Download Differential induction of cytotoxicity and apoptosis by influenza virus

Journal
of General Virology (1997), 78, 2821–2829. Printed in Great Britain
...........................................................................................................................................................................................................................................................................................
Differential induction of cytotoxicity and apoptosis by influenza
virus strains of differing virulence
Graeme E. Price,1 Harry Smith2 and Clive Sweet1
Microbial Molecular Genetics and Cell Biology Group1, School of Biological Sciences and The Medical School2,
The University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK
Two influenza viruses of differing virulence, clone
7a (virulent for humans and ferrets) and A/Fiji
(attenuated for both species), induced differential
levels of cytotoxicity (as measured by release of
lactate dehydrogenase from cells) and apoptosis (as
determined by altered nuclear morphology or flow
cytometry) in MDCK and U-937 cells. Clone 7a
induced more apoptosis and cytotoxicity than A/Fiji,
despite the fact that its infectious virus yields were
Introduction
Influenza A virus, a major cause of morbidity and mortality
in humans, is primarily a pathogen of the upper respiratory
tract. Infection results in both respiratory effects (sneezing,
hyperproduction of mucus, sore throat) and constitutional
effects (fever, myalgia, anorexia) (Sweet & Smith, 1980). In
humans and the ferret model, constitutional symptoms arise by
the production of ‘ endogenous pyrogen ’ (probably proinflammatory, pyrogenic cytokines) following interaction
between virus and inflammatory cells in the upper respiratory
tract (URT), the primary site of virus replication (Sweet et al.,
1979 ; Tinsley et al., 1987). In the ferret, influx of inflammatory
cells into the URT begins approximately 24 h post-infection
(p.i.) and correlates with the onset of fever (Sweet et al., 1979).
The clinical effects of H3N2 influenza viruses in man are more
severe than those of H1N1 viruses (Wright et al., 1980), and in
ferrets a H3N2 strain (clone 7a) induced a significantly higher
and longer fever than a H1N1 strain (A}Fiji) (Coates et al.,
1986). Since both viruses replicated similarly in the URT and
produced similar inflammatory responses, this indicated a
greater fever-producing capacity of clone 7a than that of
A}Fiji, an observation confirmed by direct injection of large
quantities of both live and UV-inactivated viruses into ferrets
Author for correspondence : Clive Sweet.
Fax ­44 121 414 6557. e-mail C.Sweet!bham.ac.uk
similar to or less than those for A/Fiji. This indicates
a greater capacity of clone 7a to induce the two
phenomena. Nevertheless the replication process
was clearly essential, since UV-irradiated virus induced little or no apoptosis, and apoptosis occurred
as a late event post-infection. The possible relevance of these findings to destruction of host cells
by influenza virus in vivo is discussed.
(Coates et al., 1986). However, attempts to demonstrate a
difference in the levels of pyrogenic cytokines induced from
phagocytic cells by these viruses failed for both nasal
inflammatory cells removed from infected febrile ferrets and
incubated in vitro, and human peripheral blood leukocytes
incubated with virus in vitro (Jakeman et al., 1991 a ; Tinsley et
al., 1987). This raised the possibility that the differences in
constitutional symptoms induced by these strains could arise
from differential damage to the respiratory epithelium, resulting in the differential release of cytokines from the site of
infection.
Influenza viruses destroy ciliated epithelial cells in the URT
(Hers & Mulder, 1961). However, the mechanism involved has
not been studied until recently, when induced programmed cell
death (apoptosis) was demonstrated in epithelial cells and
monocytes in vitro (Fesq et al., 1994 ; Hinshaw et al., 1994 ;
Takizawa et al., 1993) and in virus pneumonia of mice (Mori et
al., 1995).
Here we report the differential cytotoxicity and apoptosis
produced by clone 7a and A}Fiji in two tissue culture systems.
Methods
+ Viruses. A}Fiji}15899}83 (H1N1) and the virulent clone 7a of the
A}PR}8}34¬A}England}939}69 (H3N2) reassortment system (Beare
& Hall, 1971) were grown for 48 h in the allantoic cavity of embryonated
10-day-old hen’s eggs. The viruses were harvested from these eggs,
aliquoted and stored at ®70 °C prior to use.
0001-4818 # 1997 SGM
CICB
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 05:34:24
G. E. Price, H. Smith and C. Sweet
+ Titration of viruses. Previously, difficulties have been encountered
in plaquing the two strains in MDCK cells (unpublished observations),
thus egg-based assay systems were selected. Stock viruses were titrated
in 10-day-old hen’s eggs as previously described (Sweet et al., 1974) and
the 50 % egg infectious dose (EID ) was calculated according to the
&!
method of Thompson (1947). The many experimental samples were
titrated in the more convenient allantois-on-shell (egg-bit) assay (Sweet
et al., 1974). Assay plates were shaken at 37 °C for 72 h before reading.
Titres were calculated as 50 % egg-bit infectious doses (EBID ).
&!
However, the egg-bit assay was less sensitive than the egg assay for both
viruses, and A}Fiji replicated less well in egg bits than clone 7a. As in
previous work (Coates et al., 1985), the results from the egg-bit assay
were converted to EID values with correction factors obtained as
&!
follows : ten replicate samples of clone 7a and A}Fiji grown in eggs or in
MDCK cells (see below) were titrated in both egg bits and eggs (Coates
et al., 1985) and the EID : EBID ratios were calculated. The results were
&!
&!
1 : 5±5 (SE 1±6) and 1 : 83 (SE 21±4) respectively for egg-grown stocks of
clone 7a and A}Fiji, and 1 : 4±2 (SE 0±9) and 1 : 113 (SE 32±2) respectively
for clone 7a and A}Fiji grown in MDCK cells. In some experiments (see
Results), samples from MDCK and U-937 cells were treated with trypsin
(Sigma ; 10 µg}ml) for 90 min prior to seeding into EBID assay plates.
