FEMS Immunology and Medical Microbiology 23 (1999) 331^341
Binding of bacteria to HEp-2 cells infected with in£uenza A virus
Omar R. El Ahmer, Muhammed W. Raza, Marie M. Ogilvie, Donald M. Weir,
C. Caroline Blackwell *
Department of Medical Microbiology, University of Edinburgh, Teviot Place, Edinburgh EN8 9AG, UK
Received 20 October 1998; received in revised form 15 December 1998; accepted 16 December 1998
Abstract
Epidemiological studies indicate influenza virus infection increases susceptibility to bacterial respiratory pathogens and to
meningococcal disease. Because density of colonisation is an important factor in the development of bacterial disease, the
objectives of the study were to use flow cytometry methods for assessment of bacterial binding and detection of cell surface
antigens to determine : (1) if HEp-2 cells infected with human influenza A virus bind greater numbers of bacteria than
uninfected cells; (2) if influenza infection alters expression of cell surface antigens which act as receptors for bacterial binding;
(3) if neuraminidase affects binding of bacteria to HEp-2 cells. There was significantly increased binding of all isolates tested
regardless of surface antigen characteristics. There were no significant differences between virus-infected and -uninfected Hep-2
cells in binding of monoclonal antibodies to Lewisb , Lewisx or H type 2. There were significant increases in binding of
monoclonal antibodies to CD14 (P 6 0.05) and CD18 (P 6 0.01). Treatment of cells with monoclonal antibodies significantly
reduced binding of Neisseria meningitidis strain C:2b:P1.2, CD14 (P 6 0.001) and CD18 (P 6 0.001). No reduction in binding
of a strain of Streptococcus pneumoniae (12F) was observed in these experiments. Neuraminidase treatment of HEp-2 cells
increased binding of monoclonal antibodies to CD14 (P 6 0.01) and CD18 (P 6 0.01). In three experiments, the increase in
binding of meningococcal strain C:2b:P1.2 to neuraminidase-treated cells was not significant, but binding of Staphylococcus
aureus strain NCTC 10655 was significant (P 6 0.05). z 1999 Federation of European Microbiological Societies. Published
by Elsevier Science B.V. All rights reserved.
Keywords : In£uenza A virus ; Meningococcus; Neisseria lactamica ; Pneumococcus ; Bordetella pertussis; Haemophilus in£uenzae ;
Moraxella catarrhalis; Staphylococcus aureus
1. Introduction
Secondary bacterial infection is the most common,
life-threatening complication of in£uenza infection
[1]. In£uenza A virus infections in children are often
complicated by acute otitis media (AOM) [2^4]. It
has been suggested that viral respiratory tract infec* Corresponding author. Tel.: +44 (131) 650-3170;
Fax: +44 (131) 650-6531; E-mail: [email protected]
tions might predispose to meningococcal disease [5].
Epidemiological studies suggest that in£uenza virus
infection increases the risk of subsequent meningococcal infection [6,7].
While many studies have been carried out on binding of bacteria in animal models [8], there were no
reports on the e¡ect of in£uenza virus infection on
binding of bacteria to human respiratory epithelial
cell lines. These studies were initiated to determine if,
as we had observed for respiratory syncytial virus
0928-8244 / 99 / $20.00 ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.
PII: S 0 9 2 8 - 8 2 4 4 ( 9 8 ) 0 0 1 5 3 - 9
FEMSIM 988 2-4-99
332
O.R. El Ahmer et al. / FEMS Immunology and Medical Microbiology 23 (1999) 331^341
(RSV) [9], infection of epithelial cells with in£uenza
A might enhance bacterial binding. The objectives of
this part of the study were: (1) to test the hypothesis
that compared with uninfected HEp-2 cells, these
cells infected with human in£uenza A virus bind
greater numbers of bacteria associated with meningitis or species associated with pneumonia, otitis media and exacerbation of chronic bronchitis; (2) to
determine if as reported for RSV-infected cells [10],
there was an increased expression of native cell surface antigens which can act as receptors for bacteria;
and (3) to determine if neuraminidase a¡ects bacterial binding to HEp-2 cells.
2. Materials and methods
2.1. Growth of HEp-2 cell line
Con£uent monolayers of HEp-2 cells were obtained in tissue culture £asks by culture in growth
medium (GM) which consisted of Eagle's minimal
essential medium (Gibco) supplemented with foetal
calf serum (FCS) (Gibco) (10%), NaHCO3 (0.85 g
l31 ), L-glutamine (2 mM), streptomycin (200 Wg
l31 ) and penicillin (100 IU ml31 ). Maintenance medium (MM) had the same components as GM, but
FCS was reduced to 1%.
