Journal of Medical Virology 78:229–242 (2006)
Dengue Virus Infection of Human Microvascular
Endothelial Cells From Different Vascular
Beds Promotes Both Common and Specific
Functional Changes
Christophe N. Peyrefitte,1 Boris Pastorino,1 Georges E. Grau,2 J. Lou,2 Hugues Tolou,1
and Patricia Couissinier-Paris1*
1
Unite´ de virologie tropicale, Institut de Me´decine Tropicale du Service de Sante´ des Arme´es, Marseille, France
2
Unite´ des Rickettsies, CNRS UMR, Faculte´ de Me´decine, IFR 48,Universite´ de la Me´diterrane´e, Marseille, France
Dengue shock syndrome (DSS), the major life
threatening outcome of severe dengue disease,
which occurs in some patients in the course of
dengue infection, is the consequence of plasma
leakage in the microvascular territories. Data
from clinical and in vitro studies suggest that an
inadequate immunological response is partly
responsible for the pathophysiology of DSS, but
few is known concerning the consequences of
direct infection of endothelial cells by dengue
virus per se. In this study, an attempt was made to
study the response of two microvascular human
cell lines originating, respectively, from liver and
dermis to infection by a dengue type 2 virus, by
analyzing the virus-induced modulation of func-
tional markers. It is shown that the two micro-
vascular cell lines exhibit both common and
specific behaviors upon infection. In particular,
LSEC and HMEC-1 replicate efficiently the low-
passage virus and respond to infection by over-
producing inflammatory mediators involved in
the cross talk with circulating immune cells.
However, direct infection modulates differently
the cell surface expression of molecules critically
involved in the interactions between endothelial
and inflammatory cells. ICAM-1 and HLA-I are up
regulated as a consequence of infection in LSEC
whereas direct infection results in downregula-
tion of ICAM-1 in HMEC-1. The present results
show that infection of human microvascular
cells by unadapted dengue virus results in both
common and specific activation patterns depend-
ing likely on the tissue origin of the cells, thus
suggesting that endothelia from different terri-
tories may contribute differently to the patho-
physiological events in the course of dengue
infection. J. Med. Virol. 78:229–242, 2006.
ß 2005 Wiley-Liss, Inc.
ß 2005 WILEY-LISS, INC.
KEY WORDS: virus; flavivirus; endothelium;
activation; plasma leakage
INTRODUCTION
Dengue viruses are arthropod-borne viruses belong-
ing to the Flaviviridae family which are responsible for
infections in humans in all intertropical regions of the
world [Guzman et al., 2004]. Dengue infection presents
through a broad spectrum of clinical pictures of which
mild acute febrile illness is the more frequent. However,
severe forms, that is dengue hemorrhage fever (DHF)
and dengue shock syndrome (DSS), the major life
threatening form of severe dengue disease, are reported
with an increasing frequency [Guzman et al., 2004].
Currently, only non-specific, symptomatic therapeutics
is available for the treatment of patients presenting with
these severe forms [Damonte et al., 2004].
The hemodynamic failure characterizing the dengue
shock syndrome (DSS) is consecutive to a transient
but massive plasma leakage in the microvasculature
[Srichaikul and Nimmannitya, 2000; Guzman and
Kouri, 2002]. Transitory endothelial dysfunctions, more
than cell destruction or irreversible alterations, likely
support the capillary leakage, as suggested by the rapid
Grant sponsor: Délégation Générale de l’Armement (DGA);
Grant number: 02 CO 013; Grant sponsor: Service de Santé des
Armées (SSA).
*Correspondence to: Dr. Patricia Couissinier-Paris, Institut de
Médecine Tropicale du Service de Santé des Armées, BP 46, Parc
du Pharo, 13007 Marseille, France.
E-mail: imtssa.vro@wanadoo.fr
Accepted 30 September 2005
DOI 10.1002/jmv.20532
Published online in Wiley InterScience
(www.interscience.wiley.com)
230
reversibility of DSS in patients under resuscitation
[Iyngkaran et al., 1995; Srichaikul and Nimmannitya,
2000], and the absence of destruction of endothelial cells
in post-mortem biopsies from patients who died from
DSS [Bhamarapravati et al., 1967; Sahaphong et al.,
1980]. Several evidences indicate that the immune
response to dengue virus plays a key role in the
pathophysiological cascade leading to plasma leakage
[Chen and Cosgriff, 2000; Lei et al., 2001; Guzman and
Kouri, 2002]. In particular, DHF and DSS have been
associated with high levels of inflammatory cytokines
in the serum of patients [Iyngkaran et al., 1995; Green
et al., 1999; Gagnon et al., 2002; Liu et al., 2002; Suharti
et al., 2003]. Moreover, in vitro studies have shown that
soluble mediators produced by monocytes or macro-
phages infected by dengue virus in different conditions,
activate endothelial cells in vitro [Anderson et al., 1997;
Carr et al., 2003]. Other immune-mediated mechanisms
involving cross-reactive auto-antibodies generated in
response to the dengue NS1 protein, could also contri-
bute to the pathogenesis of secondary dengue infection
by inducing limited apoptosis or cytolysis of endothelial
cells [Lin et al., 2003].
The possibility that dengue virus infection may affect
directly the functions of microvascular endothelial cells
is much debated. Indeed, no definitive conclusion can be
drawn from ex vivo studies due to the very limited
number of histological data produced from post-mortem
biopsies from DSS patients [Bhamarapravati et al.,
1967; Sahaphong et al., 1980; Jessie et al., 2004]. In the
absence of relevant animal models, in vitro infection of
endothelial cells represent an alternative approach for
the identification of cell alterations that could be impli-
cated in the pathological process leading to DSS.
The ability of dengue viruses to infect endothelial cells
in vitro has already been demonstrated by others using
different sources of endothelial cells and viruses
[Andrews et al., 1978; Bunyaratvej et al., 1997;
Avirutnan et al., 1998; Huang et al., 2000; Jessie et al.,
2004; Talavera et al., 2004]. However, endothelial cells
from large vessels, used most often in those studies, that
is primary HUVEC or ECV304 cell line, exhibit
important phenotypic and functional differences com-
pared to endothelial cells from microvascular territories
where plasma leakage largely occurs [Bender et al.,
1994; Mason et al., 1997; Lidington et al., 1999;
Murakami et al., 2001; Kieda et al., 2002; Chi et al.,
2003]. In particular, primary HUVEC have variable and
unstable phenotype [Klein et al., 1995; Vermot-Des-
roches et al., 1995], whereas the endothelial origin of the
large vessel-derived ECV 304 cell line is highly con-
troversial thus questioning its relevance as endothelial
cell model [Brown et al., 2000; Drexler et al., 2002]. In
this regard, those large vessels-derived endothelial cells
appear poorly relevant to the pathophysiology of dengue
shock syndrome. On other hand, the diversity of viruses
used in those previous studies, that is from wild-type to
highly adapted dengue viruses, may have introduced an
element of variability between the results previously
published, making their interpretation difficult.
J. Med. Virol. DOI 10.1002/jmv
Peyrefitte et al.
The present study aimed at examining different
aspects of the response of two human microvascular
endothelial cell lines to infection by a low-passage DEN-
2 virus strain representative of the new American
genotype associated with recent dengue outbreaks
[Letmeyer et al., 1999]. The two cell lines studied, that
is the liver sinusoidal endothelial cell line LSEC and the
dermal endothelial cell line HMEC-1, are endothelia
representative of tissues or organs known to be affected
in the course of dengue infection [Boonpucknavig et al.,
1979; Desruelles et al., 1997; Mohan et al., 2000; Wahid
et al., 2000; Pancharoen et al., 2002]. The study focused
on functional markers generally modulated in the
course of endothelial activation, that is adhesion mole-
cules and endothelial-derived inflammatory mediators
[Krishnaswamy et al., 1999; Martinez-Mier et al., 2001;
Galley and Webster, 2004], and more specifically on
markers altered in the course of DSS [Hober et al., 1993;
Raghupathy et al., 1998; Juffrie et al., 2001].