&!
Appropriate titrations showed that the EID : EBID ratios were lower
&!
&!
for these trypsin-treated samples, namely 1 : 1±8 (SE 0±6) for clone 7a and
1 : 14 (SE 5±1) for A}Fiji.
+ Production of UV-inactivated virus. Allantoic fluid harvested
from 50 infected eggs was clarified from cell debris by low-speed
centrifugation (1000 g, 30 min at 4 °C) and then spun at 80 000 g for
90 min at 4 °C to pellet the virus. The pellets were soaked in TSE buffer
(0±85 % saline, 0±01 M Tris–HCl pH 7±4, 1±3 mM EDTA) overnight at
4 °C and 2 vols 57 % sucrose (w}v) added to give a final concentration
of 38 % (w}v). The suspension was layered onto 38 % sucrose (w}v) and
spun at 80 000 g for 60 min at 4 °C. The virus-containing supernatant
was removed, diluted with TSE to give a suspension of 30 % sucrose
(w}v) and spun overnight at 80 000 g at 4 °C. Final virus pellets were
resuspended in Dulbecco’s A PBS pH 7±3 and the protein concentration
was determined using a Coomassie Plus protein assay kit (Pierce &
Warriner, Chester, UK). The suspension was irradiated with UV light
(250 nm) at a distance of 7±5 cm for 7 min (giving a light intensity of
120 µW}cm#) in an open petri-dish on ice (Pickering et al., 1992).
Inactivation was confirmed by double passage in 10-day-old eggs
(Coates et al., 1986). Inactivated virus stocks were aliquoted and stored at
®70 °C prior to use.
+ Cell cultures. Madin–Darby canine kidney (MDCK) epithelial cells
and U-937 human histiocytic cells were used. MDCK cells were grown
in Eagle’s minimum essential medium, supplemented with 10 % newborn
bovine serum (NBCS), 4 mM -glutamine, 0±7 % sodium bicarbonate,
100 IU}ml penicillin, 100 µg}ml streptomycin and 50 µg}ml gentamycin. Cells were maintained in the above medium with NBCS
concentration reduced to 2 %. U-937 cells were grown in RPMI 1640
medium, supplemented with 10 % NBCS, 4 mM -glutamine, 100 IU}ml
penicillin, 100 µg}ml streptomycin, 50 µg}ml gentamycin and 2 mM
sodium pyruvate (complete RPMI). All media components were obtained
from Gibco. Both cell types were incubated at 37 °C in a 5 % CO
#
atmosphere.
+ Infection of cell cultures. MDCK cells were seeded into either
96-or 48-well flat-bottomed plates and allowed to become confluent.
Monolayers were washed twice with warmed PBS and the appropriate
volume of virus suspension, diluted in PBS (50 µl or 100 µl for 96- and
48-well plates respectively) to give the required m.o.i., was added
CICC
to each well. Controls contained only PBS. After 60 min at 37 °C,
cells were washed three times in warmed PBS, and maintenance medium
(200 µl and 300 µl for 96- and 48-well plates) was added.
U-937 cells were pelleted and resuspended to 4¬10' cells}ml in
complete RPMI (without serum). One ml of this suspension was
transferred to culture tubes and washed twice with serum-free complete
RPMI (SF-cRPMI). One ml of virus suspension diluted to the required
m.o.i. in SF-cRPMI was then added. After 60 min at 37 °C, cells
were washed three times with 5 ml SF-cRPMI and resuspended in 2 ml
SF-cRPMI prior to the addition of 2 ml complete RPMI to give a final
concentration of 1¬10' cells}ml in complete RPMI with 5 % serum.
Cells were then re-incubated for the required time.
When MDCK cells were treated with inactivated viruses, virus
suspensions were diluted in PBS to give 100 and 10 pg viral protein per
cell (this would be equivalent to approximately 10 000 and 1000 EID
&!
per cell respectively of infectious virus). Control cells had PBS added.
+ Cytotoxicity assays. Cytotoxicity assays were conducted with the
Cytotox 96 (Promega), which enzymatically assays lactate dehydrogenase (LDH) released from damaged cells into the supernatant. Control
samples were set up to correct for LDH in the medium (serum contains
LDH at low levels) and for spontaneous release of LDH from uninfected
cells. Maximum release of LDH from uninfected cells was measured after
lysis of cells by adding 9 % Triton X-100 to give a final concentration of
0±9 % in the medium. To correct for this addition of lysis solution to the
cell suspension, an equal volume of this solution was added to medium
(volume correction control). For experiments with MDCK cells, supernatants were taken, diluted 1 : 2 and incubated with the assay substrate at
room temperature for 15 min before the reaction was stopped by
addition of 1 M acetic acid. For experiments with U-937 cells a similar
procedure was used, but cells were sedimented by centrifugation,
supernatants were assayed undiluted and the reaction was stopped after
30 min at room temperature. Percentage cytotoxicity values were
calculated by substitution of absorbance values at 492 nm into the
equation : % Cytotoxicity ¯ [(experimental®medium control)®(spontaneous®medium control)}(maximum release®volume correction
control)]¬100.