2.2. Growth of in£uenza virus
The MDCK cell line from the European Collection of Animal Cell Cultures (ECACC), CAMR, Salisbury, UK) was used in the studies for the growth of
in£uenza virus. Con£uent monolayers of MDCK
cells were obtained in tissue culture £asks by incubation in GM with 10% FCS. The monolayer was
rinsed twice with Dulbecco's phosphate-bu¡ered saline A (DPBS). In£uenza virus type A, a recent human isolate obtained from the Clinical Virology
Laboratory, University of Edinburgh, was adsorbed
to cells for 60 min at 37³C. The inoculum was removed and replaced with serum free MM supplemented with 2 Wg ml31 trypsin (crystalline, Sigma)
which is required for virus preparation in this cell
line. The cells were incubated at 37³C for 3 days.
About two-thirds of the medium in the £ask was
discarded and the £ask was frozen at 370³C and
thawed to lyse the cells. The suspension was centrifuged at 800Ug for 10 min. Aliquots of the supernatant were stored at 370³C.
Samples of the supernatant £uids were tested for
the presence of in£uenza virus by the standard haemagglutination technique with guinea pig erythrocytes. The virus preparation with the highest haemagglutination titre was divided into 2-ml stock
aliquots and frozen at 370³C until used.
2.3. Detection of in£uenza-infected cells
Coverslips with con£uent monolayers of HEp-2
cells in a 24-well plate were prepared and infected
with in£uenza virus from a stock aliquot (200 Wl)
at di¡erent dilutions (1/2, 1/5 and 1/10) for 1 h at
37³C, and the inoculum was replaced with 1 ml serum free MM. The cells were incubated at 37³C for
16^24 h at which time they were washed and ¢xed
with acetone for 5 min. Fixed coverslips were
mounted (cells uppermost) on slides. The percentage
of infected cells was determined in a one-step direct
immuno£uorescence assay by incubating the coverslips with 25 Wl monoclonal antibodies speci¢c for
the virus conjugated to £uorescein isothiocyanate
(FITC) (Imagen, Dako) as per the manufacturer's
instructions. The slides were examined on an ultraviolet microscope (Microstar IV, Reichert, Bu¡alo,
New York, USA) for the proportion of £uorescent
cells per 100 cells counted. Dilutions of stock aliquots for which s 75% infection of HEp-2 cells
was observed were used for inoculation in the binding experiments.
2.4. Bacteria
2.4.1. Gram-negative isolates
Strains of Neisseria meningitidis were obtained
from several sources: C:2b:P1.2 was provided by
Dr. R.J. Fallon the Meningococcal Reference Laboratory,
Ruchill
Hospital,
Glasgow,
UK;
NG:2b:P1.10, and B:2b:P1.10 was obtained from
Dr. G. Tzanakaki, National School of Public Health,
Athens, Greece. The lipooligosaccharide immunotype strains were obtained from Dr. W.D. Zollinger,
Walter Reed Army Medical Institute, Washington,
DC (Table 1). One strain of Neisseria lactamica
(LO1) and two isolates of Moraxella catarrhalis
FEMSIM 988 2-4-99
O.R. El Ahmer et al. / FEMS Immunology and Medical Microbiology 23 (1999) 331^341
Table 1
Standard lipooligosaccharide strains of Neisseria meningitidis
Strains
Serogroup
Serotype
Subtype
Immunotype
L1
L2
L3
L4
L5
L6
L7
L8
L9
L10
L11
Ll2
C
C
B
C
B
B
B
B
A
A
A
A
NT
2c
2a
11
4
5
9
8,19
21
21
21
21
P1.2
P1.1
P1.5,2
P1.16
P1.NT
P1.7,1
P1.7,1
P1.7,1
P1.10
P1.?
P1.10
P1.NT
L1,8
L2
L(3,7,9)
L4
L5
L6
L(3,7,9)
L(3,7,9),8
L(3,7,9)
L10
L11
L12
(MC1 and MC2) were obtained from Dr. H. Young,
Department of Medical Microbiology, Edinburgh
University; MC1 grew on Modi¢ed New York
City Medium (MNYC) containing antibiotics selective for the pathogenic neisseriae. MC2 did not grow
on this medium. There was one strain of Haemophilus in£uenzae type b and two non-typable isolates
obtained from local infants. Two Bordetella pertussis
strains, 8002 (¢mbriate, type 1,2) and 250815 (non¢mbriate, type 1,3), were supplied by Dr. N.W. Preston, Department of Microbiology, University of
Manchester.
2.4.2. Gram-positive isolates
Three strains of Streptococcus pneumoniae associated with meningitis (types 7F, 12F, and 18C) were
kindly provided by Dr. L.E. Smart, The Meningococcal and Pneumococcal Reference Laboratory,
Stobhill Hospital, Glasgow, UK. Five isolates of S.
pneumoniae (types 42, 23, 10, 6, and 33) were obtained from respiratory specimens by Clinical Bacteriology, University of Edinburgh. Staphylococcus
aureus NCTC 10655 was obtained from the National
Collection of Type Cultures.
2.4.3. Culture media and conditions
N. meningitidis, N. lactamica and MC1 were
grown on MNYC. The antibiotic-sensitive isolate
(MC2) and H. in£uenzae were grown on boiled
blood agar (BBA). B. pertussis was grown on charcoal agar (Difco, UK). S. aureus was cultured on
nutrient agar. Pneumococci were grown on Colum-
333
bia blood agar with horse blood (BA). Except for B.
pertussis which required 3^5 days incubation, each
strain was grown overnight at 37³C in a humidi¢ed
atmosphere with 10% CO2 .