The present results show that direct infection of the
two microvascular cells lines LSEC and HMEC-1 with a
wild-type DEN-2 virus, triggers activation in both cell
lines. However, infection has different effects on those
two types of endothelial cells with regard to critical
markers involved in the cross talk between endothelial
and immune cells. These data suggest that vascular
endothelia activated by direct infection could contribute
differentially to the pathogenesis of DSS.
MATERIALS AND METHODS
In Vitro Virus Propagation
All work with infectious virus was carried out in a bio
safety-level 3 laboratory. The DEN-2 virus strain used
in this study, DEN2/H/IMTSSA-MART/98-703 (Gen-
Bank accession number AF 208496), was isolated in our
laboratory from a patient exhibiting a classical dengue
fever syndrome. The viral genome was sequenced
entirely [Tolou et al., 2000] and phylogenetic analysis
showed that this strain belonged to the new American
genotype which originate from Southeast Asian geno-
types [Letmeyer et al., 1999]. The batch of viral
inoculum used in this study was prepared by two
passages in C6/36 cells (from Aedes albopictus; ATCC
clone CRL 1660). Briefly, permissive insect cells were
grown at 288C in Leibowitz’s L15 medium (Biowhittaker,
Verniers, Belgium) supplemented with 1% final
L-glutamin (Biowhittaker), 5% final fetal calf serum
(FCS) (Biowhittaker) and 2% final tryptose phosphate
broth (Eurobio, Les Ulis, France). Infection was carried
out by seeding C6/36 cells in complete Leibowitz’s
medium without FCS and inoculating them with the
virus at a multiplicity of infection (MOI) of 1, for 1 hr at
288C. Complete medium containing 5% FCS was added
and cells were incubated further for 5 days. Cell culture
supernatants were collected and viral titers were
determined by plaque assay on Vero cells according to
the dilution method previously reported [Luria et al.,
1978]. Viral titres were estimated by the equation of
maximum likelihood [Kleczkowski, 1968].
Response of Endothelial Cells to Dengue Virus
LSEC and HMEC-1 Microvascular Cell Lines
LSEC and HMEC-1 cell lines have been established
and characterized by Salmon et al. [2000] and by Ades
et al. [1992], respectively. The LSEC line was estab-
lished from primary liver sinusoidal endothelial cells by
transfection using a plasmid containing both the human
telomerase gene and the SV40 large T sequence [Salmon
et al., 2000]. The HMEC-1 line was established by
transfection of primary endothelial cells isolated from
human dermis using a plasmid encoding the SV40 large
T [Ades et al., 1992a]. Both microvascular cell lines
exhibit endothelial properties from the primary cells
they had been derived from, that is typical morphology,
expression of von Willebrand factor, uptake of acety-
lated LDL, expression of adhesion molecules, ability to
form tube-like structures into Matrigel, and are stable
through a high number of in vitro passages regarding
endothelial properties [Ades et al., 1992; Salmon et al.,
2000].
Culture of Microvascular Cell Lines
The LSEC and HMEC-1 lines were cultured in flasks
or culture plates previously coated with 0.2% gelatin
(Sigma Aldrich, Saint Quentin Fallavier, France) in
PBS. LSEC were maintained in CMRL-1066 (Invitro-
gen/Life technologies, Cergy Pontoise, France) supple-
mented with 2 mM final L-glutamin (Cambrex) and 10%
FCS (Cambrex, Verviers, Belgium) while HMEC-1 were
cultured in EGM-2 MV medium (Cambrex) supplemen-
ted with 2 mM L-glutamin, 10% FCS, and commercial
single quots of recombinant human epidermal growth
factor and hydrocortisone.
Infection of Endothelial Cells
With Dengue Type 2 Virus
Despite their stability over a large number of passages
[Ades et al., 1992; Salmon et al., 2000], the HMEC-1 and
LSEC lines were used at constant passage number
1
for all experiments, that is passage 18
1 for HMEC-1
and passage 21
1 for LSEC.
Briefly, the LSEC and HMEC-1 lines were seeded in
6-well plates 48 hr before infection. LSEC were cultured
in CMRL1066 medium supplemented with 2 mM
glutamin and 10% heated FCS, while HMEC-1 were
deprivated of epidermal growth factor and hydrocorti-
sone and maintained in EGM-2 MV medium (Cambrex)
supplemented with only 2 mM L-glutamin and 10%
heated FCS. Two days after seeding, when cultures have
reached confluency, cells were infected with super-
natant of infected C6/36 cells at MOIs 1 and 0.1 for 2 hr
at þ378C, in a 5% CO2 atmosphere. Different control cell
cultures were included: untreated endothelial cells
(EC), EC incubated with supernatant from uninfected
C6/36, EC incubated with viral inoculum inactivated by
heating at 808C for 30 min [Avirutnan et al., 1998] or
with corresponding heat-treated (808C, 30 min) super-
natants from uninfected C6/36 cells. After 2 hr, the
inoculum were removed and all the cell monolayers
except the ‘‘untreated EC’’ control, were washed four
231
times with 8 ml per well of Dulbecco’s Phosphate
Buffered Saline (DPBS, Cambrex) supplemented with
5% final FCS. Finally, 4 ml of fresh medium was added in
all wells and cultures were further incubated at 378C. At
times indicated, supernatants and cells were harvested
for analysis of viral and cellular parameters. The
supernatants were centrifuged (5 min, 3,000g, at
þ208C), then either directly used for virus titration on
Vero cells as described before, or mixed (1 vol/2 vol) with
lysis buffer (Roche) and frozen at 208C for further viral
RNA extraction. Cell pellets to be analyzed for the
detection of viral RNA negative strands were directly
lysed in 300 ml of RLT-buffer (Qiagen, Courtaboeuf,
France), then frozen at 208C until further RNA
extraction.
Extraction of Viral RNA
Total dengue virus RNA was extracted from cell
culture supernatants or cell pellets using the High Pure
Viral RNA kit (Roche Diagnostics, Meylan, France) or
the RNeasy kit (Qiagen), respectively. RNAs extracted
from cell pellets were directly digested on column with
DNaseI. Total RNAs obtained from either 200 ml of
supernatants or from cell pellets were then subjected to
RT-PCR as described thereafter.
Quantification of Positive Strand Viral RNAs
Quantification of viral RNAs was carried out as
follows: after extraction as described before, the RNAs
were amplified using the Platinium quantitative RT-
PCR Thermoscript One-Step System (Invitrogen-Life
Technologies). Briefly, 2.5 ml of the RNAs were amplified
in a 20 ml reaction mix containing 3.6 pM of each of the
two PCR primers (F and R), 7 pM of dengue-specific
fluorogenic probe [Warrilow et al., 2002], 10 ml of 2X
Thermoscript Reaction mix, 1 ml of RNAse OUT and
0.5 ml of Thermoscript Plus/Platinium Taq Mix in a
Lightcycler thermocycler (Roche Diagnostics, Meylan,
France). PCR reactions were performed according to the
following protocol: after initial reverse transcription at
508C for 30 min and 5 min denaturation at 958C, cDNAs
were submitted to 45 amplification cycles with sequen-
tial steps at 958C for 15 sec and 608C for 60 sec.
Indirect Immunofluorescence Staining of
Dengue-Infected Cells
After being detached by brief exposure to a trypsin-
EDTA solution (Cambrex) and washed, endothelial cells
were seeded on 12-well fluorescence slides (VWR
Internationnal, Fontenay-sous-Bois, France) and fixed
with acetone for 30 min at 208C. Fixed cells were then
rinsed once with PBS containing 0.05% Tween 20, once
with PBS alone and then incubated 30 min at 378C in the
presence of a solution of BSA at 1% in PBS. A solution of
mouse monoclonal antibody specific for the E protein of
dengue virus (ab 9202, Abcam, Cambridge, UK) was
then added for 1 hr at 378C. Cells were then washed
twice in PBS and incubated for 1 hr at 378C in the pre-
J. Med. Virol. DOI 10.1002/jmv
232
sence of an Alexa fluor 488-conjugated goat anti-mouse
IgG polyclonal antibody (Molecular Probes, The Nether-
lands). After three washes in PBS, cells were finally
counterstained during 5 min at room temperature using
a solution of Evans blue in PBS. Slides were finally
mounted to be examined under an Olympus BH-2
fluorescence microscope. The mean percentage of infec-
ted cells was determined from at least three fields,
and represented the number of fluorescent cells out of
counterstained cells.