+ Determination of the percentage of apoptotic and infected
cells by fluorescence microscopy. Cell monolayers growing on
10 mm glass coverslips in 48-well plates were fixed for 10 min with either
acid ethanol (1 part glacial acetic acid, 2 parts ethanol) for acridine orange
staining or methanol for immunofluoresence. For acridine orange staining,
fixed cells were washed twice in McIlvain’s buffer (71 mM Na HPO ,
#
%
65 mM citric acid), stained for 10 min with 0±1 % acridine orange in
McIlvain’s buffer and rinsed twice more before mounting. For immunofluoresence, fixed monolayers were treated with propidium iodide
(0±1 mg}ml) as a nuclear stain and washed in PBS. Then monoclonal
antibody (1 : 250) to the nucleoprotein (NP) of the X-31 (H3N2) strain of
influenza virus (kindly provided by A. Douglas, NIMR, Mill Hill, London,
UK) was added and allowed to bind at 37 °C for 30 min before washing
with PBS. Finally a fluorescein isothiocyanate (FITC)-conjugated rabbit
anti-mouse IgG (Sigma ; 1 : 100) was added and allowed to bind at 37 °C
for 30 min. After washing in PBS, the coverslips were removed, mounted
in triethylenediamine–glycerol to reduce fading (Johnson et al., 1982) and
examined by fluorescence microscopy. Fields were selected at random by
moving the monolayer slightly out of focus when changing field, and at
least 1000 cells were counted for each coverslip examined.
Apoptotic cells were identified by their characteristic nuclear
morphology as revealed by the DNA-binding stains used, which showed
small, brightly staining bodies (green in the case of acridine orange, red
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 05:34:24
Influenza virus cytotoxicity and apoptosis
for propidium iodide) and lack of a cohesive nuclear structure. In contrast,
non-apoptotic cells were identified by their large, generally ovoid nuclei.
Infected cells were identified by the presence of FITC label (green
fluorescence) in the cytoplasm. Good agreement of percentages of
apoptotic and infected cells was seen between replicate samples (at least
three for each treatment), with replicate to replicate variation generally
within 10–15 %. Acridine orange and propidium iodide staining was
directly comparable for detection of apoptotic cells.
+ Detection of apoptosis by DNA laddering. MDCK cells were
infected at an m.o.i. of 10 EID per cell as described above, washed three
&!
times in PBS and incubated in maintenance medium. At intervals,
supernatants containing detached cells were removed from the adherent
monolayers and suspended cells collected by centrifugation at 3000 g for
5 min. DNA was extracted separately from both the adherent cell
monolayer and the sedimented cells by addition of 400 µl cell lysis buffer
(10 mM Tris pH 8±0, 1 mM EDTA, 0±1 % SDS), overnight digestion at
37 °C with 100 µg Proteinase K (Sigma) and extraction and precipitation
with phenol–chloroform and iso-propyl alcohol according to standard
methods (Sambrook et al., 1989). Pellets were resuspended in 30 µl
distilled water and treated with RNase A before electrophoresis on 1 % or
1±8 % agarose gels. Bands were visualized with ethidium bromide under
UV light.
+ Determination of the percentage apoptotic cells by flow
cytometric analysis. This was used for U-937 cells as fluorescence
microscopy methods were not convenient because of the small size of
these cells. U-937 cells which had been exposed to virus and control cells
were sedimented and fixed with 80 % ethanol in PBS. They were then
resuspended in counting buffer (0±0024 % EDTA, 1 % Tween 20 in PBS)
to give a final concentration of 5¬10& cells per 200 µl and shaken gently
to prevent clumping. A fluorescent DNA-binding dye, 7-amino actinomycin D (7-AAD, Sigma ; 125 ng}ml), was added and the cells analysed
using a flow cytometer (EPICS Elite, Coulter) set to discriminate between
phases of the cell cycle. At least 10 000 events (cells) were counted for
each sample. The apoptotic cell population was identified by the reduced
cellular DNA content (as assessed by reduced intensity of 7-AAD) in
apoptotic cells, compared to control cells, which have a characteristic
peak corresponding to the G phase of the cell cycle (see Results). The
"
percentage of apoptotic cells was calculated by comparing the number of
cells in the apoptotic population with the total number of cells counted,
using WinMDI software (Scripps Institute, La Jolla, USA).
+ Statistical analysis. Data were collated and mean values analysed
for statistical significance by the Student’s t-test.
Results
Differential cytotoxicity and apoptosis produced by
clone 7a and A/Fiji in MDCK cells
In preliminary experiments with inocula ranging from
100–0±1 EID per cell and incubation times of 24, 36, 48 and
&!
72 h, clone 7a was more cytotoxic for MDCK cells than was
A}Fiji (Fig. 1). Only for an inoculum of 100 EID per cell was
&!
the maximum toxicity for A}Fiji (ca. 40 %) similar to clone 7a
(ca. 45 %), but it occurred later in the infection. For all other
inocula, the maxima for A}Fiji were well below those for clone
7a, and at an inoculum of 0±1 EID per cell A}Fiji showed no
&!
toxicity, in contrast to clone 7a, which was still appreciably
toxic (ca. 20 %).
Fig. 1. Percentage cytotoxicity seen with MDCK cells at various times p.i.
with 100 (a), 10 (b), 1 (c) and 0±1 (d) EID50 per cell of clone 7a
(continuous line) and A/Fiji (dashed line) influenza viruses. Error bars
show standard deviations from the mean of 4 replicates at each timepoint.