2.5. Bacterial binding assay
Overnight monolayer cultures of HEp-2 cells in
culture £asks (75 cm3 ) were infected with in£uenza
A (5 ml of a 1:5 dilution of the virus suspension) in
serum-free MM adsorbed for 1 h. The £uid was replaced with 25 ml serum free MM and incubated
overnight at 37³C. The monolayers were rinsed twice
with sterile DPBS+A and harvested by adding 5 ml
of EDTA (0.05% v/v) per £ask at 37³C for 5^10 min.
MM (5 ml) was added to the cells to counteract
EDTA activity. The cells were centrifuged at
460Ug for 7 min and resuspended in MM without
antibiotics, counted and adjusted to 1U106 cells
ml31 as described in Section 2.1.
Binding studies with FITC-labelled bacteria were
carried out and analysed by £ow cytometry as described previously [9,11].
2.6. Detection of host cell surface antigens on HEp-2
cells infected with in£uenza virus
The monoclonal antibodies, the dilutions of each
used in these experiments and their sources are listed
in Table 2. Uninfected and in£uenza-infected HEp-2
cells (200 Wl, 1U106 ml31 ) were incubated for 60 min
at 37³C with 200 Wl of antibodies to the host antigens. The cells were washed twice with PBS and incubated for 30 min in an orbital shaker at 37³C with
FITC-labelled secondary antibodies to detect binding of their respective primary antibody:rabbit antimouse IgM (Sigma) diluted 1 in 100 in PBS; rabbit
Table 2
Monoclonal antibodies to host cell antigens
Antibodya
b
Lewis
CD14
CD15
CD18
H type 2
a
Source
a
SAPU
SAPU
SAPU
Serotec
Serotec
Isotype
Dilution
IgM
IgG1
IgM
IgG2b
IgM
1/10
1/2
1/20
1/20
1/10
Scottish Antibody Production Unit.
FEMSIM 988 2-4-99
334
O.R. El Ahmer et al. / FEMS Immunology and Medical Microbiology 23 (1999) 331^341
anti-mouse IgG (Sigma) diluted 1 in 200 in PBS; or
rabbit anti-rat IgG (Serotec) diluted 1 in 100 in PBS.
The cells were washed twice with PBS by centrifugation, resuspended in 1% bu¡ered paraformaldehyde and analysed by £ow cytometry within 48 h
of preparation. Controls included cells incubated
without either primary or secondary FITC-labelled
antibodies and cells incubated only with the secondary antibody.
2.7. Inhibition of bacterial binding by pre-treatment of
in£uenza A-infected HEp-2 cells with antibodies
to host cell antigens
In£uenza A-infected HEp-2 cells were treated with
monoclonal antibody to CD14 or CD18. The attachment assays with FITC-labelled Gram-negative species (N. meningitidis, M. catarrhalis, H. in£uenzae
type b) or Gram-positive species (S. pneumoniae
and S. aureus), were performed and analysed as described previously [9,11].
2.8. The e¡ect of pre-treatment of HEp-2 cells with
neuraminidase on bacterial binding
HEp-2 cell (200 Wl, 1U106 cells ml31 ) were incubated with neuraminidase type III from Vibrio cholerae (Sigma) (30 Wl, 1 unit ml31 ) for 60 min at 37³C.
The cells were washed twice with PBS. The FITClabelled N. meningitidis strain (C:2b:P1.2) (200 bac-
teria:cell) or S. aureus strain (NCTC 10655) (200
bacteria:cell) were incubated with equal volumes of
HEp-2 cells and HEp-2 cells pre-treated with neuraminidase at 37³C for 30 min. The cells were washed
three times at 460Ug with PBS and analysed by £ow
cytometry.
2.9. The e¡ect of pre-treatment of HEp-2 cell with
neuraminidase on detection of host cell surface
antigens
HEp-2 cell (200 Wl, 1U106 cells ml31 ) were incubated with neuraminidase as described in Section 2.8.
The cells were washed twice with PBS, incubated for
60 min at 37³C with the same dilutions of antibodies
as in Section 2.6 and detection of the bound antibodies carried out as described above.
