﻿M A J O R A R T I C L E
Differences in Cytokine Production in Human
Macrophages and in Virulence in Mice
Are Attributable to the Acidic Polymerase
Protein of Highly Pathogenic Influenza
A Virus Subtype H5N1
Saori Sakabe,1,5
Ryo Takano,1
Tokiko Nagamura-Inoue,2
Naohide Yamashita,3
Chairul A. Nidom,6
Mai thi Quynh Le,7
Kiyoko Iwatsuki-Horimoto,1
and Yoshihiro Kawaoka1,4,5,8
1
Division of Virology, Department of Microbiology and Immunology, 2
Department of Cell Processing and Transfusion, Research Hospital, 3
Department
of Advanced Medical Science, and 4
International Research Center for Infectious Diseases, Institute of Medical Science, University of Tokyo, and
5
ERATO Infection-Induced Host Responses Project, Saitama, Japan; 6
Faculty of Veterinary Medicine, Tropical Disease Centre, Airlangga University,
Surabaya, Indonesia; 7
National Institute of Hygiene and Epidemiology, Hanoi, Vietnam; and 8
Department of Pathobiological Sciences, School of
Veterinary Medicine, University of Wisconsin-Madison
(See the editorial commentary by Donis and Cox, on pages 208­10 and editorial commentary Hirsch, on page 207.)
Background. The pathogenesis of influenza A virus subtype H5N1 (hereafter, "H5N1") infection in humans
is not completely understood, although hypercytokinemia is thought to play a role. We previously reported that
most H5N1 viruses induce high cytokine responses in human macrophages, whereas some H5N1 viruses induce
only a low level of cytokine production similar to that induced by seasonal viruses.
Methods. To identify the viral molecular determinants for cytokine induction of H5N1 viruses in human
macrophages, we generated a series of reassortant viruses between the high cytokine inducer A/Vietnam/
UT3028II/03 clone 2 (VN3028IIcl2) and the low inducer A/Indonesia/UT3006/05 (IDN3006) and evaluated cyto-
kine expression in human macrophages.
Results. Viruses possessing the acidic polymerase (PA) gene of VN3028IIcl2 exhibited high levels of hyper-
cytokinemia-related cytokine expression in human macrophages, compared with IDN3006, but showed no sub-
stantial differences in viral growth in these cells. Further, the PA gene of VN3028IIcl2 conferred enhanced
virulence in mice.
Conclusions. These results demonstrate that the PA gene of VN3028IIcl2 affects cytokine production in
human macrophages and virulence in mice. These findings provide new insights into the cytokine-mediated path-
ogenesis of H5N1 infection in humans.
Keywords. H5N1 virus; Hypercytokinemia; cytokine; human macrophages; virulence; Microarray; Immune response.
Highly pathogenic avian influenza A virus subtype
H5N1 (hereafter, "H5N1") have been circulating among
wild birds and domestic poultry worldwide since 2003
[1, 2]. The total number of laboratory-confirmed cases
of human H5N1 infection has climbed to >600 cases,
with a mortality rate of 60%, posing a true threat to
human health [3]. Although the high pathogenicity of
these viruses is not completely understood, hypercyto-
kinemia and systemic viral replication have been report-
ed for patients infected with these viruses [4­9].
Alveolar macrophages are considered central players
in both innate and adaptive immune responses to re-
spiratory infection and can be infected with H5N1
Received 7 March 2012; accepted 14 June 2012; electronically published 4
October 2012.
Correspondence: Yoshihiro Kawaoka, DVM, PhD, Institute of Medical Science,
University of Tokyo, Shirokanedai 4-6-1, Minato-ku, Tokyo 108-8639, Japan
(kawaoka@ims.u-tokyo.ac.jp).
The Journal of Infectious Diseases 2013;207:262­71
© The Author 2012. Published by Oxford University Press on behalf of the Infectious
Diseases Society of America. All rights reserved. For Permissions, please e-mail:
journals.permissions@oup.com.
DOI: 10.1093/infdis/jis523
262 · JID 2013:207 (15 January) · Sakabe et al
viruses [10]. In in vitro studies, human monocyte­derived
macrophages infected with H5N1 viruses produced higher
levels of cytokines than those infected with seasonal viruses,
suggesting cytokine production contributes to the high patho-
genicity of H5N1 viruses in humans [1, 11­15]. However, we
previously found that some H5N1 viruses induce a level of
cytokine production similar to that induced by seasonal viruses
in human monocyte­derived macrophages [16], indicating that
the ability to induce cytokines is virus-strain dependent.
Previous findings regarding the role of nonstructural
protein 1 (NS1) in cytokine induction are controversial; in one
study, NS1 was shown to affect cytokine production in human
macrophages [13], whereas another study was unable to
confirm NS1's contribution to cytokine induction in these
cells [17]. PB2-E627K, which is widely known as a mutation
that enhances viral growth in mammalian cells and some
mammals, is reported to enhance cytokine production [13,
18]. Thus, determining the potential molecular markers that
enhance cytokine production could extend our knowledge of
the basis for the high virulence of H5N1 viruses in humans.
