indeed
NKp46 interacts with H1 and H5 at different binding sites.
NKp46 interaction with H1 of the H1N1-2009 virus leads to
direct NK cytotoxicity. In June 2009 the World Health Orga-
nization (WHO) raised the influenza pandemic to level 6,
stating that the current H1N1-2009 infection is the first pan-
demic of this century and the first in the last 41 years. To test
whether the H1 HA of the H1N1-2009 virus would be recog-
nized by NKp46, we infected Jeg3 cells with 19 different H1N1-
2009 viruses that were isolated from Israeli patients. The in-
fected cells were tested for binding of NKp46-Ig and NKp46
T225V-Ig proteins (Fig. 6A) and for killing by NK cells (Fig.
6B). Interestingly, we found that NKp46-Ig binds to all cells
infected with the 19 different H1N1-2009 viruses (Fig. 6A) and
the binding was correlated with the efficiency of Jeg3-mediated
killing (Fig. 6B). Little or no binding of T225V-Ig was observed
FIG. 2. Jeg3 cells infected with avian influenza virus are not killed
by NK cells. (A) Killing assay. Jeg3 cells, either uninfected or infected
with A/PR8 or H5 (avian) influenza viruses, were tested in killing
assays by using bulk NK cells. The effector to target ratio (E:T) is
indicated in the figure. (B) The killing of Jeg3 cells is mediated by the
interaction with HA. Jeg3 cells infected or not with A/PR8 influenza
virus were preincubated with anti-H1 MAb, H28-E23 or with anti-
CD99 MAb (negative control) and were tested in killing assays by bulk
NK cells at an E:T ratio of 10:1. The figure shows the results for one
representative experiment of three performed. Error bars indicate the
VOL. 84, 2010 NK CELL RECOGNITION OF AVIAN AND SWINE INFLUENZA VIRUS3995
FIG. 3. The killing of the avian-infected cells is possible when several NK activating receptors are engaged. (A) FACS staining of Jeg3 cells for NKG2D
ligands. Staining was performed with anti-MICA, MICB, and ULBP1-3 MAbs. (B) FACS staining of Hep3b cells for NKG2D and NCR ligands. Staining was
performed with anti-MICA, MICB, ULBP1-3 MAbs, NKp30-Ig, NKp44-Ig, and NKp46-Ig fusion proteins (indicated in the x axis of the figure). (C) Killing
experiment of Hep3b cells infected or not with A/PR8 or with H5N1 viruses by bulk NK cells. The various E:T ratios are indicated in the figure. (D) FACS
line) and with either a negative control MAb (blue line) or with the secondary MAb (black line). In the two right histograms the GFP intensity of the indicated
the results for one representative experiment out of three performed. Error bars indicate standard deviations.
(Fig. 6A), confirming our previous observations (4) that the
glycosylation at position 225 of NKp46 is important for the
NKp46-H1 interactions. Thus, while the avian influenza virus
recognition by NKp46 is unable to directly induce NK cytotox-
icity, the recognition of the swine influenza virus by NKp46
plays a critical role in the killing of the infected cells.
Engagement of NKG2D is essential for in vitro and in vivo
NKp46-mediated killing of avian influenza virus. If indeed
the NKp46-H5 interactions are insufficient to induce direct
NKp46-mediated killing by themselves, then by blocking the
NKG2D receptor we should be able to reduce the NKp46-
mediated killing of the H5-infected cells. To test this assump-
tion, we performed a redirected killing assay in which we in-
fected the P815 cell line with H5N1 avian influenza virus or
with the human influenza virus strain A/H3N2. We did not use
the A/PR8 or the H1N1-2009 viruses since these viruses could
not infect the P815 cells (data not shown). As expected, the
uninfected and the H5-infected P815 cells were not killed when
we used a control MAb, since uninfected P815 cells do not
express lysis ligands for NK cells and, as shown above, the
interaction between NKp46 and H5 is insufficient by itself to
induce direct killing. In contrast, the H3N2-infected P815 cells
were killed by NK cells, since the H3-NKp46 interaction leads
to direct NKp46-mediated NK cell killing (4) (Fig. 7A). When
we induced the redirected killing of the uninfected P815 cells
with anti-NKG2D MAb, an efficient killing of the P815 cells
was observed, since NKG2D is a potent killer receptor and its
triggering leads to direct cytotoxicity (Fig. 7A). Importantly,
when we incubated either the H3N2 or the H5N1-infected
P815 cells with anti-NKG2D MAb an increased and equivalent
redirected killing of both infected cells was observed (Fig. 7A).