The experiments were then extended to examine apoptosis
as well as cytotoxicity, and the percentage of infected cells and
virus yield were also measured. Fig. 2 shows how apoptotic
cells were recognized and Table 1 summarizes the results of
four experiments using an inoculum of 10 EID per cell
&!
examined at 24 h p.i. The percentage cytotoxicity varied
between experiments ; in these experiments from about
14–30 % for clone 7a and 3–15 % for A}Fiji, but in each
experiment the difference between clone 7a and A}Fiji was
statistically significant (P % 0±01). The percentage of apoptotic
cells also varied between experiments from about 27–40 % for
clone 7a and between 11 % and 22 % for A}Fiji in the
experiments cited in Table 1. Again, differences between the
viruses were significant (P % 0±05) in each experiment. Variation in the levels of cytotoxicity and apoptosis induced by
clone 7a and A}Fiji was apparent between experiments, but
within each experiment clone 7a induced significantly more
cytotoxicity and apoptosis than A}Fiji. The reasons for this are
unknown, but could include variations in susceptibility between passage number of cells, cell density at infection or LDH
content of the serum used. Fragmentation of chromosomal
DNA is a characteristic feature of apoptosis, and ‘ apoptotic
ladders ’ were detectable in DNA extracted from adherent
clone 7a-infected cells at 12 h p.i., but not in either A}Fijiinfected or control cells, or culture supernatants at this time
(not shown). At 24 h p.i. and subsequent times, laddering was
seen in apoptotic cells in culture supernatants (Fig. 3) and
adherent cells infected with either virus, and at 48 h p.i. in
control cells to some extent. The DNA fragments were found
to be on average 189 bp multimers, which is in agreement with
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 05:34:24
CICD
G. E. Price, H. Smith and C. Sweet
Fig. 2. Cells stained with acridine orange
(A, C and E) or propidium iodide and
FITC-labelled anti-NP antibody (B, D and
F). Clone 7a-infected cells (10 EID50 per
cell ; 24 h p.i.) are shown in A and B,
A/Fiji-infected cells (10 EID50 per cell ;
24 h p.i.) are shown in C and D, and
control cells are shown in E and F. Large
arrows in A, B, C and D indicate apoptotic
cells ; note the characteristic nuclear
condensation and blebbing. Small double
arrows in B and D indicate the more
granular appearance of NP with A/Fijiinfected cells (D) in comparison to the
diffuse staining seen with clone 7a (B).
other reports. The percentages of infected cells were high and
not significantly different for the two viruses. Finally, and
importantly, the yields of both viruses were not significantly
different, except in experiment 2 when the yield for A}Fiji was
greater than for clone 7a. It is important to note that these
experiments were essentially carried out under single-cycle
conditions with an inoculum that infected over 90 % of the
cells. Further (multicycle) replication would be minimal, as
liberated virus would be largely non-infectious for MDCK cells
since trypsin was not included in the growth medium of these
cultures.
The differences between the viruses might have reflected
different optimum times for sampling, and thus the experiments
described above were repeated, but samples were taken at
various times p.i. The results of a typical experiment are shown
in Table 2. Mean percentage cytotoxicities increased with time
CICE
from about 3 % at 2 h p.i. to about 70 % at 36 h p.i. for clone 7a
and from about 1–36 % for A}Fiji. Whilst again percentage
cytotoxicities differed between experiments, at all time-points
except 6 h p.i. the mean percentage cytotoxicity induced by
clone 7a was greater than that induced by A}Fiji, but only in
one case (24 h p.i.) was this significant. Apoptosis was not
evident until 12 h p.i., then the mean percentages of apoptotic
cells increased with time, and those for clone 7a were
significantly higher than for A}Fiji at all time-points. Although
the mean percentage of cells infected with A}Fiji was lower
than that for clone 7a (80 % compared to 100 %), this was not
significant in most cases. Clone 7a-infected cells showed very
bright fluorescence, but cells infected with A}Fiji gave lessintense staining, with a more speckled pattern than the diffuse
staining seen with clone 7a (Fig. 2). Although NP was
detectable in the cytoplasm earlier for clone 7a-infected cells
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 05:34:24
Influenza virus cytotoxicity and apoptosis
Table 1. Levels of cytotoxicity and apoptosis in MDCK cells 24 h p.i. with 10 EID50 per cell of clone 7a or A/Fiji
Clone 7a did not significantly differ from A}Fiji, except a, P ! 0±05 ; b, P ! 0±02 ; c, P ! 0±01 ; and d, P ! 0±001. Other figures are not significantly
different between strains.
% Cytotoxicity (³SD)*
Expt 1
Expt 2
Expt 3
Expt 4
% Cells showing apoptosis (³SD)†
% Cells infected (³SD)‡
Virus yield
(log10EID50 per
ml³SD)§
7a
Fiji
7a
Fiji
Control
7a
Fiji
Control
7a
Fiji
17±8³7±9c
13±9³2±4d
29±4³6±8d
28±7³4±9d
6±9³1±9
3±4³0±8
14±9³3±5
7±8³0±9
26±7³7±7d
26±6³2±5b
39±9³13±5a
34±2³7±4b
10±9³3±4
10±5³0±8
21±9³9±0
19±8³8±3
2±5³3±4
2±5³0
1±0³1±8
0±6³0±1


93±6³5±4
89±6³3±1


85±2³14±3
91±9³6±7


0±2³0±3
0³0
5±4³0±4
4±9³0±4c
5±3³0±4

5±6³0±2
5±9³0±4
5±7³0±4

* LDH release as a percentage of that obtained from fully lysed cells (see Methods).
† Percentage of apoptotic cells was calculated by examination of nuclear morphology (as described in Methods), by staining with acridine orange
in experiments 1 and 2, and propidium iodide in experiments 3 and 4. At least 1000 cells were counted for each sample.
‡ Percentage of infected cells was calculated by counting cells staining positive for NP by immunofluorescence (see Methods). At least 1000 cells
were counted for each sample.
§ EID per ml estimated from EBID measurement (see Methods) on virus not treated with trypsin.
&!
&!
, Not determined.