We previously evaluated the ability of several H5N1 viruses
to induce cytokines in human macrophages and found that A/
Vietnam/UT3028II/03 clone 2 (VN3028IIcl2) was a high cyto-
kine inducer and A/Indonesia/UT3006/05 (IDN3006) a low
cytokine inducer [16]. Here, we generated a series of reassor-
tant viruses between these 2 viruses and attempted to identify
the viral molecular determinants that confer high cytokine in-
duction in human macrophages. Further, to reveal the molec-
ular networks affected by the virus infection, we examined the
entire gene transcriptional profiles in human macrophages by
using microarray analysis. To clarify whether differential ex-
pression of cytokines mediated by H5N1 viruses contributes
to pathogenicity in mammals, we compared the virulence in
mice infected with the high cytokine inducer to that in mice
infected with the low cytokine inducer.
MATERIALS AND METHODS
Ethics Statement
Our research protocol for the use of human-derived macro-
phages was approved by the Research Ethics Review Committee
of the Institute of Medical Science, the University of Tokyo (ap-
proval numbers 18-15-0129 and 19-24-200430). Our research
protocol for the use of mice followed the University of Tokyo's
Regulations for Animal Care and Use, which was approved by
the Animal Experiment Committee of the Institute of Medical
Science, the University of Tokyo (approved numbers 19­29).
Cells
A549 and HEK293T cells were maintained in Dulbecco's
modified Eagle's medium (DMEM) with 10% fetal calf serum.
Madin-Darby canine kidney (MDCK) cells were maintained
in minimal essential medium supplemented with 5% newborn
calf serum. Human monocyte­derived macrophages were pre-
pared as described previously [16]. Briefly, peripheral blood
mononuclear cells were independently separated from the
buffy coat of healthy donors. Monocytes were purified by ad-
herence and were allowed to differentiate for 14 days in
Roswell Park Memorial Institute (RPMI) 1640 medium sup-
plemented with 5 ng/mL of recombinant human granulocyte-
macrophage colony-stimulating factor and 10% human serum
derived from the corresponding blood donors. All cells were
cultured at 37°C in 5% CO2.
Viruses
In this study, we used 2 H5N1 viruses with different
cytokine induction abilities: A/Vietnam/UT3028II/03 clone 2
(VN3028IIcl2, clade 1) and A/Indonesia/UT3006/05 (IDN3006,
clade 2.1.3). Viruses were isolated in MDCK cells, and virus
stocks were prepared from the supernatants of MDCK cells.
Plasmid Constructs and Reverse Genetics
Reverse genetics systems for the VN3028IIcl2 and IDN3006
viruses were established as previously described [19]. Briefly,
the complementary DNA of genes from these viruses were
cloned into the pHH21 vector. The 8 plasmids for the synthe-
sis of viral RNA and 4 expression plasmids for IDN3006 or
VN3028IIcl2 virus-derived basic polymerase 2 (PB2), basic
polymerase 1 (PB1), acidic polymerase (PA), and nucleopro-
tein (NP) proteins were transfected into HEK293T cells. Two
days later, the supernatant of the transfected cells was harvest-
ed and used to infect MDCK cells for the preparation of virus
stocks. All virus stocks were sequenced to ensure the absence
of unwanted mutations and were titrated by using the plaque
assay with MDCK cells prior to use.
Virus Infection of Human Macrophages
Differentiated macrophages were seeded on 24-well plates (BD
Falcon), with 1 × 105
cells per well. Cells were infected with
viruses at a multiplicity of infection of 2. The virus inoculum
was removed, and the cells were washed 3 times and then in-
cubated in serum-free RPMI 1640 medium supplemented
with 0.3% bovine serum albumin. At 6, 12, and 24 hours after
infection, the supernatant was collected for virus titration and
cytokine measurement; the cells were collected for use in viral
gene and protein expression assays.
Virus Infection of Human Pulmonary Epithelial Cells
A549 cells were seeded on 6-well plates, with 7.5 × 105
cells
per well. Cells were infected with viruses at a multiplicity of
infection of 0.0001. The virus inoculum was removed, and the
cells were washed 3 times and then incubated in serum-free
DMEM supplemented with 0.3% bovine serum albumin. At 6,
12, 24, 48, 72, 96, and 144 hours after infection, viruses in the
Cytokine Response in Human Macrophages · JID 2013:207 (15 January) · 263
supernatants were collected and titrated by means of plaque
assays in MDCK cells.
Cytokine Measurements
The concentrations of 7 cytokines (interleukin-6 [IL-6],
CXCL10 [IP-10], CCL3 [MIP-1], CCL4 [MIP-1], CCL5
[RANTES], tumor necrosis factor  [TNF-], and CXCL9
[MIG]) in supernatants from human macrophages infected
with H5N1 viruses were determined by using the Bio-Plex
human x-plex (Bio-Rad Laboratories) and by using the Bio-
Plex protein array system (Bio-Rad Laboratories) according to
the manufacturer's instructions.