Thus, when NKp46 interacts with H5, an efficient NK killing of
the H5-infected cells is observed only when the NKG2D re-
ceptor is also engaged (either by MAb as seen in Fig. 7A or
through the interaction with MICB, as seen in Fig. 3E).
Next, we tested the ability of the anti-NKG2D MAb to block
the NKp46-mediated killing of the Jeg3/MICB cells. For this
purpose, we infected the Jeg3/MICB cells with A/PR8 or with
H5 viruses and tested their killing by NK cells that were pre-
incubated with anti-NKG2D or with control MAb. As can be
seen in Fig. 7B, the anti-NKG2D MAb did not block the killing
of the H1-infected Jeg3/MICB cells since the H1-NKp46 in-
teractions are still intact and, as shown above, such interactions
FIG. 4. Glycosylation at amino acids Asp216 and Thr225 of NKp46 are involved in its binding to H5-infected cells. (A) Amino acid sequences
of human NKp46. The first Ig and the second Ig domains are indicated in green and in blue, respectively. The transmembrane domain is
underlined, and the three putative glycosylation sites are indicated in bold red. (B) H5N1 (avian) influenza virus-infected or noninfected Jeg3 cells
were stained with NKp46-Ig (black and blue lines, respectively) and with the indicated mutated proteins (uninfected [purple] and infected [red]).
The figure shows the results of one representative experiment of three performed.
FIG. 5. The binding sites of H1 and H5 to NKp46 are distinct.
Jeg3/MICB cells were infected with A/PR8 or with H5 (avian) viruses,
and their killing by bulk NK cells was tested in the presence or absence
of the indicated anti-NKp46 MAbs. The E:T ratio is 60:1. The figure
shows the results for one representative experiment of three per-
formed. Error bars indicate standard deviations.
VOL. 84, 2010 NK CELL RECOGNITION OF AVIAN AND SWINE INFLUENZA VIRUS3997
lead to direct cytotoxicity. In contrast, blocking of the NKG2D
interaction in the H5-infected Jeg3/MICB cells significantly
reduced the killing, further supporting the observations that
the killing induced by the H5-NKp46 interaction is possible
only when other lysis ligands are engaged (Fig. 7B).
To test whether the NKp46 interactions with H5 and the
H1N1-2009 would be important under physiological condi-
tions, we infected our NCR1gfp/gfp(the mouse homologue of
NKp46 ) knockout mice (KO) and the corresponding
C57BL/6 wild-type (WT) mice with the H5N1 and the H1N1-
2009 viruses. Mice were sacrificed 5 and 6 days after infection
(for H5N1 and H1N1-2009, respectively), and the amounts of
viruses that were present in the lungs were measured as de-
scribed in Materials and Methods. To test the in vivo involve-
ment of NKG2D in influenza virus eradication, we also in-
fected additional groups of mice with both viruses and injected,
prior to the infection, the anti-NKG2D MAb C7 that is known
to block NKG2D function in vivo. As shown in Fig. 7C, in the
absence of NKp46 (KO mice) the presence of both H5N1 and
H1N1-2009 viruses was observed in the lungs of the infected
animals, whereas in the WT mice, few or no viruses were found
(Fig. 7C). Strikingly and in agreement with our in vitro results,
blocking of the NKG2D receptor in the NKp46 KO mice had
no effect on the H1N1-2009 virus titers, probably because for
this virus NKp46 is the major killer receptor involved in the
killing of the infected cells. In contrast, when NKG2D was
blocked prior to H5 infection, a 2-fold increase in the virus titer
was observed, suggesting that NKG2D is involved in vivo in the
killing of the H5 virus. It seems, however, that the NKG2D
activity is secondary to that of NKp46 because in both viruses
when the NKG2D interactions were blocked in vivo in the WT
mice, viruses were still not detected in the lungs (Fig. 7C).