Fig. 3. DNA laddering induced by influenza virus in MDCK cells. DNA
extracted from control and infected (10 EID50 per cell) cell supernatants
24 h p.i. run on a 1±8 % agarose gel and visualized with ethidium bromide.
Samples of 15 µl DNA extracted from supernatants of 1¬106 cells were
loaded onto each lane. This is representative of the DNA content from
cells released into the medium following different treatments. Lane A,
marker (1 kb ladder ; Gibco) ; lane B, clone 7a-infected ; lane C, A/Fijiinfected ; lane D, control cells ; lane E, high molecular mass marker (uncut
bacteriophage λDNA ; Gibco). Note the increased intensity of bands from
clone 7a-infected cell supernatants compared to A/Fiji-infected
supernatants, indicating more extensive apoptosis.
than for A}Fiji-infected cells (6 h p.i. vs 12 h p.i), cells infected
with either virus had NP present in the nucleus at 6 h p.i. The
presence of NP in the nucleus would be masked by staining
DNA with propidium iodide for the detection of apoptosis.
Thus, the apparently lower number of infected cells with A}Fiji
did not account for the observed differences in cytotoxicity or
apoptosis. With regard to virus yield, where significant
differences were observed, yields for A}Fiji were greater than
those for clone 7a. The apparently high virus yield 2 h p.i.
probably represents eluted attached virus that was not
removed by washing. In these experiments, residual inoculum
was not neutralized by antibody and remaining attached virus
was not removed by neuraminidase treatment, as these
procedures may have affected the induction of apoptosis.
Nevertheless, it is clear that virus yields increase significantly
at time-points after 2 h. Interestingly, yields were similar only
after applying the EBIB : EID conversion factors (see
&!
&!
Methods) to correct for the poor ability of A}Fiji to replicate
in egg bits. Further protease treatment of virus with trypsin
prior to assay in egg bits enhanced A}Fiji titres by about 1
log , whereas similar treatment of clone 7a enhanced titres by
"!
only about 0±2 log , suggesting that under single-cycle
"!
conditions A}Fiji produces as many virus particles as clone 7a,
but these particles are largely non-infectious, with uncleaved
haemagglutinin.
When both viruses were inactivated by UV irradiation,
treatment of MDCK cells with 100 and 10 pg}cell of viral
protein (corresponding to 10 000 and 1000 EID per cell of
&!
infectious virus) for 12, 24 and 48 h produced only low levels
(! 10 %) of cytotoxicity and apoptosis, and any differences in
production of apoptosis between the viruses were not apparent
(two experiments, results not shown).
Differential cytotoxicity and apoptosis produced by
clone 7a and A/Fiji in U-937 cells
Using flow cytometry cell cycle analysis protocols, it was
possible to discriminate between apoptotic and normal cells.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 05:34:24
CICF
G. E. Price, H. Smith and C. Sweet
Table 2. Levels of cytotoxicity and apoptosis in MDCK cells at various times p.i. with 10 EID50 per cell of clone 7a or
A/Fiji
Differences between clone 7a and A}Fiji were not significant, except a, P ! 0±05 ; b, P ! 0±01 ; and c, P ! 0±001.
% Cytotoxicity (³SD)*
Time
(h p.i.)
2
6
12
24
36
7a
Fiji
3±1³3±7
0±7³2±8
3±8³7±8
5±8³2±0
11±7³6±5
7±1³4±8
63±6³3±7b 30±9³6±5
69±9³3±4 36±3³27±6
% Cells showing apoptosis (³SD)†
7a
Fiji
Control
0³0
0³0
48±3³2±1c
97±8³1±5c
97±4³3±0b
0³0
0³0
1±7³1±7
30±5³6±3
35±8³13±5
0³0
0³0
0³0
0³0
1±8³0±7
% Cells infected (³SD)‡
7a
Fiji
All cells stain faintly
89±9³2±7c
0³0
99±7³0±3
81±3³0±3
100³0
80±4³7±8
100³0a
60±5³10±5
Virus yield
(log10EID50 per
ml³SD)§
Control
7a
Fiji
0³0
0³0
0³0
0³0
0³0
5±1³0±4b
5±5³0±5
6±4³0±1b
6±7³0±2
6±2³0±1c
6±5³0±3
5±6³0±7
7±3³0±3
7±5³0±4
6±7³0±3
* LDH release as a percentage of that obtained from fully lysed cells (see Methods).
† Percentage of apoptotic cells was calculated by examination of nuclear morphology (as described in Methods), by staining with propidium
iodide. At least 1000 cells were counted for each sample.
‡ Percentage of infected cells was calculated by counting cells staining positive for NP by immunofluorescence (see Methods). At least 1000 cells
were counted for each sample.
§ EID per ml estimated from EBID (with trypsin added) measurements (see Methods).
&!
&!
respectively. Cytotoxicity levels were lower than for MDCK
cells, with no significant cytotoxicity being seen until 48 h p.i.,
when that induced by clone 7a was significantly greater than
that induced by A}Fiji (Table 3). The virus yields were low,
and not significantly different between the two strains. It is
possible that the virus yields at all time-points represent
residual inoculum, since it is known that influenza virus is
poorly productive in monocytic cells (Rodgers & Mims, 1982),
and virus was found to be stable in culture medium at 37 °C
over this time period (data not shown).
Discussion
Fig. 4. Flow cytometric analysis of apoptosis in U-937 cells. U-937 cells
infected with 10 EID50 per cell of clone 7a (solid line) or A/Fiji (dashed
line) influenza viruses and control cells (shaded) were taken at 24 h p.i.,
fixed and stained with 7-AAD prior to cell cycle analysis. The first marker
(M1) denotes the apoptotic (sub-G1) cell population with reduced cellular
DNA content as determined by reduced 7-AAD binding (see Methods),
whilst M2 shows the non-apoptotic population. Note the increased sub-G1
peak with both viruses (clone 7a " A/Fiji) in comparison to control cells.