In Vitro Activation of Microvascular Endothelial
Cell Lines by Pro-Inflammatory Cytokines
Activation of LSEC and HMEC-1 by the inflammatory
cytokine TNF was carried out initially to determine the
basal and the cytokine-activated states of the two cell
lines in our culture conditions. Briefly, endothelial cells
grown at confluence in gelatin-coated 6-well plates were
cultured in their respective culture medium supple-
mented or not with 20 ng/ml of recombinant human TNF
(PeProtech, TEBU, France). Cultures were detached
at different times (24, 48, 72 hr) by brief exposure to a
commercial solution of trypsin-EDTA (Cambrex),
washed twice with ice-cold PBS containing 5% FCS
(Cambrex), and used for cell surface phenotyping as
described below.
Similarly, the ability of endothelial cell lines to
modulate cell expressionof HLA-I antigens, a marker
of interest in flavivirus infections [Kesson et al., 2002],
was assessed by exposing HMEC-1 and LSEC lines to
stimulation by recombinant human IFN-g (hIFN-g,
PeProtech, TEBU, France) at 500 and 2,000 U/ml.
HLA-I surface levels were analyzed by cytometry 3 days
after addition of hIFN-g, a time corresponding to the
peak of expression in response to this cytokine.
Flow Cytometry Analysis of Cell
Surface Activation Markers
All antibodies used were tested previously at different
dilutions to determine the optimal concentration for
these assays. Preliminary experiments were also desig-
ned to verify that the cell surface antigens studied were
insensitive to the trypsin-EDTA treatment in the condi-
tions used (data not shown).
Briefly, infected and control cells were harvested at
times indicated after infection. After detachment by
brief exposure of adherent cells to a commercial solution
of trypsin-EDTA, cells were washed twice with ice-cold
PBS containing 5% FCS, and incubated at þ48C for 1 hr
in the presence of FITC- or PE-conjugated mouse
monoclonal antibodies directed against the human cell
surface antigens of interest, diluted in a ‘‘staining
buffer’’ solution made of ice-cold PBS supplemented
with 5% of heat-inactivated human AB serum, 5% of
FCS and 0.01% of sodium azide. Antibodies used were
specific for ICAM-1 (BD Biosciences Pharmingen, Le
Pont de Claix, France), VCAM-1 (eBioscience, Clinic-
sciences, Montrouge, France), DC-SIGNR (R&D System,
Abingdon, UK) and HLA-I (Caltag laboratories, Tebu
J. Med. Virol. DOI 10.1002/jmv
Peyrefitte et al.
international, Le Perray en Yvelines, France), respec-
tively. Corresponding FITC- or PE- conjugated isotypic
control monoclonal antibodies (Caltag laboratories,
Tebu international) were used as negative control in
all immunophenotyping experiments. After two washes
in ice-cold ‘‘staining buffer,’’ cells were fixed in a freshly
prepared solution of 0.5% paraformaldehyde (Sigma
Aldrich) in PBS and rapidly analyzed on a 4-colors FACS
Calibur cytometer (Becton Dickinson, Pont de Claix,
France). All data were processed using the BD Cell-
QuestTM Pro software. Results are presented as mean
fluorescence intensity (MFI) that corresponds to the
geometric mean value of fluorescence calculated for a
given cell surface marker.
Chemokine and Cytokine Quantitation
by Sandwich ELISA
IL-1b, IL-6, IL-8, and RANTES present in super-
natants from dengue virus-infected microvascular EC or
from control cultures were quantified using DuoSet
R&D System ELISA kits, according to manufacturer’s
protocols. Supernatants were tested pure and 2 fold-
diluted, in duplicate. Supernatants still exhibiting
saturating OD signals were tested further at higher
dilutions.
Statistical Analysis
All viral and cellular parameters were assessed in
three independent experiments for the two studied
microvascular cell lines. For statistical comparisons of
the two groups, unpaired Student’s t-test was used. A
P-value of <0.05 was considered to be significant.
RESULTS
Susceptibility of Liver and Dermis-Derived
Microvascular Endothelial Cell Lines to
Infection by a DEN-2 Low-Passage Isolate
The two human microvascular endothelial cell lines
studied were infected with the dengue type 2 virus
strain DEN2/H/IMTSSA-MART/98-703 [Tolou et al.,
2000] at MOI 1 or 0.1, and assessed for infection and
replication efficiency at days 1, 2, 4, 5, and 7 times after
infection. As shown by indirect immunofluorescence
assay detecting the viral E protein, the percentage of
infected cells did not exceed 10% for the LSEC and 1% for
the HMEC-1 at the higher MOI of 1 (Fig. 1) in the course
of infection kinetics. No cytopathic effect was observed
in infected cultures compared to control cell lines in the
course of kinetic studies, as assessed by both microscopic
observation and trypan blue viability staining (data not
shown).
LSEC and HMEC-1 Microvascular Endothelial
Cell Lines Efficiently Replicate the
Low-Passage DEN-2 Isolate
Despite a 10-fold difference in infection rates between
the LSEC and the HMEC-1, the two microvascular cell
lines actively replicated the virus, as shown in Figure 2.
Response of Endothelial Cells to Dengue Virus
Fig. 1. Infection rates of the microvascular cell lines HMEC-1 and
LSEC as determined by indirect immunofluorescence staining of
infected cells. a: Percentage of infected cells detected in cultures of
HMEC-1 and LSEC from day 1 to 7 after exposure to the non-adapted
DEN-2 virus strain. The percentage of infected microvascular
endothelial cells after exposure to MOI 0.1 (opened diamond and
opened square for HMEC-1 and LSEC, respectively) or to MOI 1 (closed
Viral titers (or infectious doses, ID) in the supernatants
of infected HMEC-1 and LSEC ranged between 104 and
105 ID/ml from day 2 to 7 after infection (Fig. 2a,b), while
viral RNA copies were comprised between 105 and 106
copies/ml for infected cultures of HMEC-1 (Fig. 2a) and
between 106 and 107 copies/ml for infected cultures of
LSEC (Fig. 2b). Thus, when related to the percentage of
infected cells in each cell line, that is 1% for HMEC-1 and
10% for LSEC, both endothelial cell lines exhibited
comparable replication efficiency.
DEN-2 Infection Induces an Increased
Production of Inflammatory Cytokines and
Chemokines by LSEC and HMEC-1
IL-1b, IL-6, IL-8, and RANTES were quantified in the
culture supernatants of infected cultures of LSEC and
HMEC-1 since high levels of those mediators were
Fig. 2. Replication of the low-passage DEN-2 virus strain in the
microvascular cell lines HMEC-1 and LSEC assessed by determination
of viral titers and quantification of viral RNA copies. Left ordinate:
Viral RNA copies/ml (continuous lines) were quantified by real-time
RT-PCR in the supernatants of infected endothelial cell at days 1, 2, 4,
5, and 7 after infection as described in Materials and Methods. RNA
copy number presented correspond respectively to MOI 1- (closed
squares) and MOI 0.1-infected cells (closed diamonds). Right ordi-
233
diamond and closed square for HMEC-1 and LSEC, respectively)
indicated on the vertical axis, is the mean
SD of at least three
representative fields. The results presented are representative of three
independent experiments. b: Detection of dengue envelope protein
by indirect immunofluorescence in day 3-infected cultures of LSEC
(upper micrograph) and HMEC-1 (lower micrograph) microvas-
cular endothelial cell lines.
reported in the course of DSS in patients [Hober et al.,
1993; Raghupathy et al., 1998; Juffrie et al., 2001;
Suharti et al., 2003] or shown to be produced by dengue-
infected endothelial cells in vitro [Avirutnan et al., 1998;
Huang et al., 2000; Talavera et al., 2004].