In conclusion, we demonstrate here both in vitro and in vivo
that NKp46 directly interacts with both the H1N1-2009 and the
avian viruses. However, whereas the interaction of NKp46 with
H1 is sufficient to induce direct killing, the NKp46 interaction
with H5 is not sufficient to induce direct killing and requires
the “help” of additional receptors.
Influenza epidemics occur every year, and every several de-
cades an influenza pandemic that claims the lives of millions
arises. Although the world’s greatest concern in the last few
years was an avian H5N1 influenza pandemic, a novel subtype
of influenza virus A (H1N1), leading to what is now considered
to be the first pandemic in the last 41 years and the first in the
21st century. This virus, H1N1-2009, first appeared in Mexico
in March 2009 and in 3 months spread to more than 80 countries
though the WHO considers the severity of this pandemic at
this time to be moderate, the severity might change over time,
as occurred in the “Spanish Flu” pandemic (27).
In contrast to the limited mortality caused by the H1N1-2009,
infection with the avian influenza virus (H5N1) results in a severe
mortality rate of ca. 60% (http://www.who.int/csr/disease/avian
possible explanation for the extreme virulence of the H5N1 virus
might be the preferential binding of the avian H5 to ?2,3-SA
FIG. 6. The H1N1-2009-NKp46 interaction is similar to other human origin influenza virus HAs. (A) Jeg3 cells infected or not infected with
19 different isolations of H1N1-2009 (the isolation number is indicated at the x axis) that were obtained from different Israeli patients. Jeg3 cells
were stained either with NKp46-Ig or with T225V-Ig (white and gray bars, respectively). (B) The cells in panel A were tested for killing by NK
cells in E:T ratio of 25:1. Error bars indicate standard deviations.
3998 ACHDOUT ET AL. J. VIROL.
residues (33), which are located mainly on proteins found at the
human lower respiratory tract (39).
We show here that, in contrast to H1N1-2009 virus and all
other influenza viruses studied thus far (4, 17, 22), the H5-
NKp46 interactions are unable to directly activate the NKp46-
mediated killing by themselves. Indeed, it has been shown that
under certain conditions NKp46 acts as a costimulating recep-
tor and not as a direct cytotoxic receptor. Bryceson et al.
demonstrated that in resting nonactivated NK cells, triggering
of NKp46 is insufficient to activate cytotoxicity or cytokine
secretion, whereas in interleukin-2-activated NK cells, NKp46
could mediate direct killing (9). Indeed, when additional li-
gands of NK activating receptors are present on the H5N1-
infected cells, or in vivo, when other activating ligands such as
the ligands for NKG2D are also upregulated due to the infec-
tion (12, 34; the present study), NKp46 interactions with H5
significantly contribute to virus killing.
We propose that the H5-NKp46 interactions are unable to
induce direct cytotoxicity, probably because H5 binds NKp46
in a site that is distinct from that of H1. We have previously
demonstrated that the glycosylation on the Thr225 residue is
critical for the interactions of H1 with NKp46 (4); we further
show here that the glycosylation at Thr225 is also important for
the interaction of NKp46 with the H1-2009 protein. In con-
trast, the recognition of H5 by NKp46 depends on both the
Asp216 and the Thr225 residues. Interestingly, both residues
are located in the stalk region of NKp46, whereas the third
glycosylated residue is located between domains 1 and 2 and
therefore might not be accessible for binding (13).