The profiles for U-937 cells infected with clone 7a and A}Fiji
are shown in Fig. 4. The results of an experiment in which
cytotoxicity and apoptosis were measured at various times p.i.
is shown in Table 3 ; these results are typical of three such
experiments. Only low levels of apoptosis were detected at 2
and 12 h p.i., but at 24 h p.i. around 30 % of clone 7a-infected
cells were undergoing apoptosis, in comparison to about 10 %
of A}Fiji-infected cells. By 48 h p.i. these figures had risen
again to about 40 % and 20 % for clone 7a and A}Fiji
CICG
Although apoptosis induced by influenza virus has been
described in many cell types, such as HeLa cells (Takizawa et al.,
1993), human monocytes (Fesq et al., 1994), avian lymphocytes
(Hinshaw et al., 1994), murine macrophages (Lowy & Dimitrov,
1995) and Vero cells (Govorkova et al., 1996), to our
knowledge this is the first report of differential induction of
apoptosis by different influenza virus strains. It is clear that for
MDCK cells clone 7a is more cytotoxic and induces more
apoptosis than A}Fiji (Fig. 1, Tables 1 and 2). Furthermore,
these differences appear not to be due to greater overall
replication of clone 7a compared to A}Fiji, because yields of
clone 7a were either similar to or less than those of A}Fiji
(Tables 1 and 2). Nevertheless, elements of the replication
process must occur for these effects to become manifest,
because UV-inactivated virus had barely any cytotoxic or
apoptotic effect, and apoptosis occurred as a late event p.i. The
experiments with U-937 cells were on a smaller scale than
those using MDCK cells. Both cytotoxicity and apoptosis
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 05:34:24
Influenza virus cytotoxicity and apoptosis
Table 3. Levels of cytotoxicity, apoptosis and infectious virus yields in U-937 cells at
various times p.i. with 10 EID50 per cell of clone 7a or A/Fiji
Differences between clone 7a and A}Fiji were not significant, except a, P ! 0±001.
% Cytotoxicity (³SD)*
Time
(h p.i.)
2
12
24
48
% Cells showing apoptosis†
Virus yield
(log10EID50
per ml³SD)‡
7a
Fiji
7a
Fiji
Control
7a
Fiji
®0±2³0±7
0±2³0±5
®1±1³1±8
17±0³0±8a
0±1³0±4
0±7³0±6
®0±2³1±3
6±6³1±7
1±8
0±7
26±4
40±7
1±2
2±2
9±9
18±0
4±4
0±5
3±2
5±2
5±8³0±3
5±7³0±1
5±7³0±1
4±7³0±6
5±3³0±2
5±5³0±4
5±7³0±1
4±8³0±1
* LDH release as a percentage of that obtained from fully lysed cells (see Methods).
† Percentage of apoptotic cells was determined by flow cytometry (see Methods). At least 10 000 cells were
counted for each sample.
‡ EID per ml estimated from EBID (with trypsin added) measurements (see Methods).
&!
&!
were less evident than for the experiments with MDCK cells,
but the pattern of clone 7a being more active than A}Fiji
without its yield being greater than that of A}Fiji was the
same. The induction of a lower yield of clone 7a from cells
showing greater apoptosis fits with the suggestion that
apoptosis may be a cellular mechanism induced to limit virus
replication (Martz & Howell, 1989 ; Teodoro & Branton, 1997).
However, influenza virus yields from MDCK cells transfected
with bcl-2, an inhibitor of apoptosis, were reduced suggesting
that apoptosis was required for maximal yield (Olsen et al.,
1996).
What component or product of influenza virus induces
apoptosis? The non-structural protein (NS1) of influenza virus
A}Turkey}Ontario}7732}66 (H5N9) has some homology to
a region of the Fas antigen, a pro-apoptotic protein. Also, coexpression of bcl-2 alters localization of NS1 and NP from the
nucleus to the cytoplasm, as well as alteration of the
haemagglutinin glycosylation pattern (Hinshaw et al., 1994).
Both NS1 and NP are produced early in infection and are
candidates for triggering apoptosis. The fact that UVinactivated virus had little apoptotic activity suggests that a
virus-induced product rather than a virion component is
involved. In the present work, the differential induction of
apoptosis by clone 7a and A}Fiji may reflect differential
distribution of viral proteins within subcellular compartments
in infected cells. NP was distributed throughout the cytoplasm
as well as in the nucleus of clone 7a-infected cells, whereas it
was largely restricted to the nucleus of A}Fiji-infected cells
until later in infection. Clone 7a derives its matrix and PB2
genes from A}PR}8}34 (H1N1), but all other genes are
derived from A}England}939}69 (H3N2) (Matsuyama et al.,
1980). Hence, it may be of interest to examine other
A}England¬PR}8 reassortants, especially as preliminary
results show that PR}8 induces an intermediate level of
apoptosis compared to that of clone 7a and A}Fiji (unpublished
data).
Is the cytotoxicity detected by release of LDH connected
with apoptosis? Probably. The release of LDH is consistent
with the observation that influenza virus enhances the leakiness
of cell membranes in conjunction with bacterial toxins (Jakeman
et al., 1991 b) and may indicate that influenza viruses are
intrinsically harmful to cell membranes either on entry into the
cell or budding from it. Apoptosis results in the formation of
apoptotic bodies surrounded by an intact plasma membrane
(Cohen et al., 1992), but later the membranes surrounding
apoptotic bodies may degenerate (Savill et al., 1993), and this
would result in the release of LDH. This is the most likely
source of the LDH detected in our in vitro cytotoxicity studies.