IL-1b was undetectable in the supernatants of
infected or mock control cultures of HMEC-1 and LSEC
at any time of infection (data not shown). As shown in
Figure 3, IL-6 was produced constitutively and variably
by the two cell lines, but IL-6 levels increased signifi-
cantly from day 4 after infection in supernatants of
infected HMEC-1 (twofold increase; P < 104) (Fig. 3a),
and from day 2 in supernatants of infected LSEC
cultures (fourfold increase; P < 104) (Fig. 3b). IL-6
levels were higher in the supernatants of cells infected
at MOI 1 compared to MOI 0.1-infected cells, but
differences were statistically not significant.
nate: Viral infectious doses (ID/ml; dotted lines) were determined
by plaque assay titration on Vero cells at days 1, 2, 4, 5, and 7 after
infection, as described in Materials and Methods. Infectious doses were
measured in culture supernatants of respectively MOI 1- (closed
squares) and MOI 0.1-infected cells (closed diamonds). All values
presented are means
SD of values obtained in three independent
experiments.
J. Med. Virol. DOI 10.1002/jmv
234
Fig. 3. Constitutive and dengue virus-induced secretion of IL-6 by HMEC-1 and LSEC microvascular
cell lines. IL-6 levels were quantified in supernatants of mock- (~), MOI 0.1- (&), and MOI 1- (^) infected
HMEC-1 and LSEC lines at days 1, 2, 4, 5, and 7 after infection, using commercial sandwich ELISA assays.
Presented results are mean values
SD obtained from three independent experiments.
The chemokines IL-8 and RANTES were also quanti-
fied in the cell culture supernatants of uninfected and
infected LSEC and HMEC-1. As shown in Figure 4, both
cell lines produced constitutively IL-8 (Fig. 4a) and
RANTES (Fig. 4b). However, dengue infection induced a
significant increase of IL-8 and RANTES production by
both LSEC (IL-8, 3.2-fold increase from day 2, P < 104;
RANTES, 2.8-fold increase from day 2, P < 104) and
HMEC-1 (IL-8, 2.4-fold increase from day 4, P < 103
and RANTES, 2.9-fold increase from day 2, P < 103).
IL-6, IL-8, and RANTES levels in LSEC and HMEC-1
cultures exposed to heat-inactivated virus inoculum
were not different from those measured in supernatants
from cultures exposed to mock medium whether heated
or not.
Basal and Cytokine-Activated Phenotype of
LSEC and HMEC-1 Reveals Functional
Heterogeneity Between the Two
Microvascular Endothelial Cell Lines
Only the HMEC-1 has been characterized previously
for cell surface expression of adhesion molecules [Ades
et al., 1992; Xu et al., 1994]. To get more knowledge on
the behavior of HMEC-1 and LSEC lines in our culture
conditions, the basal and activated phenotype of both
HMEC-1 and LSEC lines was determined in response to
the inflammatory cytokine TNF, of which effects on the
expression of critical adhesion molecules like ICAM-1
and VCAM-1 are well established [Gerritsen et al., 1993;
Haraldsen et al., 1996].
As shown in Figure 5a, resting HMEC-1 and LSEC
express constitutively ICAM-1 although basal levels of
ICAM-1 are higher on LSEC compared to HMEC-1.
When activated with TNF for 24 hr, the two micro-
vascular cell lines respond by over-expressing ICAM-1.
However, TNF-induced induced a 13-fold increase of
ICAM-1 mean fluorescence intensity (MFI) at the sur-
face of HMEC-1 (P < 104) but only a 1.9-fold increase of
ICAM-1 MFI on LSEC (P < 104).
J. Med. Virol. DOI 10.1002/jmv
Peyrefitte et al.
As shown in Figure 5b, VCAM-1 was expressed
constitutively at very low to undetectable levels on the
two microvascular cell lines, but was highly induced
on LSEC after stimulation by TNF (4.6-fold increase
of VCAM-1 MFI; P < 104). Differently, HMEC-1
responded to TNF by slightly increasing VCAM-1
(twofold increase of VCAM-1 MFI), as shown in
Figure 5b.
Finally, since HLA-I antigens were reported to be
over-expressed during some flavivirus infections
[Kesson et al., 2002], we determined the HLA-I cell
surface levels on both resting and hIFN-g-stimulated
endothelial cells [Johnson, 2002]. As shown in
Figure 5c, the upregulation of HLA-I induced by 500
and 2,000 U/ml of hIFN-g on LSEC and HMEC-1 was
comparable (respectively, 3.4- and 3.7-fold increase of
HLA-I MFI on LSEC; 3- and 3.5-fold increase of HLA-I
MFI on HMEC-1).
Direct Infection by the Low-Passage DEN-2
Strain Differently Modulates the Expression of
Cell Surface Molecules on LSEC and HMEC-1
Endothelial Cell Lines
Immunophenotyping of the two microvascular cell
lines was carried out at different times after infection
(2 hr, 4 hr, day 1, 2, 4, 5, and 7) to assess whether dengue
infection affected the cell surface levels of antigens like
ICAM-1, VCAM-1, and HLA antigens which are related
to endothelial activation [Biederman, 2001; Galley and
Webster, 2004] or modulated in the course of flavivirus
infection [Kesson et al., 2002]. The expression of other
cell surface antigens, in particular of the endothelial-
restricted DC-SIGN related molecule DC-SIGNR (or L-
SIGN) [Navarro-Sanchez et al., 2003; Tassaneetrithep
et al., 2003], was also assessed.
In the very early times of infection, that is 2 and 4 hr,
no detectable variation of the cell surface level of any
studied antigens was observed on LSEC and HMEC-1.
However, detectable variations of ICAM-1 or/and
Response of Endothelial Cells to Dengue Virus 235

Fig. 4. a: Constitutive and dengue virus-induced secretion of IL-8 and RANTES by HMEC-1 and LSEC
microvascular cell lines. IL-8 (a) and RANTES (b) levels were quantified in supernatants of mock (~), MOI
0.1 (&), and MOI 1 (^) infected HMEC-1 and LSEC lines using commercial sandwich ELISA assays. The
mean values
SD of three independent experiments are presented.

HLA-I cell surface levels appeared as soon as
day 1 or 2 after infection, on HMEC-1 and LSEC,
respectively.
As shown in Figure 6a,b, HLA-I MFI increased by
2.35-fold (P < 0.05) and 1.71-fold (P < 0.05) on L-SEC
infected at MOI 1 and MOI 0.1, respectively, as soon as
2 days after infection. Virus-induced over-expression of
HLA-I on infected LSEC was maximal at day 5 post
infection (3.37- and 3.58-fold increase of HLA-I MFI in
MOI 1- and MOI 0.1-infected cultures; P < 0.05) and
remained significantly elevated until day 7. Infection
also induced a significant over-expression of ICAM-1
on MOI 1- and MOI 0.1-infected LSEC compared to
uninfected cells (Fig. 6c,d), between day 4 (1.65- and
1.47-fold increase of ICAM-1 MFI, respectively, on MOI
1- and MOI 0.1-infected LSEC; P < 0.05) and day 5 after
infection (1.38- and 1.39-fold increase of ICAM-1 MFI,
respectively, on MOI 1- and MOI 0.1-infected LSEC,
P < 0.05).
Infection of the HMEC-1 cell line by the DEN-2 viral
strain resulted in a distinct pattern of regulation of
ICAM-1 and HLA-I. HLA-I cell surface levels were
unchanged between day 1 and 7 on cells from infected
cultures compared to corresponding uninfected cells
(data not shown), while a slight increase could be
observed at day 7, of which significance is unknown.
Differently from LSEC, infection of HMEC-1 at both
MOIs resulted in a significant decrease of ICAM-1 cell
surface levels compared to uninfected cells: as shown in
Figure 6e,f, ICAM-1 MFI decreased significantly as soon
as day 1 after infection (1.7- and 1.63-fold decrease,
respectively, on MOI 1- and MOI 0.1-infected HMEC-1,
P < 0.05) until day 5 p.i (1.92- and 1.63-fold decrease of
ICAM-1 MFI, respectively, on MOI 1- and MOI 0.1-
infected HMEC-1; P < 0.05).