The invading viruses and the immune system are in constant
battle. The influenza viruses, similar to other viruses, have
developed several ways to escape from immune system recog-
nition. For example, a high frequency of mutations enables the
virus to constantly change its surface glycoproteins and thereby
FIG. 7. Blocking of NKG2D activity during infection. (A) Redirected killing assay.35S-labeled P815 cells were infected or not with H5 (avian
virus) or H3N2 (human influenza virus) and then assayed in a redirected killing assay with human NK cells at E:T of 3:1 in the presence of either
anti-NKG2D MAb or control IgG1 MAb. (B) Bulk NK cells were preincubated with anti-NKG2D MAb or with anti-CD99 MAb (negative control)
and were tested in killing assays at an E:T ratio of 10:1 against Jeg3/MICB cells that were infected or not infected with A/PR8 or H5N1 influenza
viruses. (C) Anti-NKG2D receptor antibody was injected or not injected into C57BL/6 mice (WT) or C57BL/6 NCR1gfp/gfpmice (KO) 6 h prior
to their intranasal inoculation with 16 TCID50of the H1N1-2009 virus (isolation number 926)/mouse or with 1.6 TCID50of the H5N1 virus/mouse.
Five or six days (for avian and H1N1-2009 infections, respectively) after infection the mice were sacrificed, their lungs were homogenized, RNA
was extracted, and the presence of virus in the infected lungs was detected by specific primers using real-time RT-PCR. The virus titer was
determined based on the viral genome copy numbers.
VOL. 84, 2010 NK CELL RECOGNITION OF AVIAN AND SWINE INFLUENZA VIRUS3999
escape from recognition by cytotoxic T lymphocytes and MAbs
(32). We showed that the influenza virus also induces the
reorganization of major histocompatibility complex (MHC)
class I proteins on the cell surface. MHC class I proteins enter
into rafts early after infection to enhance the recognition by
NK inhibitory receptors (1, 2). On the other hand, we suggest
that NKp46 has evolved to use the conserved ability of various
human HAs to bind ?2,6-SA and “utilize” this property to
directly kill the infected cells in an ?2,6-SA-NKp46-dependent
manner (4). This is probably why the swine influenza virus,
which interacts with ?2,6-SA (19), is directly killed in an
NKp46-dependent manner. Unfortunately, due to the lack of
specific blocking MAbs, we could not directly demonstrate
NKp46-dependent killing of the swine influenza viruses. How-
ever, the direct correlation observed between the binding of
NKp46 and the killing of the H1N1-2009 influenza virus-in-
fected cells, together with the impressive enhancement of viral
load observed in the absence of NKp46, strongly suggests that
similar to other human influenza viruses (4), the interaction
between NKp46 and the swine H1 virus is crucial for the killing
of the infected cells.
With regard to the avian influenza virus, we suggest that
NKp46 did not evolve to interact with avian HA, and therefore
the NKp46 interactions with H5 through its ?2,3-SA are insuf-
ficient to activate the NKp46-mediated killing of NK cells. Our
final conclusion is that it might be that the avian H5N1 virus is
so deadly not only due to its ability to infect the lower respi-
ratory track but perhaps also due to its inability to directly
activate NKp46. It is therefore possible that if genetic changes
in the H5N1 virus would occur to enable its transmission from
human to human, the emerging virus would not be more vio-
lent but may be less dangerous, since the infected cells would
be recognized directly by NKp46 and the virus might be better
eliminated, in a similar manner to that of the H1N1-2009
This study was supported by grants from the U.S.-Israel Binational
Science Foundation, the Israeli Cancer Research Foundation, the Is-
raeli Science Foundation, the Israel Science Foundation (Morasha),
the Association for International Cancer Research, the DKFZ-MOST,
and an Israel-Croatia research grant (all to O.M.). O.M. is a Crown
Professor of Molecular Immunology.