This is supported by the observation that cytotoxicity
generally increased significantly after apoptosis had begun. In
vivo, apoptotic bodies would probably be engulfed by
professional phagocytes before membrane breakdown occurred.
LDH release has been suggested as a marker for replication
in the assessment of antiviral compounds active against
influenza virus (Watanabe et al., 1995). Our findings suggest
caution in drawing this conclusion, because for our two strains
the levels of replication (which were essentially similar) cannot
be correlated with LDH release (which was clearly far greater
for clone 7a). In a similar vein, the cytopathic effect (CPE) of
influenza virus is sometimes used as a marker for virus
replication, for example in TCID determinations. But in this
&!
work, production of apoptosis (which is basically responsible
for CPE) was different for the two strains for similar levels of
replication.
Differential induction of apoptosis between strains has
been seen with Theiler’s murine encephalomyelitis virus
(Jelachich & Lipton, 1996) and reovirus (Tyler et al., 1995). In
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 05:34:24
CICH
G. E. Price, H. Smith and C. Sweet
each case the more virulent strain also induced more apoptosis,
and in the latter case this was associated with reduced
replication (Rodgers et al., 1997). For reovirus, production of
pro-inflammatory cytokines also correlated with capacity to
induce apoptosis (Farone et al., 1996). Since previous work has
shown that clone 7a is more virulent for ferrets than A}Fiji and
more rapidly induces a greater and longer fever (Smith &
Sweet, 1988), the effects observed in this work may well have
some connection with pathogenicity. Apoptosis may, for
example, trigger the release of active pro-inflammatory or
pyrogenic cytokines. Several cellular enzymes are activated
during apoptosis, including interleukin (IL)-1β converting
enzyme (ICE), which cleaves pro-IL-1β into active IL-1β
(Kumar, 1995). IL-1β is secreted by inflammatory cells and is a
pyrogen which induces fever in several animal species
(Dinarello, 1992). The fact that cytotoxic effects and apoptosis
have been shown in a human histiocytic lymphoma cell line (U937) is important because cytokines are readily inducible in
such cells, and genes involved in both apoptosis and cytokine
production are inducible by the same transcriptional regulators
(e.g. NF-IL6 and NF-κB) (Siebenlist et al., 1994 ; Wada et al.,
1995). Two strains which differ in their ability to induce
apoptosis may be useful tools in the search for the viral
component or cellular products involved in these pathways.
We should like to thank Mr Graham Davies, Department of Anatomy
for assistance with flow cytometry and Claire Grattan and Parveen
Akhtar for assistance with inactivated viruses and the LDH assay
respectively. G. E. P. was supported by Glaxo-Wellcome Research and
Development, whilst C. G. and P. A. were supported by Wellcome Trust
and SGM vacation studentships.
Govorkova, E. A., Murti, G., Meignier, B., Detaisne, C. & Webster,
R. G. (1996). African green monkey kidney (Vero) cells provide an
alternative host cell system for influenza A and influenza B viruses. Journal
of Virology 70, 5519–5524.
Hers, J. F. P. & Mulder, J. (1961). Broad aspects of the pathology and
pathogenesis of human influenza. American Review of Respiratory Disease
83, 84–97.
Hinshaw, V. S., Olsen, C. W., Dybdahl-Sissoko, N. & Evans, D. (1994).
Apoptosis : a mechanism of cell-killing by influenza A and influenza B
viruses. Journal of Virology 68, 3667–3673.
Jakeman, K. J., Bird, C. R., Thorpe, R., Smith, H. & Sweet, C. (1991 a).
Nature of the endogenous pyrogen (EP) induced by influenza viruses :
lack of correlation between EP levels and content of the known pyrogenic
cytokines, interleukin 1, interleukin 6 and tumour necrosis factor. Journal
of General Virology 72, 705–709.
Jakeman, K. J., Smith, H. & Sweet, C. (1991 b). Influenza virus
enhancement of membrane leakiness induced by staphylococcal α toxin,
diphtheria toxin and streptolysin S. Journal of General Virology 72,
111–115.
Jelachich, M. L. & Lipton, H. L. (1996). Theiler’s murine encephalomyelitis virus kills restrictive but not permissive cells by apoptosis.
Journal of Virology 70, 6856–6861.
Johnson, G. D., Davidson, R. S., McNamee, K. C., Russell, G., Goodwin,
D. & Holborow, E. J. (1982). Fading of immunofluorescence during
microscopy : a study of the phenomenon and its remedy. Journal of
Immunological Methods 55, 231–242.
Kumar, S. (1995). ICE-like proteases in apoptosis. Trends in Biochemical
Sciences 20, 198–202.
Lowy, R. J. & Dimitrov, D. S. (1995). Induction of apoptosis in J774.1
macrophages by influenza-virus. Molecular Biology of the Cell 6, 2063.
Martz, E. & Howell, D. M. (1989). CTL : virus control cells first and
cytolytic cells second ? Immunology Today 10, 79–86.
Matsuyama, T., Sweet, C., Collie, M. H. & Smith, H. (1980). Aspects of
virulence in ferrets exhibited by influenza virus recombinants of known
genetic constitution. Journal of Infectious Diseases 141, 351–361.
References
Mori, I., Komatsu, T., Takeuchi, K., Nakakuki, K., Sudo, M. & Kimura, Y.
(1995). In vivo induction of apoptosis by influenza virus. Journal of
Beare, A. S. & Hall, T. S. (1971). Recombinant influenza A viruses as
General Virology 76, 2869–2873.
live vaccines for man. Lancet ii, 1271–1273.