2.10. Analysis of samples by £ow cytometry
For each sample analysed, two parameters were
obtained: (1) the percentage of cells in the population that showed £uorescence levels higher than the
background (cells incubated without FITC-labelled
bacteria or cells without primary antibody to detect
host cell antigens); (2) the mean levels of £uorescence in the population of cells with bound bacteria
or (monoclonal antibodies) re£ecting the mean number of bacteria (antibody molecules) bound per cell
on a logarithmic scale. The mean £uorescence chan-
Table 3
Results of six experiments on binding of Gram-negative species to uninfected HEp-2 cells and HEp-2 cells infected with in£uenza A virus
(200 bacteria :cell, except for B. pertussis, 500:cell)
Species
Type
Mean BI of
uninfected cells
Mean BI of
in£uenza A-infected cells
In£uenza A
as % uninfected cells (95% CI)
P
N. meningitidis
C :2b:P1.2
NG :2b:P1.10
B :2b:P1.10
27832
31756
28906
47300
47472
50794
552 (402^759)
490 (361^665)
656 (493^873)
0.001
0.001
0.001
H. in£uenzae
type b
87D (NT)
241D (NT)
37635
31810
30570
54588
52785
47900
540 (463^630)
166 (158^175)
157 (145^169)
0.001
0.001
0.001
N. lactamica
LO1
38972
53087
396 (336^467)
0.001
B. pertussis
type 1,2
type 1,3
39742
42937
56948
59330
387 (286^524)
367 (270^500)
0.001
0.001
FEMSIM 988 2-4-99
O.R. El Ahmer et al. / FEMS Immunology and Medical Microbiology 23 (1999) 331^341
335
Table 4
Results of six experiments on binding of M. catarrhalis isolates to uninfected HEp-2 cells and HEp-2 cells infected with in£uenza A virus
(400 bacteria :cell)
Isolates
BI uninfected cells
BI infected cells
Infected as % uninfected cells (95% CI)
P
MC1
MC2
MC3
MC4
MC5
MC6
MC7
MC8
37419
39743
47743
49491
54145
41145
54416
38276
52442
60986
56143
68211
66768
62384
67595
63355
288
421
197
321
335
323
314
442
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
(261^318)
(354^500)
(165^234)
(258^399)
(299^375)
(260^402)
(252^392)
(256^764)
nel values for positive cells were obtained from a
conversion table of log £uorescence supplied by the
manufacturer (Coulter).
A binding index (BI) for each sample in the study
was calculated by multiplying the percentage of positive cells and the mean £uorescence of positive cells.
In this study, the data were presented as binding
indices to re£ect the £uorescence of the whole population of cells examined as the virus infection might
a¡ect the percentage of cells binding bacteria (antibodies) and the mean numbers of bacteria (antibodies) bound.
on paired t-tests applied to logarithms of the binding
indices.
2.11. Statistical methods
3.2. Attachment of Gram-negative or Gram-positive
species to in£uenza A-infected HEp-2 cells
The data were assessed by estimate of relative
binding of the bacteria to in£uenza A-infected
HEp-2 cells compared with uninfected cells based
3. Results
3.1. Determination of in£uenza A-infected HEp-2
cells
The proportion of HEp-2 cells infected with in£uenza A at 24 h post inoculation was in excess of 75%
in each experiment.
Tables 3 and 4 summarise the results of six experiments to assess the e¡ect of infection with in£uenza
Table 5
Results of six experiments on binding of S. pneumoniae and S. aureus to uninfected HEp-2 cells and HEp-2 cells infected with in£uenza
A virus (200 bacteria :cell)
Species
Type/strain
Mean BI of
uninfected cells
Mean BI of
in£uenza A-infected cells
In£uenza A
as % uninfected cells (95% CI)
P
S. pneumoniaea
7F
12F
18C
25249
23415
15526
32099
33237
30267
233 (162^335)
254 (171^379)
309 (254^378)
0.01
0.01
0.001
S. pneumoniaeb
6
10
23
33
42
12675
14224
13206
29415
12078
24442
31165
28681
44974
28551
275
289
263
213
262
(228^333)
(254^329)
(238^290)
(200^227)
(236^290)
0.001
0.001
0.001
0.001
0.001
S. aureus
NCTC 10655
32756
45330
362 (292^450)
0.001
a
b
Meningitis strains.
Respiratory isolates.
FEMSIM 988 2-4-99
336
O.R. El Ahmer et al. / FEMS Immunology and Medical Microbiology 23 (1999) 331^341
Table 6
Results of ¢ve experiments on binding of N. meningitidis LOS immunotype strains to uninfected HEp-2 cells and HEp-2 cells infected
with in£uenza A virus (400 bacteria:cell)
Strains
Phenotype
L1
L2
L3
L4
L5
L6
L7
L8
L9
L10
L11
L12
C :NT:P1.2
C :2C:P1.1
B :2a:P1.5,2
C :11:P1.16
B :4:NT
B :5:P1.7,1
B :9:P1.7,1
B :8,19:P1.7,1
A:21:P1.10
A:
A:21:P1.10
A:21:NT
L1,8
L2
L(3,7,9)
L4
L5
L6
L(3,7,9)
L(3,7,9),8
L(3,7,9)
L10
L11
L12
Mean BI of
uninfected cells
Mean BI of
in£uenza A-infected
cells
In£uenza A
as % uninfected cells
(95% CI)
P
20071
15232
19477
17592
16825
19326
21243
26127
20828
21095
19610
20195
40747
38222
49190
37622
38237
39217
41701
45959
43547
37424
36935
35556
360
471
716
318
412
407
414
372
691
322
376
327
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.01
0.01
0.001
0.01
0.01
A virus on binding to HEp-2 cells by Neisseria, Haemophilus, Moraxella, and Bordetella species. Each
isolate was tested at a ratio of 200 bacteria per cell
except for M. catarrhalis (400 bacteria:cell) and B.
pertussis (500 bacteria:cell). All the bacterial species
showed enhanced binding to cells infected with in£uenza A compared with uninfected HEp-2 cells.