1. Achdout, H., T. I. Arnon, G. Markel, T. Gonen-Gross, G. Katz, N. Lieber-
man, R. Gazit, A. Joseph, E. Kedar, and O. Mandelboim. 2003. Enhanced
recognition of human NK receptors after influenza virus infection. J. Immu-
2. Achdout, H., I. Manaster, and O. Mandelboim. 2008. Influenza virus infec-
tion augments NK cell inhibition through reorganization of major histocom-
patibility complex class I proteins. J. Virol. 82:8030–8037.
3. Arnon, T. I., H. Achdout, O. Levi, G. Markel, N. Saleh, G. Katz, R. Gazit, T.
Gonen-Gross, J. Hanna, E. Nahari, A. Porgador, A. Honigman, B. Plachter,
D. Mevorach, D. G. Wolf, and O. Mandelboim. 2005. Inhibition of the
NKp30 activating receptor by pp65 of human cytomegalovirus. Nat. Immu-
4. Arnon, T. I., H. Achdout, N. Lieberman, R. Gazit, T. Gonen-Gross, G. Katz,
A. Bar-Ilan, N. Bloushtain, M. Lev, A. Joseph, E. Kedar, A. Porgador, and
O. Mandelboim. 2004. The mechanisms controlling the recognition of tu-
mor- and virus-infected cells by NKp46. Blood 103:664–672.
5. Arnon, T. I., M. Lev, G. Katz, Y. Chernobrov, A. Porgador, and O. Mandel-
boim. 2001. Recognition of viral hemagglutinins by NKp44 but not by
NKp30. Eur. J. Immunol. 31:2680–2689.
6. Arnon, T. I., G. Markel, and O. Mandelboim. 2006. Tumor and viral recog-
nition by natural killer cells receptors. Semin. Cancer Biol. 16:348–358.
7. Biassoni, R. 2008. Natural killer cell receptors. Adv. Exp. Med. Biol. 640:
8. Brandt, C. S., M. Baratin, E. C. Yi, J. Kennedy, Z. Gao, B. Fox, B. Haldeman,
C. D. Ostrander, T. Kaifu, C. Chabannon, A. Moretta, R. West, W. Xu, E.
Vivier, and S. D. Levin. 2009. The B7 family member B7-H6 is a tumor cell
ligand for the activating natural killer cell receptor NKp30 in humans. J. Exp.
9. Bryceson, Y. T., M. E. March, H. G. Ljunggren, and E. O. Long. 2006.
Synergy among receptors on resting NK cells for the activation of natural
cytotoxicity and cytokine secretion. Blood 107:159–166.
10. Cerwenka, A., and L. L. Lanier. 2001. Natural killer cells, viruses and cancer.
Nat. Rev. Immunol. 1:41–49.
11. Chen, Y., B. Perussia, and K. S. Campbell. 2007. Prostaglandin D2 sup-
presses human NK cell function via signaling through D prostanoid receptor.
J. Immunol. 179:2766–2773.
12. Draghi, M., A. Pashine, B. Sanjanwala, K. Gendzekhadze, C. Cantoni, D.
Cosman, A. Moretta, N. M. Valiante, and P. Parham. 2007. NKp46 and
NKG2D recognition of infected dendritic cells is necessary for NK cell
activation in the human response to influenza infection. J. Immunol. 178:
13. Foster, C. E., M. Colonna, and P. D. Sun. 2003. Crystal structure of the
human natural killer (NK) cell activating receptor NKp46 reveals structural
relationship to other leukocyte receptor complex immunoreceptors. J. Biol.
14. Gazit, R., R. Gruda, M. Elboim, T. I. Arnon, G. Katz, H. Achdout, J. Hanna,
U. Qimron, G. Landau, E. Greenbaum, Z. Zakay-Rones, A. Porgador, and O.
Mandelboim. 2006. Lethal influenza infection in the absence of the natural
killer cell receptor gene Ncr1. Nat. Immunol. 7:517–523.