RNA Isolation and Integrity
Infected cells were lysed with TRIzol Reagent (Invitrogen),
and chloroform-isoamyl alcohol (5:1) was added. The cells
were then vortexed and centrifuged for 15 min at 12 000 × g at
4°C. The resulting aqueous layer was subjected to RNA extrac-
tion by using RNeasy Mini kit columns (Qiagen) according to
the manufacturer's instructions. Isolated total RNA integrity
was assessed by determining UV 260/280 absorbance ratios
and by examining 28S/18S ribosomal RNA bands with an
Agilent 2100 bioanalyzer (Agilent Technologies) according to
the manufacturer's instructions. Only samples with high-
quality RNA (ie, those with RNA integrity numbers >9.0)
were used for microarray analysis.
Microarray Analysis and Bioinformatics
Cy3-labeled complementary RNA probe synthesis was initiat-
ed with 50 ng of total RNA by using the Agilent Low Input
Quick Amp Labeling kit, one color (Agilent Technologies).
The Agilent SurePrint G3 Human GE 8 × 60 K microarrays
(G4851) were used according to the manufacturer's instruc-
tions. Slides were scanned with an Agilent's High-Resolution
Microarray Scanner, and image data were processed by using
Agilent Feature Extraction software, version 10.7.3.1. All data
were subsequently uploaded into GeneSpring GX, version
11.5, for data analysis. In accordance with proposed MIAME
(minimum information about a microarray experiment) stan-
dards, microarray data gained in this study are publically
available at the Gene Expression Omnibus (GEO) database
(http:www.ncbi.nlm.nih.gov/geo/) under the accession number
GSE40711. For the microarray data analysis, each gene expres-
sion array data set was normalized to the in silico pool for
mock-infected cells (n = 3).
Statistically significant differences in gene expression between
IDN3006 and VN3028IIcl2 and between IDN3006 and
IDN3006/cl2PA were determined by using 1-way analysis of
variance (ANOVA) (P < .05) with the Tukey honestly significant
difference (HSD) post hoc test and the Benjamini-Hochberg
false discovery rate correction. Differentially expressed genes
were further filtered to include genes whose expression changed
at least 2.0-fold relative to the level in the IDN3006-infected
group. Only the genes whose expression showed at least 2.0-
fold changes and statistical significance (P < .05) in 2 pairs of
comparisons involving IDN3006 versus VN3028IIcl2 and
IDN3006 versus IDN3006/cl2PA were assigned to a Gene On-
tology group or were uploaded into Ingenuity Pathway Analysis
(IPA; Ingenuity Systems) to identify the functions and pathways
that were enriched in the gene set. To identify the gene network
that enriched the differentially expressed genes, genes were con-
nected in a de novo network based on known interactions in
IPA. Genes that were not directly or indirectly connectable to
each other were not analyzed further. For all analysis, P values
were calculated by using the Fisher exact test to identify biolog-
ical functions and pathways that were significant (P < .05).
Quantitative Real-Time Polymerase Chain Reaction (PCR)
Analysis
Quantitative real-time PCR was performed to determine the
expression levels of viral PA, hemagglutinin (HA), NP, matrix
(M), and NS genes. RNA was reverse-transcribed with oligo-
nucleotide dT and SuperScript III reverse transcriptase (Invi-
trogen). Quantitative real-time PCR was conducted with the
SYBR Green PCR master mix (Invitrogen) and performed on
the ABI7900HT Fast Real-Time PCR System platform. Target
gene messenger RNA (mRNA) levels were normalized to 18S
human ribosomal RNA according to the 2-Ct
calculation
[20]. Primer sequences are available on request.
Western Blotting
Infected cells were lysed with 2 × sample buffer (Invitrogen),
and samples were loaded onto precast gels for 4%­20% sodium
dodecyl sulfate­polyacrylamide gel electrophoresis (Bio-Rad
Laboratories). Proteins were transferred electrophoretically to
polyvinylidene fluoride membrane (Millipore) in transfer buffer
(100 mM Tris, 190 mM glycine, and 10% methanol). The
membranes were blocked for 1 hour at room temperature with
Blocking One (Nacalai Tesque) and were then incubated with
primary antibodies for 1 hour at room temperature. Anti-PA
and anti-NS1 mouse monoclonal antibodies (1:10 000) and
anti-H5N1 rabbit polyclonal antibodies (1:10 000) for detection
of HA, NP, and M1 were used. After being washed 3 times with
phosphate-buffered saline containing 0.05% Tween-20 (PBS-T),
the membranes were incubated with anti-mouse (1:10 000) or
anti-rabbit (1:10 000) secondary antibodies conjugated with
horse radish peroxidase (GE Healthcare) for 1 hour at room
temperature. After 3 more washes with PBS-T, specific proteins
were detected by using the ECL Prime Western Blotting Detec-
tion System (GE Healthcare). Photography was conducted with
the VersaDoc Imaging System (Bio-Rad laboratories).