Binding of S. aureus, S. pneumoniae strains associated with meningitis and pneumococcal strains associated with respiratory infections to uninfected and
in£uenza-infected cells was assessed in six experiments in which ratios of 200 bacteria per cell were
used. In£uenza virus-infected HEp-2 cells had significantly higher binding indices compared with binding
indices for uninfected cells (Table 5).
3.3. Binding of N. meningitidis LOS immunotype
strains to in£uenza A-infected HEp-2 cells
Previous studies with RSV serotypes A and B
(298^436)
(328^531)
(493^1070)
(163^384)
(277^615)
(282^587)
(303^564)
(295^471)
(482^991)
(224^462)
(231^612)
(220^485)
found no di¡erence in enhancement of binding associated with serogroup, serotype or subtype of the
meningococcal strains tested. Strains with di¡erent
LOS immunotype antigen were assessed for binding
to determine if the carbohydrate composition of the
endotoxin a¡ected binding to virus-infected cells. In
¢ve experiments, each of 12 immunotype strains exhibited signi¢cantly enhanced binding to cells infected with in£uenza A virus compared to uninfected
cells (Table 6).
3.4. Expression of cell surface antigens on HEp-2
cells infected with in£uenza virus
In three experiments, the binding indices for
monoclonal antibodies directed towards cell surface
antigens indicate there was no signi¢cant di¡erence
in binding of monoclonal antibodies to Lewisb , Lewisx or H type 2 to HEp-2 cells or HEp-2 cells infected
with in£uenza virus. There were signi¢cant increases
Table 7
Binding indices for monoclonal antibodies to host cell surface antigens on HEp-2 cells and in£uenza A-infected HEp-2 cells
Antibodies
Mean BI of uninfected cells
Mean BI of in£uenza A-infected cells
95% CI
P
Lewisb
Lewisx (CD15)
H type 2
CD14
CD18
1782
1241
790
1933
2173
1496
1143
680
10984
12901
64^111
69^123
67^126
182^2188
514^684
NS
NS
NS
0.05
0.001
FEMSIM 988 2-4-99
O.R. El Ahmer et al. / FEMS Immunology and Medical Microbiology 23 (1999) 331^341
337
Fig. 1. The e¡ect of pre-treatment of in£uenza A-infected HEp-2 cells with monoclonal antibody to CD14 or CD18 on binding of Gramnegative species.
in binding of anti-CD14 (P 6 0.05, 95% CI 182^
2188) and anti-CD18 (P 6 0.001, 95% CI 514^684)
monoclonal antibodies to HEp-2 cells infected with
in£uenza virus compared to uninfected HEp-2 cells
(Table 7).
3.5. Inhibition of bacterial binding by anti-CD14 or
anti-CD18
In four experiments, binding of N. meningitidis
(C:2b:P1.2), M. catarrhalis strains MC1 and MC2
and the type b H. in£uenzae isolate to in£uenza Ainfected HEp-2 cells was inhibited by anti-CD14 or
anti-CD18. The results showed signi¢cant reduction
in binding by anti-CD14 for each species tested: N.
meningitidis (P 6 0.001, 95% CI 55^62); M. catarrhalis (MC1) (P 6 0.001, 95% CI 47^61); MC2
(P 6 0.01, 95% CI 52^69); and H. in£uenzae type b
(P 6 0.001, 95% CI 43^52). A similar pattern was
observed for anti-CD18: N. meningitidis (P 6 0.001,
95% CI 45^53); M. catarrhalis (MC1) (P 6 0.01, 95%
CI 37^62); MC2 (P 6 0.001, 95% CI 50^64); and H.
in£uenzae type b (P 6 0.001, 95% CI 39^45) (Fig. 1).
With Gram-positive species, no signi¢cant reduction in binding was observed. Binding indices for S.
aureus to in£uenza A-infected cells treated with antiCD14 (95% CI 87^106) or anti-CD18 (95% 90^106)
showed no di¡erences compared with untreated cells.
A similar pattern was observed with S. pneumoniae
(12F) with anti-CD14 (95% CI 95^107) or anti-CD18
(95% CI 96^108).
3.6. The e¡ects of neuraminidase treatment on
bacterial binding
In three experiments, the binding of S. aureus
strain (NCTC 10655) and N. meningitidis strain
(C:2b:P1.2) to HEp-2 cells pre-treated with neuraminidase were compared with their binding indices
to untreated HEp-2. The statistical analysis showed
the increases observed were not signi¢cant for N.
meningitidis strain C:2b:P1.2 (P s 0.05, 95% CI 99^
125), but marginally signi¢cant for S. aureus
(P 6 0.05, 95% CI 105^139).