15. Hindiyeh, M., V. Levy, R. Azar, N. Varsano, L. Regev, Y. Shalev, Z. Gross-
man, and E. Mendelson. 2005. Evaluation of a multiplex real-time reverse
transcriptase PCR assay for detection and differentiation of influenza viruses
A and B during the 2001–2002 influenza season in Israel. J. Clin. Microbiol.
16. Horimoto, T., and Y. Kawaoka. 2006. Strategies for developing vaccines
against H5N1 influenza A viruses. Trends Mol. Med. 12:506–514.
17. Jarahian, M., C. Watzl, P. Fournier, A. Arnold, D. Djandji, S. Zahedi, A.
Cerwenka, A. Paschen, V. Schirrmacher, and F. Momburg. 2009. Activation
of natural killer cells by Newcastle disease virus hemagglutinin-neuramini-
dase. J. Virol. 69:8420–8428.
18. Kilbourne, E. D. 2006. Influenza pandemics of the 20th century. Emerg.
Infect. Dis. 12:9–14.
19. Maines, T. R., A. Jayaraman, J. A. Belser, D. A. Wadford, C. Pappas, H.
Zeng, K. M. Gustin, M. B. Pearce, K. Viswanathan, Z. H. Shriver, R. Raman,
N. J. Cox, R. Sasisekharan, J. M. Katz, and T. M. Tumpey. 2009. Transmis-
sion and pathogenesis of swine-origin 2009 A(H1N1) influenza viruses in
ferrets and mice. Science 24:484–487.
20. Maines, T. R., K. J. Szretter, L. Perrone, J. A. Belser, R. A. Bright, H. Zeng,
T. M. Tumpey, and J. M. Katz. 2008. Pathogenesis of emerging avian influ-
enza viruses in mammals and the host innate immune response. Immunol.
21. Manaster, I., and O. Mandelboim. 2008. The unique properties of human
NK cells in the uterine mucosa. Placenta 29(Suppl. A):S60–S66.
22. Mandelboim, O., N. Lieberman, M. Lev, L. Paul, T. I. Arnon, Y. Bushkin,
D. M. Davis, J. L. Strominger, J. W. Yewdell, and A. Porgador. 2001.
Recognition of haemagglutinins on virus-infected cells by NKp46 activates
lysis by human NK cells. Nature 409:1055–1060.
23. Mandelboim, O., P. Malik, D. M. Davis, C. H. Jo, J. E. Boyson, and J. L.
Strominger. 1999. Human CD16 as a lysis receptor mediating direct natural
killer cell cytotoxicity. Proc. Natl. Acad. Sci. U. S. A. 96:5640–5644.
24. Mandelboim, O., H. T. Reyburn, M. Vales-Gomez, L. Pazmany, M. Colonna,
G. Borsellino, and J. L. Strominger. 1996. Protection from lysis by natural
killer cells of group 1 and 2 specificity is mediated by residue 80 in human
histocompatibility leukocyte antigen C alleles and also occurs with empty
major histocompatibility complex molecules. J. Exp. Med. 184:913–922.
25. Moretta, A., C. Bottino, M. Vitale, D. Pende, C. Cantoni, M. C. Mingari, R.
Biassoni, and L. Moretta. 2001. Activating receptors and coreceptors in-
volved in human natural killer cell-mediated cytolysis. Annu. Rev. Immunol.
26. Munster, V. J., E. de Wit, J. M. van den Brand, S. Herfst, E. J. Schrauwen,
T. M. Bestebroer, D. van de Vijver, C. A. Boucher, M. Koopmans, G. F.
Rimmelzwaan, T. Kuiken, A. D. Osterhaus, and R. A. Fouchier. 2009. Patho-
genesis and transmission of swine-origin 2009 A(H1N1) influenza virus in
ferrets. Science 24:971–976.