Minigenome Assay
Human macrophages or A549 cells were transfected with the
expression plasmids for PB2, PB1, PA, and NP (0.25 g each),
264 · JID 2013:207 (15 January) · Sakabe et al
pPoll-NP(0)luc2(0) [21], and ptkRluc (0.025 g each) by using
the jetPEI-Macrophage DNA transfection reagent (Polyplus
transfection) or the Lipofectamine LTX and Plus Reagent (Invi-
trogen), respectively. The cells were then harvested 48 or 24
hours after transfection, respectively. Luciferase activity was as-
sessed by using a dual-luciferase reporter assay system (Promega)
according to the manufacturer's instructions. Photinus luciferase
activity was standardized against Renilla luciferase activity.
Animal Experiments
Female BALB/c mice aged 5­6 weeks ( Japan SLC) were used
in the study. To investigate viral pathogenicity in mice, 4 mice
per group were anaesthetized with isoflurane and intranasally
infected with 50 L of serial dilutions of viruses, creating
doses ranging from 100
to 103
plaque-forming units (PFU).
Mice were monitored daily for morbidity and mortality for up
to 14 days after infection. The percentage changes in body
weights were calculated by comparing the weight of each
mouse at each time point to its weight on day 0. The 50%
mouse lethal dose for each virus was calculated by using the
Reed-Muench method [22]. To determine virus growth in
mouse organs, 9 mice per group were intranasally infected
with 102
PFU of IDN3006 or IDN3006/cl2PA. Three mice per
group were euthanized on days 2, 4, and 7 after infection, and
lungs, nasal turbinates, livers, spleens, kidneys, intestines, and
brains were collected. Virus in organs was titrated by using
plaque assays in MDCK cells.
Statistical Analysis
Statistically significant differences in cytokine expression, virus
growth, luciferase activity, and viral mRNA expression were
determined by using 1- or 2-way ANOVA with the post hoc
Tukey HSD test. In using 2-way ANOVA, we first confirmed
that there were no interaction effects between 2 factors, and
then we determined whether there were statistically significant
differences in virus growth or viral mRNA levels among the
viruses during infection. Statistically significant differences in
mouse survival were determined by using the log-rank test.
For all statistical tests, values were considered to be signifi-
cantly different when the P value was <.05.
Accession Numbers
The viral nucleotide sequences described in this study are
available from GenBank under the following accession
numbers: HM114456 for the PA gene of VN3028IIcl2 and
JX235398 for IDN3006.
RESULTS
Contribution of the PA Gene to Cytokine Induction in Human
Macrophages
To determine the molecular basis for the induction of cyto-
kines in human macrophages, we used 2 H5N1 viruses that
differed in their ability to induce cytokines in human macro-
phages. We first generated wild-type VN3028IIcl2 (high cyto-
kine inducer) and IDN3006 (low cytokine inducer) by using
reverse genetics and confirmed the difference in their abilities
to induce cytokines by infecting human monocyte­derived
macrophages with them and measuring the expression
levels of 7 cytokines (IL-6, IP-10, MIP-1, MIP-1, RANTES,
TNF-, and MIG; Figure 1). VN3028IIcl2 induced signifi-
cantly higher expression levels of all of the cytokines tested
than did IDN3006 in human macrophages. To identify the
gene segments responsible for the difference in cytokine in-
duction between these 2 viruses, we generated 2 reassortant
viruses: VN3028IIcl2(3P + NP), which possesses the polymer-
ase complex genes (PB2, PB1, and PA) and the NP gene from
VN3028IIcl2 and its remaining genes from IDN3006, and
VN3028IIcl2(HA + NA + M + NS), which possesses the HA,
neuraminidase (NA), M, and NS genes from VN3028IIcl2 and
its remaining genes from IDN3006. VN3028IIcl2(3P + NP)
induced statistically significantly higher expression levels of all
of cytokines tested than did IDN3006, whereas VN3028IIcl2
(HA + NA + M + NS) did not (Figure 1A). These results indi-
cate that the PB2, PB1, PA, and NP genes contribute to the
high cytokine induction property of the VN3028IIcl2 virus in
human macrophages.
To identify the gene segments responsible for the high cyto-
kine induction among those that encode the proteins in the ri-
bonucleoprotein complex (ie. PB2, PB1, PA, and NP), we
generated a series of reassortants possessing a single gene
segment derived from VN3028IIcl2 in the IDN3006 backbone
(eg, IDN3006/cl2PB2 indicates a virus that possesses the
PB2 gene from VN3028IIcl2 and its remaining 7 genes from
IDN3006) and evaluated the cytokine expression levels
(Figure 1B). The reassortants that possessed the single PB2,
PB1, or NP gene from VN3028IIcl2 did not induce cytokine
levels similar to those produced by the parental VN3028IIcl2.
However, the cytokine expression induced by IDN3006/cl2PA
was as high as that induced by VN3028IIcl2; the expression
levels of RANTES and TNF- induced by IDN3006/cl2PA were
even statistically significantly higher than those induced by
VN3028IIcl2. Thus, the PA gene of VN3028IIcl2 was responsi-
ble for the high cytokine induction in human macrophages.