VCAM-1 surface levels were unchanged on infected
LSEC and HMEC-1 cell lines during the whole kinetic
studies, remaining comparable to basal expression
J. Med. Virol. DOI 10.1002/jmv
236
Fig. 5. Phenotypic heterogeneity in the expression of ICAM-1,
VCAM-1, and HLA-I on resting and cytokine- activated LSEC and
HMEC-1. Cell surface expression of the adhesion molecules ICAM-1
and VCAM-1, and of HLA-I antigens was assessed on HMEC-1 and
LSEC before (resting state) and after 24 hr-activation by TNF, as
described in Materials and Methods. a: Cell surface expression of
ICAM-1 on HMEC-1 (left graph) and LSEC (right graph). Green
curves represent the cell surface expression of ICAM-1 on unactivated
resting endothelial cells. Red curves represent the cell surface
expression of ICAM-1 on TNF-activated endothelial cells. b: Cell
surface expression of VCAM-1 on HMEC-1 (left graph) and LSEC (right
graph). Green curves represent the cell surface expression of ICAM-1
on unactivated endothelial cells. Red curves represent the cell surface
Fig. 6. Infection by the low-passage DEN-2 virus strain results in
increased expression of HLA-I and ICAM-1 on LSEC and decreased
expression of ICAM-1 on HMEC-1. a, c, e: Fluorescence curves
representing the cell surface expression of HLA-I (a) and ICAM-1 (b)
on LSEC and the cell surface expression of ICAM-1 on HMEC-1 (c) are
presented for mock-treated cultures (green curve), MOI 0.1-infected
cultures (red curve) and MOI 1-infected cultures (orange curve).
Results are representative of three independent experiments. Fluor-
escence graphs corresponding to isotypic controls have been subtracted
from those of corresponding markers for each culture condition using
specific functions included in the Cell Quest Pro software (Becton
J. Med. Virol. DOI 10.1002/jmv
Peyrefitte et al.
expression of VCAM-1 on TNF-activated endothelial cells. c: Expres-
sion profile of HLA-I antigens on HMEC-1 (left graph) and LSEC (right
graph). Green curves represent the basal expression level of HLA-I
antigens on unactivated endothelial cells. Red and orange curves
represent the cell surface levels of HLA-I antigens after 72 hr of culture
of HMEC-1 and LSEC in the presence of hIFN-g at respectively 500 U/
ml (red curve) and 2,000 U/ml (orange curve). For all histograms, dotted
lines represent the fluorescence intensity curves corresponding to the
fixation of isotypic control antibodies for each condition. Only one
representative control curve is presented for each graph since the geo-
metric values of fluorescence of isotypic control antibodies were similar
and control curves overlapped each other.
DickinsonTM). b, d, f: For each histogram, white bars exhibit a basal
value of 1 corresponding to the mean fluorescence intensity of the
studied antigen on mock-treated cells, relatively to its own value.
Increase or decrease of mean fluorescence intensities of ICAM-1 and
HLA-I antigens in response to infection is expressed as the ratio
between the mean fluorescence intensity of studied antigens on MOI
0.1-infected cells (sparse points) or MOI-1 infected cells (dense points)
relatively to the mean fluorescence intensity of the same antigen on
mock treated cells (white bars). Increase or decrease of mean
fluorescence intensity of the studied antigens was analyzed at days 1,
2, 4, 5, and 7 after infection.
Response of Endothelial Cells to Dengue Virus 237

Fig. 6.

J. Med. Virol. DOI 10.1002/jmv

238
levels observed on uninfected control cultures (data not
shown).
Finally, the endothelial-restricted DC-SIGNR mole-
cule remained undetectable at the surface of infected
and control cultures of LSEC and HMEC-1 in the whole
study (data not shown).
DISCUSSION
Vascular endothelial cells are central effectors in the
pathological process of vascular leakage leading to DSS.
The present study aimed at identifying phenotypic
changes induced in two human microvascular endothe-
lial cell lines from distinct vascular beds, as a conse-
quence of direct infection by a well-characterized and
non-adapted DEN-2 virus. The rationale for this study
was based on two observations: first, only partial
features of the response of endothelial cells to infection
by dengue virus have been addressed in previous
studies; second: interpretation and comparison between
data reported in these studies have been strongly
limited by the diversity of the virus strains (different
serotypes, highly adapted virus, virus isolate from
patient) and of cellular models used (mostly primary
endothelial cells or cell line from large vessels).
The two endothelial cell lines studied, that is LSEC
and HMEC-1, were derived from liver sinusoidal
endothelium [Salmon et al., 2000] and dermis [Ades
et al., 1992; Xu et al., 1994], respectively, and have
retained most morphological and functional properties
of the primary endothelial cells from which they have
been derived. The availability of such lines allowed us to
address the consequences of infection in endothelial
cells relevant to the pathophysiology of capillary
leakage. The virus chosen for the present study was
characterized entirely in our laboratory and was shown
to belong to the new American genotype encountered in
recent epidemics of dengue [Rico-Hesse et al., 1997;
Tolou et al., 2000]. A short amplification was carried out
in insect cells to produce the viral inoculum, in order to
maintain the virus as close as possible to the initial
isolate, whereas evidences indicate that in vitro adapta-
tion occurs mostly in mammalian cells [Lee et al., 1997;
Thayan et al., 1997; Chen et al., 2003].
Infection studies using this low-passage DEN-2 strain
showed that 10% of LSEC and only 1% of HMEC-1 were
infected. Previous studies reported higher infection
rates, that is 10–25% of infected cells in cultures of
primary endothelial cells from human umbilical vein
(HUVEC) [Avirutnan et al., 1998; Warke et al., 2003],
and close to 100 % of infected cells in studies using the
ECV304 cell line as cellular model of endothelial cells
[Avirutnan et al., 1998; Bosch et al., 2002]. Several
factors related both to cell and/or virus properties may
influence the efficiency of infection and explain such
differences. Both primary HUVEC or the ECV304 cell
line often used as models in dengue infection studies
[Anderson et al., 1997; Avirutnan et al., 1998; Huang
et al., 2000; Bosch et al., 2002] are derived from large
vessels, and exhibit functional and phenotypic proper-
J. Med. Virol. DOI 10.1002/jmv
Peyrefitte et al.
ties very different from those of endothelial cells
associated with microvascular territories [Bender
et al., 1994; Mason et al., 1997; Lidington et al., 1999;
Murakami et al., 2001; Kieda et al., 2002; Chi et al.,
2003]. Differences in the expression pattern of cell
surface molecules important for virus adsorption, like
virus co-receptors, that is heparan sulfate [Chen et al.,
1997; Germi et al., 2002; Lin et al., 2002], DC-SIGNR
[Navarro-Sanchez et al., 2003; Tassaneetrithep et al.,
2003], or other unindentified molecules, may influence
the efficiency of adsorption of a same dengue virus on
different types of cells [Diamond et al., 2000; Bielefeldt-
Ohmann et al., 2001]. This may explain in particular the
100% infection rate currently observed with the ECV304
cell line [Brown et al., 2000; Drexler et al., 2002] the
susceptibility of which to dengue infection depends on
the expression of three cell surface proteins of unknown
identity [Wei et al., 2003].
While heparan sulfate distribution was not investi-
gated in this study, the expression of the endothelial-
restricted C-type lectin DC-SIGNR [Bashirova et al.,
2001; Pöhlmann et al., 2001], was assessed. DC-SIGNR
is an homolog of the dengue co-receptor DC-SIGN
[Navarro-Sanchez et al., 2003; Tassaneetrithep et al.,
2003] also described as an attachment factor for other
viruses [Pöhlmann et al., 2001, 2003; Simmons et al.,
2003; Marzi et al., 2004], and has been shown to be
expressed specifically on some types of endothelial cells
[Bashirova et al., 2001]. In all infection experiments,
DC-SIGNR remained undetectable at the surface of
uninfected and infected LSEC and HMEC-1. Since this
molecule was first reported on primary liver sinusoidal
endothelial cells [Bashirova et al., 2001], its absence at
the surface of the liver–endothelial LSEC line was
unexpected, suggesting that either DC-SIGNR has been
lost due to immortalization, or alternatively that its
expression is dependent on in vivo microenvironment.