27. Neumann, G., T. Noda, and Y. Kawaoka. 2009. Emergence and pandemic
potential of swine-origin H1N1 influenza virus. Nature 459:931–939.
28. Pappaioanou, M. 2009. Highly pathogenic H5N1 avian influenza virus: cause
of the next pandemic? Comp. Immunol. Microbiol. Infect. Dis. 32:287–300.
29. Pessino, A., S. Sivori, C. Bottino, A. Malaspina, L. Morelli, L. Moretta, R.
Biassoni, and A. Moretta. 1998. Molecular cloning of NKp46: a novel mem-
ber of the immunoglobulin superfamily involved in triggering of natural
cytotoxicity. J. Exp. Med. 188:953–960.
30. Pogge von Strandmann, E., V. R. Simhadri, B. von Tresckow, S. Sasse, K. S.
4000ACHDOUT ET AL.J. VIROL.
Reiners, H. P. Hansen, A. Rothe, B. Boll, V. L. Simhadri, P. Borchmann, Download full-text
P. J. McKinnon, M. Hallek, and A. Engert. 2007. Human leukocyte antigen-
B-associated transcript 3 is released from tumor cells and engages the
NKp30 receptor on natural killer cells. Immunity 27:965–974.
31. Raulet, D. H. 2003. Roles of the NKG2D immunoreceptor and its ligands.
Nat. Rev. Immunol. 3:781–790.
32. Scholtissek, C. 1994. Source for influenza pandemics. Eur. J. Epidemiol.
33. Shinya, K., M. Ebina, S. Yamada, M. Ono, N. Kasai, and Y. Kawaoka. 2006.
Avian flu: influenza virus receptors in the human airway. Nature 440:435–
34. Siren, J., T. Sareneva, J. Pirhonen, M. Strengell, V. Veckman, I. Julkunen,
and S. Matikainen. 2004. Cytokine and contact-dependent activation of
natural killer cells by influenza A or Sendai virus-infected macrophages.
J. Gen. Virol. 85:2357–2364.
35. Stern-Ginossar, N., N. Elefant, A. Zimmermann, D. G. Wolf, N. Saleh, M.
Biton, E. Horwitz, Z. Prokocimer, M. Prichard, G. Hahn, D. Goldman-Wohl,
C. Greenfield, S. Yagel, H. Hengel, Y. Altuvia, H. Margalit, and O. Mandel-
boim. 2007. Host immune system gene targeting by a viral miRNA. Science
36. Stevens, J., O. Blixt, L. M. Chen, R. O. Donis, J. C. Paulson, and I. A.
Wilson. 2008. Recent avian H5N1 viruses exhibit increased propensity for
acquiring human receptor specificity. J. Mol. Biol. 381:1382–1394.
37. Stevens, J., O. Blixt, T. M. Tumpey, J. K. Taubenberger, J. C. Paulson, and
I. A. Wilson. 2006. Structure and receptor specificity of the hemagglutinin
from an H5N1 influenza virus. Science 312:404–410.
38. Sutherland, C. L., N. J. Chalupny, and D. Cosman. 2001. The UL16-binding
proteins, a novel family of MHC class I-related ligands for NKG2D, activate
natural killer cell functions. Immunol. Rev. 181:185–192.
39. van Riel, D., V. J. Munster, E. de Wit, G. F. Rimmelzwaan, R. A. Fouchier,
A. D. Osterhaus, and T. Kuiken. 2006. H5N1 virus attachment to lower
respiratory tract. Science 312:399.
40. Yokoyama, W. M., S. Kim, and A. R. French. 2004. The dynamic life of
natural killer cells. Annu. Rev. Immunol. 22:405–429.
VOL. 84, 2010 NK CELL RECOGNITION OF AVIAN AND SWINE INFLUENZA VIRUS4001