Evaluation of the Effect of Viral Replicative Properties on
Cytokine Induction
The PA subunit plays a role in virus replication in infected
cells [23, 24]. To evaluate whether the high cytokine induction
phenotype of VN3028IIcl2 is linked to high viral replication
in infected cells, as observed with viruses possessing the PB2-
E627K mutation [13], we compared the replication in human
macrophages of IDN3006, VN3028IIcl2, and IDN3006/cl2PA
(Figure 2) by determining virus titers, polymerase activity,
viral protein (PA, HA, NP, M1, and NS1) expression, and viral
Cytokine Response in Human Macrophages · JID 2013:207 (15 January) · 265
Figure 1. Cytokine production in human macrophages. Human monocyte­derived macrophages were infected with a series of reassortants between
IDN3006 and VN3028IIcl2 at a multiplicity of infection of 2. At 12 hours after infection, the supernatants of the infected cells were collected, and the
concentrations of 7 cytokines (interleukin-6 [IL-6], CXCL10 [IP-10], CCL3 [MIP-1], CCL4 [MIP-1], CCL5 [RANTES], tumor necrosis factor  [TNF-], and
CXCL9 [MIG]) were measured. All values were normalized to the value of IDN3006. A, Contributions of a set of 4 viral genes to the cytokine production
were evaluated. B, Single viral gene segments responsible for the cytokine induction were determined. The values are means ± SD (n = 3). *P < .05,
compared with the IDN3006 group (1-way analysis of variance with the post hoc Tukey honestly significant difference test). Abbreviations: HA, hemag-
glutinin; M, matrix; NA, neuraminidase; NP, nucleoprotein; NS, nonstructural.
266 · JID 2013:207 (15 January) · Sakabe et al
mRNA gene (PA, HA, NP, M, and NS) expression. Among the
3 viruses tested, there were no statistically significant differences
in the virus titers in the supernatants of infected cells, al-
though IDN3006/cl2PA replicated slightly better than IDN3006
or VN3028IIcl2 at 6 hours after infection (Figure 2A). Al-
though VN3028IIcl2 showed statistically significantly higher
polymerase activity as compared to IDN3006 (Figure 2B),
IDN3006/cl2PA exhibited polymerase activity similar to that
of IDN3006, suggesting that the PA gene of VN3028IIcl2 did
not substantially affect the polymerase activity. While only PA
gene expression was statistically significantly higher in the
IDN3006-infected cells as compared to cells infected with
VN3028IIcl2 or IDN3006/cl2PA, there were no substantial dif-
ferences in the expression levels of other viral proteins or
genes tested among the 3 viruses (Figures 2C and 2D). Thus,
we found no appreciable differences in the viral replicative
properties of these 3 viruses even though they differed in their
cytokine-inducing abilities in human macrophages. These
results indicate that the high cytokine induction by VN3028IIcl2
was not related to efficient virus growth in human macrophages.
The Role of the PA Gene of VN3028IIcl2 in the Differential
Regulation of Host Genes Involved in Hypercytokinemia
To gain further insights into how the PA gene of VN3028IIcl2
affects cytokine production in human macrophages, we examined
all of the gene expression levels in the cells infected with
IDN3006, VN3028IIcl2, and IDN3006/cl2PA. The gene ex-
pression of IP-10, MIP-1, RANTES, TNF-, and MIG con-
firmed the differential upregulation at the transcriptional level
in cells infected with VN3028IIcl2 or IDN3006/cl2PA, com-
pared with those infected with IDN3006 that was observed at
the protein level, although the differential upregulation of IL-6
and MIP-1 was confirmed only in cells infected with
IDN3006/cl2PA, compared with IDN3006 (Figures 1 and 3A).
Microarray analyses further identified 129 genes that were stat-
istically significantly differentially expressed in the human
macrophages infected with VN3028IIcl2 and IDN3006/cl2PA
as compared to IDN3006 (Supplementary Table 1). Gene On-
tology analysis revealed that these genes were enriched with
cytokine-related genes (Figure 3B). IPA network analyses iden-
tified a gene network composed of 48 genes that were signifi-
cantly differentially expressed in the macrophages infected
with VN3028IIcl2 and IDN3006/cl2PA as compared to
IDN3006 (Figure 3C). Of note, this network was enriched
with genes related to hypercytokinemia, represented by
IFNA4, IFNA8, IFNA14, IFNB1, IL12A, IL29, and CXCL10.
These results indicate that the PA amino acid differences are
responsible for the differential expression of the various cyto-
kines related to hypercytokinemia at the gene transcriptional
level in human macrophages.