Regarding the critical role of DC-SIGN in dengue virus
infection in some types of cells [Navarro-Sanchez et al.,
2003; Tassaneetrithep et al., 2003], one can speculate
that its absence at the surface of LSEC and HMEC-1
could explain, at least partly, the low percentages of
infection observed for both LSEC and HMEC-1.
However, Talavera et al. [2004] have recently shown
that HMEC-1 cells could be infected up to 40% of total
cells in culture when exposed to a dengue 2 virus strain
of unknown genotype, suggesting that virus entry may
not depend exclusively on DC-SIGNR in those cells. The
marked difference of infection levels observed for the
endothelial cell line HMEC-1 in our experiments
compared to that of Talavera et al. [2004] may be due
to different factors of which differences of infectivity
related to the virus genotype, adaptation of virus strain
through multiple passages [Chen et al., 2003; Cologna
and Rico-Hesse, 2003; Edgil et al., 2003], methods
employed to prepare the viral batch used to infect cells.
Other reasons, independent on the viral strain may also
explain at least partly those differences, of which cell
culture conditions, infection methods, sensitivity of the
detection methods used.
Response of Endothelial Cells to Dengue Virus
Despite the low rate of infection observed for both
LSEC and HMEC-1, both microvascular endothelial cell
lines replicate the low-passage DEN-2 strain efficiently
and at comparable levels. Surprisingly, the mean virus
production, that is 105 I.D/ml estimated at 2 days of
infection, is comparable to that reported for HUVEC
infected by the DEN 2 16681 laboratory strain, a
prototype virus strain with high replication efficiency
[Huang et al., 2000; Edgil et al., 2003]. This suggests
that microvascular endothelial cells could act as an
important replication site during dengue virus infec-
tion.
The ability of a virus to induce or not some cytopathic
effects in host cells is an important aspect of the host
cell–virus interaction. While actively and efficiently
replicating the DEN-2 strain, infected cultures of
HMEC-1 and LSEC were devoid of detectable cytopathic
effect when compared to corresponding uninfected
control cultures. Comparable observations have been
reported by others, respectively, on HUVEC or on
HMEC-1 infected by dengue virus [Bunyaratvej et al.,
1997; Talavera et al., 2004].
Microvascular endothelial cells are an active interface
between blood immune cells and tissues. Many stimuli
of which vasoactive compounds but also several patho-
gens [Shen et al., 1997; Knight et al., 1999; Sundstrom
et al., 2001; Caruso et al., 2002; De Maula et al., 2002;
Hippenstiel and Suttorp, 2003] induce endothelial cells
to change from a physiological state to a pro-inflamma-
tory and pro-adhesive state. This is most often char-
acterized by increased production of inflammatory
soluble mediators and by over-expression of cell surface
molecules, that is adhesion molecules, HLA antigens,
and others, involved in the cross talk between endothe-
lial cells and circulating immune cells.
Such effects could be observed as a consequence of
direct exposure of LSEC and HMEC-1 to the DEN-2 low
passage isolate. Both microvascular endothelial cell
lines over-produced IL-6, IL-8, and RANTES, whereas
this could not be attributed exclusively to infected cells
since cytokines were quantified in the total super-
natants of cell cultures containing no more than 1 and
10% of infected cells for HMEC-1 and LSEC lines,
respectively. Intracellular staining of cytokines com-
bined to cell surface staining of viral antigens should
allow to define if only infected cells over-produce those
soluble factors in response to dengue infection. How-
ever, the increase of cytokines levels observed in
infected cultures of LSEC and HMEC-1 was comparable
to that reported for dengue-infected HUVEC or ECV304
[Huang et al., 2000; Bosch et al., 2002] that contain a
much higher percentage of infected cells. Differences in
absolute concentration of IL-6, IL-8, and RANTES
produced by HMEC-1 and LSEC, that is 5- to 30-fold
lower concentrations between the former and the latter,
might reflect the 10-fold difference of infection rate
between LSEC and HMEC-1, or alternatively might be
due to functional differences between the two lines.
Increased production of IL-6, IL-8, and RANTES
by LSEC and HMEC-1 in response to infection is in
239
agreement with results from previous studies carried
out using different dengue virus strains and types of
endothelial cells. All reported an over-production of IL-6
[Huang et al., 2000], IL-8 [Avirutnan et al., 1998;
Talavera et al., 2004], or RANTES [Avirutnan et al.,
1998], thus suggesting that increased secretion of these
inflammatory mediators is a common feature of endo-
thelial cells response to dengue viruses, whatever the
origin of viruses and the tissue-specificity of endothelial
cells.
The increased secretion of pro-inflammatory media-
tors by dengue-infected microvascular endothelial cells
should be considered with respect to markers increased
in the course of systemic inflammation and particularly
in dengue shock syndrome. High levels of IL-6, one of the
most significant marker of severity in systemic inflam-
mation [Reinhart et al., 2002] have been associated with
the most severe forms of dengue [Hober et al., 1993;
Juffrie et al., 2001]. High levels of the CXC chemokine
IL-8 [Baggiolini et al., 1995] were also associated with
severe dengue infection [Raghupathy et al., 1998], and
IL-8 has been shown to increase the permeability of
dengue-infected endothelial cells expressing adequate
IL-8 receptors [Salcedo et al., 2000; Talavera et al.,
2004], thus contributing potentially to the pathophysio-
logical mechanisms of plasma leakage.
The increased secretion by dengue-infected endothe-
lial cells of the CC chemokine RANTES may enhance the
local inflammatory response by facilitating the transmi-
gration of inflammatory Th1-type lymphocytes to endo-
thelium-associated tissues [Kawai et al., 1999].
Variations in the cell surface levels of adhesion
molecules and HLA antigens is another feature of
endothelial activation. Analysis were first carried out
to characterize more precisely the phenotypic changes
occurring in HMEC-1 and LSEC lines in response to
mediators known to activate endothelial cells like TNF
or IFNg. The results show that resting and TNF-
activated LSEC and HMEC-1 have distinct pattern of
expression of adhesion molecules and HLA-I antigens.
Whereas LSEC exhibit higher constitutive levels of
ICAM-1 compared to HMEC-1, they only slightly up-
regulate ICAM-1 in response to TNF, compared to
HMEC-1 which strongly over-express ICAM-1 in res-
ponse to this cytokine. Differently, TNF-induces high
levels of VCAM-1 on LSEC but only low levels of VCAM-
1 at the surface of the HMEC-1. Finally, LSEC exhibit
higher constitutive levels of HLA-I than HMEC-1 while
both cell lines respond similarly to stimulation by hIFN-
g, that is about threefold increase in HLA-I mean
fluorescence intensity. Those results underline the
functional heterogeneity of the two microvascular endo-
thelial cell lines studied, a critical parameter to consider
when investigating the contribution of one or another
type of vascular endothelium to a pathophysiological
process [Bender et al., 1994; Murakami et al., 2001].
Interestingly, such a functional heterogeneity was
also observed in the response of LSEC and HMEC-1 to
infection by the low-passage DEN-2 strain. Indeed,
direct infection of LSEC resulted in an increased
J. Med. Virol. DOI 10.1002/jmv
240
expression of ICAM-1 and HLA-I at the surface of the
total cell population. Whereas more pronounced effects
could have been identified by looking specifically at
infected endothelial cells, this could not be done due to
the poor quality of double staining with antibodies
directed to both studied markers and viral antigens.
The observed increased of ICAM-1 at the surface of
LSEC as a result of infection is an interesting feature
that differs from previous study on HUVEC infected by
the highly adapted Dengue type 2 strain 16681, showing
no change in ICAM-1 cell surface expression in response
to dengue virus infection [Anderson et al., 1997].
However, the virus-induced over-expression of ICAM-1
and HLA-I observed on LSEC is comparable to that
reported on HUVEC infected with another flavivirus,
that is West Nile virus [Shen et al., 1997], whereas in
this latter model, maximal upregulation of ICAM-1 was
observed in the very early times of infection. Thus
common signaling pathways may be involved in the
upregulation of those molecules and particularly of
HLA-I antigens, during flavivirus infections [Kesson
and King, 2001; Momburg et al., 2001; Cheng et al.,
2004]. The functional consequences of increased expres-
sion of ICAM-1 and HLA-I on infected LSEC could be of
importance for the interactions of those cells with
inflammatory immune cells, by contributing to enhance
the recruitment of circulating immune cells to the pro-
inflammatory endothelium [Martinez-Mier et al., 2001].