3.7. The e¡ects of neuraminidase treatment on
detection of host cell surface antigens
In three experiments, the binding indices for
monoclonal antibodies directed to cell surface antigens indicate there was no signi¢cant di¡erence in
binding of monoclonal antibodies to Leb , Lex or H
type 2 to HEp-2 cells compared with HEp-2 cells
treated with neuraminidase. There were signi¢cant
increases in binding of anti-CD14 (P 6 0.01, 95%
CI 220^350) and anti-CD18 (P 6 0.01, 95% CI
FEMSIM 988 2-4-99
338
O.R. El Ahmer et al. / FEMS Immunology and Medical Microbiology 23 (1999) 331^341
256^359) monoclonal antibodies to HEp-2 cells
treated with neuraminidase compared to untreated
HEp-2 cells (Fig. 2).
4. Discussion
4.1. Binding of bacteria to in£uenza-infected HEp-2
cells
HEp-2 cells infected with human in£uenza A virus
showed increased adherence of respiratory tract bacteria associated with meningitis, pneumonia and exacerbation of chronic bronchitis. In£uenza virus-infected HEp-2 cells showed no changes in expression
of Lewis antigens or H type 2, but signi¢cantly increased binding of monoclonal antibodies for the cell
surface antigens CD14 and CD18 compared to uninfected HEp-2 cells. In contrast, RSV infection enhanced binding of Lewisx as well as to CD14 and
CD18 [10].
Pre-treatment of the virus-infected cells with antiCD14 or anti-CD18 signi¢cantly reduced binding of
Gram-negative species tested, but not the Gram-positive species. Treatment of uninfected HEp-2 cells
with neuraminidase enhanced bacterial binding.
Neuraminidase treatment signi¢cantly enhanced
binding of S. aureus. This was suggested to be due
to removal of sialic acid from sialyl Lewisx as Lewisx
has been shown to be a receptor for S. aureus [12];
however, there were no changes in binding of antiLewisx to neuraminidase-treated cells. The enhanced
binding observed with N. meningitidis and neuraminidase-treated cells was not signi¢cant. This could be
due to small number of experiments. Binding of antiCD14 and anti-CD18 to the neuraminidase-treated
cells was signi¢cantly enhanced.
4.2. Secondary respiratory infections and in£uenza
Several possible mechanisms for promoting bacterial superinfection by in£uenza have been proposed.
These include: decreased tracheal clearance function
[13,14]; impairment in pulmonary and systemic bacterial clearance mechanisms [15]. Studies by Sanford
et al. [16,17] and Davison and Sanford [18] suggest
the possibility that in£uenza infection predisposes
the host to bacterial colonisation, and subsequent
disease by enhancing the adherence of some bacteria
to virus-infected respiratory tract cells. This hypothesis is, in fact, consistent with clinical, epidemiological, and experimental evidence which indicates that
in£uenza A infection predisposes the lungs to secondary pneumonia due to S. aureus, S. pneumoniae,
and to a lesser extent, H. in£uenzae [14]. Vaccination
against in£uenza virus type A before an in£uenza A
Fig. 2. Binding of monoclonal antibodies to host cell surface antigens on HEp-2 cells treated with PBS or neuraminidase.
FEMSIM 988 2-4-99
O.R. El Ahmer et al. / FEMS Immunology and Medical Microbiology 23 (1999) 331^341
epidemic has been shown to be e¡ective in preventing the development of AOM in children aged between 1 and 3 years [19]. In the present study, each
of the bacterial strains associated with AOM or
pneumonia showed enhanced binding to in£uenzainfected cells. If in£uenza neuraminidase has the
same e¡ect as that used in our study, this might be
an additional mechanism by which the virus infection increases accessibility of host cell antigens to
which bacteria bind.
4.3. Bacterial meningitis and in£uenza
Many children with bacterial meningitis had signs
and symptoms of upper respiratory infection, suggesting that viral infection could lead to secondary
bacterial invasion [20]. While recent epidemiological
studies found little evidence to suggest that RSV is
associated with outbreaks of meningococcal infection
[21], there is evidence that in£uenza virus infection
might increase the risk of subsequent meningococcal
infection [6,7]. Both RSV and in£uenza virus enhanced binding of all meningococcal isolates tested,
regardless of serogroup, serotype, subtype or immunotype.
In£uenza virus is also known to a¡ect cellular immunity adversely [22]. It has been suggested that
in£uenza-induced coughing causes aerosolisation
and, therefore, increased transmission of N. meningitidis. All of these ¢ndings could, alone or in combination with increased density of colonisation by the
bacteria, account for the association between these
two infections.
4.4. Host cell receptors for bacteria and in£uenza
infection
CD14 and CD18 are receptors for Gram-negative
bacteria such as Escherichia coli [23,24]. In£uenza
virus-infected HEp-2 cells showed signi¢cantly increased binding of monoclonal antibodies for the
cell surface antigens CD14 and CD18 compared to
uninfected HEp-2 cells. This might partly explain
how virus infection contributes to enhanced binding
by upregulation of expression of host cell surface
antigens that can act as receptors for the Gram-negative bacteria.