Figure 2. Viral growth properties in human macrophages. Human monocyte­derived macrophages were infected with IDN3006, VN3028IIcl2, or
IDN3006/cl2PA at a multiplicity of infection of 2. A, At 6, 12, and 24 hours after infection, the virus titers in the supernatants were determined by means
of plaque assays in Madin-Darby canine kidney cells. B, The 4 viral protein expression plasmids (basic polymerase 2 [PB2], basic polymerase 1 [PB1],
acidic polymerase [PA], and nucleoprotein [NP]) of IDN3006, VN3028IIcl2, or IND3006/cl2PA together with pPollNP(0)luc2(0) for the production of virus-
like RNA encoding the reporter luciferase gene were transfected into human monocyte­derived macrophages and assayed for luciferase activities 48
hours after transfection at 37°C. The values were standardized to the value for the ribonucleoprotein complex protein activity of IDN3006. C and D, Viral
proteins (PA, hemagglutinin [HA], NP, matrix [M], matrix 1 [M1], nonstructural [NS], and nonstructural 1 [NS1]; C) and messenger RNA (mRNA; PA, HA,
NP, M, and NS; D) expression levels were measured by means of Western blotting and quantitative real-time polymerase chain reaction, respectively.
The values are means ± SD (n = 3). *P < .05 among the 3 groups (2-way analysis of variance with the post hoc Tukey honestly significant difference test).
Cytokine Response in Human Macrophages · JID 2013:207 (15 January) · 267
Contribution of Cytokine Induction to Pathogenicity in Mice
To examine whether the VN3028IIcl2 PA gene that conferred
high cytokine induction is associated with pathogenicity in
mammals, we compared the pathogenicity in mice of
IDN3006 with that of IDN3006/cl2PA. Four mice per group
were infected with 100
­103
PFU of each virus, and mortality
and morbidity were observed for 14 days after infection
(Figure 4). Mice infected with 102
PFU of IDN3006/cl2PA ex-
hibited 100% mortality, whereas 75% of mice infected with
IDN3006 survived. However, there were no statistically
Figure 3. Genes differentially expressed between low and high cytokine inducer H5N1 viruses in human macrophages. Human monocyte­derived
macrophages were infected with IDN3006, VN3028IIcl2, or IDN3006/cl2PA at a multiplicity of infection of 2. At 6 hours after infection, the cells were
harvested and subjected to microarray analysis. A, Gene expression levels determined by microarray are shown for the 7 cytokines tested in Figure 1.
Yellow indicates P < .05, compared with the IDN3006-infected group (1-way analysis of variance [ANOVA] with the Tukey honestly significant difference
(HSD) post hoc test). All values were normalized to the values of the mock group. B, A total of 129 genes were selected by 1-way ANOVA with the
Tukey HSD post hoc test (P < .05) and by filtering the genes whose expression changed at least 2.0-fold relative to the level in the IDN3006-infected
group. The set of genes differentially regulated in both pairs of VN3028IIcl2 and IDN3006/cl2PA as compared to IDN3006 was functionally annotated
by means of Gene Ontology (GO) grouping. Statistical significance was determined by using the Fisher exact test (P < .05). C, The differentially expressed
genes were connected in a de novo network based on known interactions in the Ingenuity Pathway Analysis (IPA) knowledgebase and functionally
annotated by use of IPA canonical pathways. Red indicates genes whose expression showed a change of >4.0-fold for both VN3028IIcl2 and IDN3006/
cl2PA, relative to IDN3006. Yellow indicates genes whose expression showed a change of >4.0-fold for IDN3006/cl2PA, relative to IDN3006. The genes
enriched with the top canonical pathway of "Role of Hypercytokinemia/hyperchemokinemia in the Pathogenesis of Influenza" are boxed.
268 · JID 2013:207 (15 January) · Sakabe et al
significant differences in mouse survival at any infectious dose
tested. The 50% mouse lethal doses of IDN3006 and
IDN3006/cl2PA were 2.1 × 102
and 5.9 PFU, respectively.
Thus, there was a trend toward increased virulence in mice for
IDN3006/cl2PA, compared with IDN3006, indicating that the
VN3028IIcl2 PA contributed to virus pathogenicity in mice.
These results suggest that the high cytokine induction pheno-
type contributes to high pathogenicity in mammals.
DISCUSSION
Here, we demonstrated that the PA gene of an H5N1 virus
confers high levels of cytokine expression in human mono-
cyte­derived macrophages. We also showed that PA enhances
pathogenicity in mammals without altered virus growth prop-
erties in human macrophages.
There are 8 amino acid differences in the PA protein
between IDN3006 (low cytokine inducer) and VN3028IIcl2
(high cytokine inducer) (Table 1). Yet, we did not find any PA
amino acid residues common to either the high or low cyto-
kine inducers examined in our previous study (Table 1) [16].
Although we attempted to identify the amino acid substitu-
tions responsible for the high cytokine production in macro-
phages, a single amino acid mutation did not substantially
change the cytokine induction level (data not shown), suggest-
ing a synergistic effect among the PA amino acids on cytokine
induction. The influenza polymerase complex has a role in in-
hibiting type I interferon in infected cells [25]. Since the PA
amino acid differences tested here did not affect the
Figure 4. Contribution of the acidic polymerase of VN3028IIcl2 to pathogenicity in mice. Four mice per group were intranasally infected with serially
diluted viruses ranging from 100
to 103
plaque-forming units (PFU) of IDN3006 or IDN3006/cl2PA, and survival (A) and body weight changes (B) were
monitored daily for 14 days after infection. The values are means ± SD (n = 4) for the mice that were alive at each time point. P values were determined
by using the log-rank test for each pair.