Over expression of HLA-I may interfere with the
development of an early innate NK response [Lobigs
et al., 2003], while enhancing the cytotoxic response
of virus-specific CD8þ T-cells towards infected cells, as
shown for cells infected by West Nile virus [Kesson and
King, 2001], thus contributing to the immunopatholo-
gical response. Despite limitations in extrapolating our
results to in vivo cellular events, the fact that LSEC
respond to dengue virus infection by acquiring features
of activated endothelium, suggests that liver-associated
endothelial cells could play a role in the initiation or the
amplification of the inflammatory cascade leading to
vascular leakage in DSS.
Differently from LSEC, infection of HMEC-1 by the
same DEN-2 low-passage strain resulted in a decreased
expression of ICAM-1, while HLA-I cell surface levels
remained unchanged compared to uninfected cells. As
for LSEC, it is likely that double cell surface staining
allowing to focus our analysis on infected cells that
represent only 1% of total HMEC-1 cultures, would have
revealed more pronounced effects of infection.
However, despite those limitations, the prolonged
decrease of ICAM-1 cell surface expression observed on
HMEC-1 cells after exposure to dengue virus, may have
functional consequences on the ability of HMEC-1
to bind circulating leucocytes. Indeed, comparable
decrease of ICAM-1 cell surface levels in a different
model, was shown to limit the recruitment of inflamma-
tory cells to this type of microvascular endothelial cells
[Lindermann et al., 2000].
The present study aimed at examining the phenotypic
changes induced by direct exposure of human micro-
J. Med. Virol. DOI 10.1002/jmv
Peyrefitte et al.
vascular endothelial cells from different vascular beds,
to a non-adapted dengue type 2 virus. The results show
that microvascular endothelial cells from two distinct
vascular beds, not only support viral replication but can
be activated through direct exposure to dengue virus,
exhibiting some common but also some cell type-specific
responses to dengue infection, suggesting that the
response of microvascular beds to dengue infection
may differ from one territory to another one.
The observed changes should now be completed by
functional studies to investigate if viral infection
modifies the permeability or other cellular mechanisms
related to vascular leakage, and by analysis of interac-
tions between dengue-infected microvascular endothe-
lial cells and immune cells in the context of dengue
infection.
REFERENCES
Ades EW, Candal FJ, Swerlick RA, George VG, Suymmers S, Bosse DC,
Lawley TJ. 1992. HMEC-1–Establishment of an immortalized
human microvascular endothelial cell line. J Invest Dermatol 99:
683–690.
Anderson R, Wang S, Osiowy C, Issekutz A. 1997. Activation of
endothelial cells via antibody-enhanced dengue virus infection of
peripheral blood monocytes. J Virol 71:4226–4232.
Andrews BS, Theofilopoulos AN, Peters CJ, Loskutoff DJ, Brandt WE,
Dixon FJ. 1978. Replication of Dengue and Junin viruses in
cultured rabbit and human endothelial cells. Infect and Immun
20:776–781.
Avirutnan P, Malasit P, Seliger B, Bhakdi S, Husmann M. 1998.
Dengue virus infection of human endothelial cells leads to
chemokine production, complement activation, and apoptosis.
J Immunol 161:6338–6346.
Baggiolini M, Loetscher P, Moser B. 1995. Interleukin-8 and the
chemokine family. Int J Immunopharmac 17:103–108.
Bashirova AA, Geijtenbeek TBH, van Duijnhoven GCF, van Vliet SJ,
Eilering JBG, Martin MP, Wu L, Martin TD, Viebig N, Knolle PA,
Kewalramani VN, van Kooyk Y, Carrington M. 2001. A dendritic
cell-specific intercellular adhesion molecule 3-grabbing noninte-
grin (DC-SIGN)-related protein is highly expressed on human liver
sinusoidal endothelial cells and promotes HIV-1 infection. J Exp
Med 193:671–678.
Bender JR, Sadeghi MM, Watson C, Pfau S, Pardi R. 1994. Hetero-
genous activation threshold to cytokines in genetically distinct
endothelial cells: Evidence for diverse transcriptional responses.
Proc Natl Acad Sci USA 91:3994–3998.
Bhamarapravati N, Tuchinda P, Boonyapaknavik V. 1967. Pathology
of Thailand haemorrhagic fever: A study of 100 autopsy cases. Ann
Trop Med Parasitol 61:500–510.
Biederman BC. 2001. Vascular endothelium: Checkpoint for inflam-
mation and immunity. News Physiol Sci 16:84–88.
Bielefeldt-Ohmann H, Meyer M, Fitzpatrick DR, Mackenzie JS. 2001.
Dengue virus binding to human leukocyte cell lines: Receptor usage
differs between cell types and virus strains. Virus Res 73:81–89.
Boonpucknavig S, Boonpucknavig V, Bhamarapravati N, Nimmanni-
tya S. 1979. Immunofluorescence study of skin rash in patients with
dengue hemorrhagic fever. Arch Pathol Lab Med 103:463–466.
Bosch I, Xhaja K, Estevez L, Raines G, Melichar H, Warke RV, Fournier
MV, Ennis FA, Rothman AL. 2002. Increased production of
interleukin-8 in primary human monocytes and in human epithe-
lial and endothelial cell lines after dengue virus challenge. J Virol
76:5588–5597.
Brown J, Readings SJ, Jones S, Fitchett CJ, Howl L, Martin A,
Longland CL, Michelangeli F, Dubrova YE, Brown CA. 2000.
Critical evaluation of ECV304 as a human endothelial cell model
defined by genetic analysis and functional responses: A comparison
with the human bladder cancer derived epithelial cell line T24/83.
Lab Invest 80:37–45.
Bunyaratvej A, Butthep P, Yoksan S, Bhamarapravati N. 1997.
Dengue viruses induce cell proliferation and morphological changes
Response of Endothelial Cells to Dengue Virus
of endothelial cells. Southeast Asian J Trop Med Public Health
28:32–37.
Carr JM, Hocking H, Bunting K, Wright PJ, Davidson A, Gamble J,
Burrel CJ, Li P. 2003. Supernatants from dengue virus type-2
infected macrophages induce permeability changes in endothelial
cell monolayers. J Med Virol 69:521–528.
Caruso A, Rotola A, Comar M, Favilli F, Galvan M, Tosetti M, Campello
C, Caselli E, Alessandri G, Grassi M, Garrafa E, Cassai E, Di Luca
D. 2002. HHV-6 infects human aortic and heart microvascular
endothelial cells, increasing their ability to secrete proinflamma-
tory chemokines. J Med Virol 67:528–533.
Chen JP, Cosgriff TM. 2000. Hemorrhagic fever virus-induced changes
in haemostasis and vascular biology. Blood Coag and Fibrinol
11:461–483.
Chen Y, Maguire T, Hileman RE, Fromm JR, Esko JD, Linhardt RJ,
Marks RM. 1997. Dengue virus infectivity depends on envelope
protein binding to target cell heparan sulfate. Nat Med 8:866–871.
Chen WJ, Wu HR, Chiou SS. 2003. E/NS1 modifications of dengue 2
virus after serial passages in mammalian and/or mosquito cells.
Intervirology 46:289–295.
Cheng Y, King NJ, Kesson AM. 2004. Major histocompatibility complex
class I (MHC-I) induction by West nile virus: Involvement of 2
signaling pathways in MHC-I upregulation. J Infect Dis 189:658–
669.
Chi JT, Chang HY, Haraldsen G, Jahnsen FL, Troyanskaya OG, Chang
DS, Wang Z, Rockson SG, van de Rijn M, Botstein D, Brown PO.
2003. Endothelial cell diversity revealed by global expression
profiling. Proc Natl Acad Sci USA 100:10623–10628.