339
Pre-treatment of in£uenza A virus-infected HEp-2
cells with monoclonal antibodies to CD14 or CD18
signi¢cantly inhibited binding of Gram-negative species tested (N. meningitidis, M. catarrhalis, and H.
in£uenzae type b). While it could be argued that
this e¡ect was due to steric hinderance, there was
no e¡ect on inhibition of binding of Gram-positive
species tested. This indicates that the endotoxins of
Gram-negative bacteria are involved in enhanced
binding of Gram-negative species observed with virus-infected cells.
4.5. E¡ect of neuraminidase
HEp-2 cells pre-treated with neuraminidase
showed signi¢cantly increased binding of S. aureus
compared to untreated HEp-2 cells, while the binding of meningococci was increased to neuraminidasetreated cells, the results were not signi¢cant. Di¡erences in binding to enzyme-treated cells and in£uenza-infected cells indicates neuraminidase contributes
to, but is not the only e¡ect on, cells that could
enhance binding.
Sialic acid is present on the structure of both
CD14 molecules [25] and CD18 molecules [26] and
neuraminidase treatment appears to make the epitope to which monoclonal reagents bind more accessible. Studies by Spooncer and colleagues [27] and
Fukuda and colleagues [28] had shown by carbohydrate analysis that two of the major carbohydrate
structures carried by neutrophil glycoproteins are
large highly branched N-linked structures: a polyfucosylated lactosaminoglycan and a sialylated fucosyllactosaminoglycan ; both of these carry the CD15
determinant. The sialylated derivative sLex is not
recognised by the CD15 antibodies unless the sialic
acid is removed. Removing of sialic acid from the
host cell antigens molecules by neuraminidase treatment might expose epitopes masked by sialic acid
residues; however, enhanced binding of monoclonal
anti CD15 was not observed.
We conclude that in addition to the various mechanisms already proposed to explain why in£uenza A
is a predisposing factor for bacterial disease of the
respiratory tract and meningitis, these studies indicate that in£uenza virus-infected epithelial cells bind
more bacteria. If a similar phenomenon occurs in
FEMSIM 988 2-4-99
340
O.R. El Ahmer et al. / FEMS Immunology and Medical Microbiology 23 (1999) 331^341
vivo, this could contribute to density of colonisation
which is an important factor in development of bacterial disease.
[11]
Acknowledgments
O.R.E. was supported by a grant from the Libyan
Ministry of Education. The work was supported in
part by grants from the Meningitis Association of
Scotland and Chest Heart and Stroke Scotland.
[12]
[13]
References
[14]
[1] Kilbourne, E.D. (1987) The epidemiology of in£uenza. In:
In£uenza (Kilbourne, E.D., Ed.), pp. 255^289. Plenum, New
York.
[2] Wright, P.F., Ross, K.B., Thompson, J. and Karzon, D.T.
(1977) In£uenza A infections in young children. Primary natural infection and protective e¤cacy of live-vaccine-induced
or naturally acquired immunity. New Engl. J. Med. 296,
829^834.
[3] Wright, P.F., Thompson, J. and Karzon, D.T. (1980) Di¡ering virulence of H1N1 and H3N2 in£uenza strains. Am. J.
Epidemiol. 112, 814^819.
[4] Henderson, F.W., Collier, A.M., Sanyal, M.A., Watkins,
J.M., Fairclough, D.L., Clyde, W.A. and Denny, F.W.
(1982) A longitudinal study of respiratory viruses and bacteria
in the etiology of acute otitis media with e¡usion. New Engl.
J. Med. 306, 1377^1383.
[5] Moore, P.S., Hierholzer, J., De Witt, W., Gount, K., Lippveld, T., Plikaytis, B. and Broome, C.V. (1990) Respiratory
viruses and mycoplasma as cofactors for epidemic group A
meningococcus meningitis. J. Am. Med. Assoc. 264, 1272^
1275.
[6] Harrison, L.H., Armstrong, C.W., Jenkins, S.R., Harmon,
M.W., Ajello, G.W., Miller, G.B. and Broome, C.V. (1991)
A cluster of meningococcal disease on a school bus following
epidemic in£uenza. Arch. Intern. Med. 151, 1005^1009.
[7] Cartwright, K.A.V., Jones, D.M., Smith, A.J., Stuart, J.M.,
Kaczmarski, E.B. and Palmer, S.R. (1991) In£uenza A and
meningococcal disease. Lancet 338, 554^557.
[8] Plotkowski, M.C., Puchelle, E., Beck, G., Jacquot, J. and
Hannoun, C. (1986) Adherence of type 1 Streptococcus pneumoniae to tracheal epithelial of mice infected with in£uenza A/
PR8 virus. Am. Rev. Resp. Dis. 134, 1040^1044.
[9] Raza, M.W., Ogilvie, M.M., Blackwell, C.C., Stewart, J., Elton, R.A. and Weir, D.M. (1993) E¡ect of respiratory syncytial virus infection on binding of Neisseria meningitidis and
Haemophilus in£uenzae type b to a human epithelial cell line
(HEp-2). Epidemiol. Infect. 110, 339^347.