Table 1. Amino Acid Differences in Acidic Polymerase Between IDN3006 and VN3028IIcl2
Amino Acid IDN3006 VN3028IIcl2 Low Cytokine Inducersa
High Cytokine Inducersa
Known Functionb
90 V M V V/M ...
94 L I L/I I ...
163 I L I/L L cRNA interaction [26]
327 A E A/E E ...
391 T R T/K K/R ...
520 Y F Y/F F ...
594 G S G/S S ...
653 S P S/P P PB1 binding [26]
Abbreviations: cRNA, complementary RNA; PB1, basic polymerase 1.
a
Cytokine induction phenotype was determined in our previous study [16]. Low cytokine inducers include IDN3006 and A/Vietnam/UT31203A/07; high cytokine
inducers include VN3028IIcl2 and A/Hong Kong/483/97.
b
Functions previously reported to be associated with these amino acid positions.
Cytokine Response in Human Macrophages · JID 2013:207 (15 January) · 269
polymerase activity in human macrophages, the host respons-
es in the infected cells may have been affected by the altered
anti-interferon activity of the polymerase complex, possibly
mediated by PA or by some as yet unknown function of PA.
Virus growth assay revealed that PA was not responsible for
the differences between the high and low cytokine inducers
with respect to viral transcription, translation, and replication
in human macrophages. Interestingly, IDN3006 showed statis-
tically significantly higher growth than IDN3006/cl2PA in
human pulmonary epithelial cells (Supplementary Figure 1).
Viruses with efficient viral replication in epithelial cells are
generally highly pathogenic to mice, represented by viruses
with the PB2-E627K mutation, which enhances viral replica-
tion in human pulmonary epithelial cells [18]. When we in-
vestigated the virus titers in mouse organs, there were no
substantial differences between IDN3006 and IDN3006/cl2PA
in terms of virus growth in any mouse organ tested (Supple-
mentary Figure 2). Our results, including virus growth, cyto-
kine production, and gene expression in human macrophages,
indicate that the ability to induce high levels of cytokines in
macrophages, rather than viral replication efficiency in respira-
tory epithelial cells, may be responsible for enhanced pathoge-
nicity in mice.
In patients infected with H5N1 viruses, high levels of cyto-
kines, represented by MCP-1, MIG, IL-8, IL-10, IL-6, IFN-, and
RANTES, are generally observed in their sera [6, 9]. Since the
PA of a high cytokine inducer was responsible for high levels of
these cytokines in macrophages, PA may be a factor in the
increased mortality mediated by hypercytokinemia in humans
infected with H5N1 viruses. However, both viruses used in this
study were isolated from patients who had lethal outcomes, sug-
gesting the involvement of other unknown virulence determinants
in the high pathogenicity of H5N1 viruses in humans. Neverthe-
less, our results have clinical implications for cytokine-mediated
pathogenesis in humans infected with H5N1 viruses.
Supplementary Data
Supplementary materials are available at The Journal of Infectious Diseases
online (http://jid.oxfordjournals.org/). Supplementary materials consist of
data provided by the author that are published to benefit the reader. The
posted materials are not copyedited. The contents of all supplementary
data are the sole responsibility of the authors. Questions or messages
regarding errors should be addressed to the author.
Notes
Acknowledgments. We thank Susan Watson for editing the manuscript.
Financial support. This work was supported by a grant-in-aid for
Specially Promoted Research and by the Japan Initiative for Global
Research Network on Infectious Diseases from the Ministry of Education,
Culture, Sports, Science, and Technology, Japan; by a grant-in-aid of Sci-
entific Research from the Ministry of Health, Labor, and Welfare, Japan;
by a grant-in-aid for Exploratory Research for Advanced Technology
(ERATO) from Japan Science and Technology Agency (JST), Japan; and
by National Institute of Allergy and Infectious Diseases Public Health
Service Research grants. S. S. and R. T. were supported by Research
Fellowships from the Japan Society for the Promotion of Science for
Young Scientists.
Potential conflicts of interest. All authors: No reported conflicts.
All authors have submitted the ICMJE Form for Disclosure of Potential
Conflicts of Interest. Conflicts that the editors consider relevant to the
content of the manuscript have been disclosed.
References
1. Guan Y, Poon LL, Cheung CY, et al. H5N1 influenza: a protean pan-
demic threat. Proc Natl Acad Sci U S A 2004; 101:8156­61.
2. Li KS, Guan Y, Wang J, et al. Genesis of a highly pathogenic and po-
tentially pandemic H5N1 influenza virus in eastern Asia. Nature
2004; 430:209­13.