Olsen, C. W., Kehren, J. C., Dybdahl-Sissoko, N. R. & Hinshaw, V. S.
(1996). Bcl-2 alters influenza virus yield, spread, and hemagglutinin
Coates, D. M., Sweet, C., Quarles, J. M., Overton, H. A. & Smith, H.
(1985). Severity of fever in influenza : studies on the relation between
viral surface antigens, pyrexia, level of nasal virus and inflammatory
response in the ferret. Journal of General Virology 66, 1627–1631.
Coates, D. M., Sweet, C. & Smith, H. (1986). Severity of fever in
influenza : differential pyrogenicity in ferrets exhibited by H1N1 and
H3N2 strains of differing virulence. Journal of General Virology 67,
419–425.
Cohen, J. J., Duke, R. C., Fadok, V. A. & Sellins, K. S. (1992). Apoptosis
and programmed cell death in immunity. Annual Review of Immunology
10, 267–293.
Dinarello, C. A. (1992). Role of interleukin-1 in infectious diseases.
Immunological Reviews 127, 119–146.
Farone, A. L., Frevert, C. W., Farone, M. B., Morin, M. J., Fields, B. N.,
Paulauskis, J. D. & Kobzik, L. (1996). Serotype-dependent induction of
pulmonary neutrophilia and inflammatory cytokine gene expression by
reovirus. Journal of Virology 70, 7079–7084.
Fesq, H., Bacher, M., Nain, M. & Gemsa, D. (1994). Programmed celldeath (apoptosis) in human monocytes infected by influenza A virus.
Immunobiology 190, 175–182.
CICI
glycosylation. Journal of Virology 70, 663–666.
Pickering, J. M., Smith, H. & Sweet, C. (1992). Influenza virus
pyrogenicity : central role of structural orientation of virion components
and involvement of viral lipid and glycoproteins. Journal of General
Virology 73, 1345–1354.
Rodgers, B. C. & Mims, C. A. (1982). Influenza virus replication in
human alveolar macrophages. Journal of Medical Virology 9, 177–184.
Rodgers, S. E., Barton, E. S., Oberhaus, S. M., Pike, B., Gibson, C. A.,
Tyler, K. L. & Dermody, T. S. (1997). Reovirus-induced apoptosis of
MDCK cells is not linked to viral yield and is blocked by Bcl-2. Journal of
Virology 71, 2540–2546.
Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning : A
Laboratory Manual, 2nd edn. Cold Spring Harbor, NY : Cold Spring
Harbor Laboratory.
Savill, J., Fadok, V., Henson, P. & Haslett, C. (1993). Phagocyte
recognition of cells undergoing apoptosis. Immunology Today 14,
131–136.
Siebenlist, U., Franzoso, G. & Brown, K. (1994). Structure, regulation
and function of NF-κB. Annual Review of Cell Biology 10, 405–455.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 05:34:24
Influenza virus cytotoxicity and apoptosis
Smith, H. & Sweet, C. (1988). Lessons for human influenza from
Tinsley, C. M., Sweet, C., Coates, D. M. & Smith, H. (1987). The local
origin of fever in influenza : differential production of endogenous
pyrogen by nasal inflammatory cells of ferrets exhibiting different levels
of fever. FEMS Microbiology Letters 42, 103–108.
pathogenicity studies with ferrets. Reviews of Infectious Diseases 10,
56–75.
Sweet, C. & Smith, H. (1980). Pathogenicity of influenza virus.
Microbiological Reviews 44, 303–330.
Sweet, C., Stephen, J. & Smith, H. (1974). Purification of influenza
viruses using disulphide-linked immunosorbents derived from rabbit
antibody. Immunochemistry 11, 295–304.
Differences in the capacity of reovirus strains to induce apoptosis are
determined by the viral attachment protein σ1. Journal of Virology 69,
6972–6979.
Sweet, C., Bird, R. A., Cavanagh, D., Toms, G. L., Collie, M. H. & Smith,
H. (1979). The local origin of the febrile response induced in ferrets
Wada, N., Matsumura, M., Oba, Y., Kobayashi, N., Takizawa, T. &
Nakanishi, Y. (1995). Transcription stimulation of the Fas-encoding
during respiratory infection with a virulent influenza virus. British Journal
of Experimental Pathology 60, 300–308.
Takizawa, T., Matsukawa, S., Higuchi, Y., Nakamura, S., Nakanishi, Y.
& Fukuda, R. (1993). Induction of programmed cell death (apoptosis) by
influenza virus infection in tissue culture cells. Journal of General Virology
74, 2347–2355.
Teodoro, J. G. & Branton, P. E. (1997). Regulation of apoptosis by viral
gene products. Journal of Virology 71, 1739–1746.
Thompson, W. R. (1947). The use of moving averages and interpolation
to estimate median effective dose. Bacteriological Reviews 11, 115–147.
Tyler, K. L., Squier, M. K. T., Rodgers, S. E., Schneider, B. E., Oberhaus, S. M., Grdina, T. A., Cohen, J. J. & Dermody, T. S. (1995).
gene by nuclear factor for interleukin-6 expression upon influenza virus
infection. Journal of Biological Chemistry 270, 18007–18012.
Watanabe, W., Sudo, K., Asawa, S., Konno, K., Yokota, T. & Shigeta, S.
(1995). Use of lactate-dehydrogenase to evaluate the antiviral activity
against influenza A virus. Journal of Virological Methods 51, 185–191.
Wright, P. F., Thompson, J. & Karzon, D. T. (1980). Differing virulence
of H1N1 and H3N2 influenza strains. American Journal of Epidemiology
112, 814–819.
Received 24 March 1997 ; Accepted 6 June 1997
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 05:34:24
CICJ