[10] Saadi, A.T., Blackwell, C.C., Raza, M.W., James, V.S., Stew-
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
art, J., Elton, R.A. and Weir, D.M. (1993) Factors enhancing
adherence of toxigenic Staphylococcus aureus to epithelial cells
and their possible role in sudden infant death syndrome. Epidemiol. Infect. 100, 507^517.
El Ahmer, O.R., Raza, M.W., Ogilvie, M.M., Blackwell, C.C.,
Weir, D.M. and Elton, R.A. (1996) The e¡ect of respiratory
virus infection on expression of cell surface antigens associated with binding of potentially pathogenic bacteria. In: Toward Anti-Adhesin Therapy of Microbial Diseases (Ofek, I.
and Kahane, I., Eds.), pp. 169^177. Plenum, New York.
Saadi, A.T., Blackwell, C.C., Essery, S.D. et al. (1996) Developmental and environmental factors that enhance binding of
Bordetella pertussis to human epithelial cells in relation to
sudden infant death syndrome (SIDS). FEMS Immunol.
Med. Microbiol. 16, 51^59.
Sweet, C. and Smith, H. (1980) Pathogenicity of in£uenza
virus. Microbiol. Rev. 44, 303^330.
Nugent, K.M. and Pesanti, E.L. (1983) Tracheal function during in£uenza infection. Infect. Immun. 42, 1102^1108.
Warshauer, D., Goldstein, E., Akers, T., Lippert, W. and
Kim, M. (1977) E¡ect of in£uenza viral infection on the ingestion and killing of bacterial by alveolar macrophage. Am.
Rev. Resp. Dis. 115, 269^277.
Sanford, B.A., Shelokov, A. and Ramsay, M.A. (1978) Bacterial adherence to virus infected cells: a cell culture mode of
bacterial superinfection. J. Infect. Dis. 137, 176^181.
Sanford, B.A., Smith, N., Shelokov, A. and Ramsay M.A.
(1979) Adherence of group B streptococci and human erythrocytes to in£uenza A virus-infected MDCK cells. Proc. Soc.
Exp. Biol. Med. 160, 226^232.
Davison, V.E. and Sanford, B.A. (1981) Adherence of Staphylococcus aureus to in£uenza A virus infected Madin^Darby
canine kidney cell cultures. Infect. Immun. 32, 118^126.
Heikkinen, T., Ruuskanen, O., Waris, M., Ziegler, T., Arola,
M. and Halonen, P. (1991) In£uenza vaccination in the prevention of acute otitis media in children. Am. J. Dis. Child.
145, 445^448.
Fulginiti, V.A. and Sieber, O.F., Jr. (1975) Acute bacterial
meningitis. In: Textbook of Pediatrics, 10th edn. (McKay,
R.J. and Nelson, W.E., Eds.), pp. 583^587. W.B. Saunders,
Philadelphia, PA.
Stuart, J.M., Cartwright, K.A.V. and Andrews, N.J. (1996)
Respiratory syncytial virus infection and meningococcal disease. Epidemiol. Infect. 117, 107^111.
Jakab, G.J. (1982) Immune impairment of alveolar macrophage phagocytosis during in£uenza virus pneumonia. Am.
Rev. Resp. Dis. 126, 778^782.
Wright, S.D. and Jong, M.T. (1986) Adhesin-promoting receptors on human macrophages recognize Escherichia coli by
binding to lipopolysaccharide. J. Exp. Med. 164, 1876^1888.
Wright, A.L., Taussing, L.M., Ray, C.G., Harrison, H.R. and
Holberg, C.J. (1989) The Tucson Children's Respiratory
Study, II: lower respiratory tract illness in the ¢rst year of
life. Am. J. Epidemiol. 129, 1232^1246.
Stelter, F., P¢ster, M., Bernheiden, M., Jack, R.S., Bu£er, P.,
Engelmann, H. and Schutt C. (1996) The myeloid di¡erentia-
FEMSIM 988 2-4-99
O.R. El Ahmer et al. / FEMS Immunology and Medical Microbiology 23 (1999) 331^341
tion antigen CD14 is N- and O-glycosylated contribution of
N-linked glycosylation to di¡erent soluble CD14 isoforms.
Eur. J. Biochem. 236, 457^464.
[26] Kerr, M.A. and Stocks, S.C. (1992) The role of CD15-(Lex )related carbohydrates in neutrophil adhesion. Histochem. J.
24, 811^826.
[27] Spooncer, E., Fukuda, M., Klock, J.C., Oates, J.E. and Dell,
341
A. (1984) Isolation and characterization of polyfucosylated
lactosaminoglycan from human granulocytes. J. Biol. Chem.
259, 4792^4801.
[28] Fukuda, M., Spooncer, E., Oates, J.E., Dell, A. and Klock,
J.C. (1984) Structure of sialylated fucosyl lactosaminoglycan
isolated from human granulocytes. J. Biol. Chem. 259, 925^
935.
FEMSIM 988 2-4-99