3. World Health Organization (WHO). Cumulative number of con-
firmed human cases of avian influenza A(H5N1) reported to WHO.
Geneva: WHO, 2012. http://www.who.int/influenza/human_animal_
interface/H5N1_cumulative_table_archives/en/index.html. Accessed 10
September 2012.
4. Hui DS. Review of clinical symptoms and spectrum in humans with
influenza A/H5N1 infection. Respirology 2008; 13(Suppl 1):S10­3.
5. de Jong MD, Bach VC, Phan TQ, et al. Fatal avian influenza A
(H5N1) in a child presenting with diarrhea followed by coma. N Engl
J Med 2005; 352:686­91.
6. de Jong MD, Simmons CP, Thanh TT, et al. Fatal outcome of human
influenza A (H5N1) is associated with high viral load and hypercyto-
kinemia. Nat Med 2006; 12:1203­7.
7. Abdel-Ghafar AN, Chotpitayasunondh T, Gao Z, et al. Update on
avian influenza A (H5N1) virus infection in humans. N Engl J Med
2008; 358:261­73.
8. Yuen KY, Chan PK, Peiris M, et al. Clinical features and rapid viral
diagnosis of human disease associated with avian influenza A H5N1
virus. Lancet 1998; 351:467­71.
9. Peiris JS, Yu WC, Leung CW, et al. Re-emergence of fatal human in-
fluenza A subtype H5N1 disease. Lancet 2004; 363:617­9.
10. van Riel D, Munster VJ, de Wit E, et al. H5N1 virus attachment to
lower respiratory tract. Science 2006; 312:399.
11. Cheung CY, Poon LL, Lau AS, et al. Induction of proinflammatory
cytokines in human macrophages by influenza A (H5N1) viruses: a mech-
anism for the unusual severity of human disease?. Lancet 2002; 360:
1831­7.
12. Lee SM, Gardy JL, Cheung CY, et al. Systems-level comparison of
host-responses elicited by avian H5N1 and seasonal H1N1 influenza
viruses in primary human macrophages. PLoS One 2009; 4:e8072.
13. Mok KP, Wong CH, Cheung CY, et al. Viral genetic determinants of
H5N1 influenza viruses that contribute to cytokine dysregulation. J
Infect Dis 2009; 200:1104­12.
14. Woo PC, Tung ET, Chan KH, Lau CC, Lau SK, Yuen KY.
Cytokine profiles induced by the novel swine-origin influenza A/H1N1
virus: implications for treatment strategies. J Infect Dis 2010; 201:
346­53.
15. Hui KP, Lee SM, Cheung CY, et al. H5N1 influenza virus-induced
mediators upregulate RIG-I in uninfected cells by paracrine effects
contributing to amplified cytokine cascades. J Infect Dis 2011; 204:
1866­78.
16. Sakabe S, Iwatsuki-Horimoto K, Takano R, et al. Cytokine production
by primary human macrophages infected with highly pathogenic
H5N1 or pandemic H1N1 2009 influenza viruses. J Gen Virol 2011;
92:1428­34.
17. Monteerarat Y, Sakabe S, Ngamurulert S, et al. Induction of TNF-
alpha in human macrophages by avian and human influenza viruses.
Arch Virol 2010; 155:1273­9.
18. Hatta M, Gao P, Halfmann P, Kawaoka Y. Molecular basis for high
virulence of Hong Kong H5N1 influenza A viruses. Science 2001;
293:1840­2.
270 · JID 2013:207 (15 January) · Sakabe et al
19. Neumann G, Watanabe T, Ito H, et al. Generation of influenza A
viruses entirely from cloned cDNAs. Proc Natl Acad Sci U S A 1999;
96:9345­50.
20. Livak KJ, Schmittgen TD. Analysis of relative gene expression data
using real-time quantitative PCR and the 2(-Delta Delta C(T))
Method. Methods 2001; 25:402­8.
21. Octaviani CP, Goto H, Kawaoka Y. Reassortment between seasonal H1N1
and pandemic (H1N1) 2009 influenza viruses is restricted by limited
compatibility among polymerase subunits. J Virol 2011; 85:8449­52.
22. Reed LJ, Muench H. A simple method of estimating fifty percent end
points. Am J Hyg 1938; 27:493­7.
23. Honda A, Ishihama A. The molecular anatomy of influenza virus
RNA polymerase. Biol Chem 1997; 378:483­8.
24. Huang TS, Palese P, Krystal M. Determination of influenza virus
proteins required for genome replication. J Virol 1990; 64:
5669­73.
25. Iwai A, Shiozaki T, Kawai T, et al. Influenza A virus polymerase inhib-
its type I interferon induction by binding to interferon beta promoter
stimulator 1. J Biol Chem 2010; 285:32064­74.
26. Naffakh N, Tomoiu A, Rameix-Welti MA, van der Werf S. Host re-
striction of avian influenza viruses at the level of the ribonucleopro-
teins. Annu Rev Microbiol 2008; 62:403­24.
Cytokine Response in Human Macrophages · JID 2013:207 (15 January) · 271