Cologna R, Rico-Hesse R. 2003. American genotype structures decrease
dengue virus output from human monocytes and dendritic cells.
J Virol 77:3929–3938.
Damonte EB, Pujol CA, Coto CE. 2004. Prospects for the therapy and
prevention of dengue virus infections. Adv Virus Res 63:239–285.
De Maula CD, Leutenegger CM, Jutila MA, MacLachlan NJ. 2002.
Bluetongue virus-induced activation of primary bovine lung
microvascular endothelial cells. Ve Immunopathol 86:147–157.
Desruelles F, Lamaury I, Roudier M, Goursaud R, Mahe A, Castanet J,
Strobel M. 1997. Cutaneo-mucous manifestations of dengue. Ann
Dermatol Venereol 124:237–241.
Diamond MS, Dianna E, Roberts TG, Lu B, Harris E. 2000. Infection of
human cells by dengue virus is modulated by different cell types and
viral strains. J Virol 74:7814–7823.
Drexler HG, Quentmeier H, Dirks WG, MacLeod RA. 2002. Bladder
carcinoma cell line ECV304 is not a model system for endothelial
cell. In Vitro Cell Dev Biol Anim 38:185–186.
Edgil D, Diamond MS, Holden KL, Paranjape SM, Harris E. 2003.
Translation efficiency determines differences in cellular infection
among dengue virus type 2 strains. Virology 317:275–290.
Gagnon SJ, Mori M, Kurane I, Green S, Vaughn DW, Kalayanarooj S,
Suntayakorn S, Ennis FA, Rothman AL. 2002. Cytokine gene
expression and production in peripheral blood mononuclear cells of
children with acute dengue virus infections. J Med Virol 67:41–46.
Galley HF, Webster NR. 2004. Physiology of the endothelium. Br J
Anaesth 93:105–113.
Germi R, Crance JM, Garin D, Guimet J, Lortat-Jacob H, Ruigrok RW,
Zarski JP, Drouet E. 2002. Heparan sulfate-mediated binding of
infectious dengue virus type 2 and yellow fever virus. Virology
292:162–168.
Gerritsen ME, Niedbala MJ, Szczepanski A, Carley WW. 1993.
Cytokine activation of human macro- and microverssel-derived
endothelial cells. Blood Cells 19:325–339.
Green S, Vaughn DW, Kalayanarooj S, Nimmannitya S, Suntayakorn
S, Nisalak A, Lew R, Innis BL, Kurane I, Rothman AL, Ennis FA.
1999. Early immune activation in acute dengue illness is related to
development of plasma leakage and disease severity. J Inf Dis 179:
755–762.
Guzman MG, Kouri G. 2002. Dengue: An update. The Lancet Infect
Diss 2:33–42.
Guzman MG, Kouri G, Diaz M, Llop A, Vasquez S, et al. 2004. Dengue,
one of the great emerging health challenges of the 21st century.
Expert Rev Vaccines 3:511–520.
Haraldsen G, Kvale D, Lien B, Farstad IN, Brandtzaeg P. 1996.
Cytokine-regulated expression of E-selectin, intercellular adhesion
molecule-1 (ICAM-1), and vascular cell adhesion molecule-1
(VCAM-1) in human microvascular endothelial cells. J Immunol
156:2258–2265.
241
Hippenstiel S, Suttorp N. 2003. Interaction of pathogens with endo-
thelium. Thromb Haemost 89:18–24.
Hober D, Poli L, Roblin B, Gestas P, Chungue E, Granic G, Imbert P,
Pecarere JL, Vergez-Pascal R, Wattre P, Maniez-Montreuil M.
1993. Serum levels of tumor necrosis factor-alpha (TNF-alpha),
interleukin-6 (IL-6), and interleukin-1 beta (IL-1 beta) in dengue-
infected patients. Am J Trop Med Hyg 48:324–331.
Huang YH, Lei HY, Liu HS, Lin YS, Liu CC, Yeh TM. 2000. Dengue
virus infect human endothelial cells and induces IL-6 and IL-8
production. Am J Trop Med Hyg 63:71–75.
Iyngkaran N, Yadav M, Sinniah M. 1995. Augmented inflammatory
cytokines in primary dengue infection progressing to shock.
Singapore Med J 36:218–221.
Jessie K, Fong MY, Devi S, Lam SK, Wong KT. 2004. Localization of
dengue virus in naturally infected tissues, by immunohistochem-
istry and in situ hybridization. J Infect Dis 189:1411–1418.
Johnson DR. 2002. Locus-specific constitutive and cytokine-induced
HLA Class I gene expression. J Immunol 170:1894–1902.
Juffrie M, Meer GM, Hack CE, Haasnoot K, Sutary A, Veerman J, Thijs
LG. 2001. Inflammatory mediators in dengue virus infection in
children: Interleukin-6 and its relation to C-reactive and secretory
phospholipase A2. Am J Trop Med Hyg 65:70–75.
Kawai T, Seki M, Hiromatsu K, Esatcott JW, Watts GFM, Sugai M,
Smith DJ, Porcelli SA, Taubman MA. 1999. Selective diapedesis of
Th1 cells induced by endothelial cell RANTES. J Immunol 163:
3269–3278.
Kesson AM, King NJ. 2001. Transcriptional regulation of major
histocompatibility complex class I by flavivirus West nile is
dependent on NF-kappaB activation. J Infect Dis 184:947–954.
Kesson AM, Cheng Y, King NJ. 2002. Regulation of immune
recognition molecules by flavivirus, West Nile. Viral Immunol 15:
273–283.
Kieda C, Paprocka M, Krawczenko A, Zalecki P, Dupuis P, Monsigny
M, Radzikowski C, Dus D. 2002. New human microvascular
endothelial cell lines with specific adhesion molecules phenotypes.
Endothelium 9:247–261.
Kleczkowski A. 1968. Estimation of virus concentration by the dilution
method. In: Maramorosch K, Koprowski H, editor. Methods in
virology. New York: Academic Press, Inc. pp 705–713.
Klein CL, Bittinger F, Köhler H, Wagner M, Otto M, Hermanns I,
Kirkpatrick CJ. 1995. Comparative studies on vascular endothe-
lium in vitro. Effects of cytokines on the expression of E-selectin,
ICAM-1 and VCAM-1 by cultured human endothelial cells obtained
from different passages. Pathobiology 63:83–92.
Knight DA, Waldman WJ, Sedmak DD. 1999. Cytomegalovirus-
mediated modulation of adhesion molecule expression by human
arterial and microvascular endothelial cells. Transplantation 68:
1814–1818.
Krishnaswamy G, Kelley J, Yerra L, Smith JK, Chi DS. 1999. Human
endothelium as a source of multifunctional cytokine: Molecular
regulation and possible role in human disease. J Interferon
Cytokine Res 19:91–104.
Lee E, Weir RC, Dalgarno D. 1997. Changes in the dengue virus major
envelope protein on passaging and their localization on the three-
dimensional structure of the protein. Virology 232:281–290.
Lei HY, Yeh TM, Liu YS, Lin YS, Chen SH, Liu CC. 2001.
Immunopathogenesis of dengue virus infection. J Biomed Sc 8:
377–388.
Letmeyer KC, Vaughn DW, Watts DM, Salas R, Villalobos I, de Chacon
Ramos C, Rico-Hesse R. 1999. Dengue virus structural differences
that correlate with pathogenesis. J Virol 73:4738–4747.
Lidington EA, Moyes DL, McCormack AM, Rose ML. 1999. A
comparison of primary endothelial cells and endothelial cell
lines for studies of immune interactions. Transpl Immunol 7:239–
246.
Lin YL, Lei HY, Lin YS, Yeh TM, Chen SH, Liu HS. 2002. Heparin
inhibits dengue-2 virus infection of five human liver cell lines.
Antiviral Res 56:93–96.
Lin CF, Lei HY, Shiau AL, Liu CC, Liu HS, Yeh TM, Chen SH, Lin YS.
2003. Antibodies from dengue patient sera cross-react with
endothelial cells and induce damage. J Med Virol 69:82–90.
Lindermann S, Sharafi M, Spiecker M, Buerke M, Fisch A, Grosser T,
Veit K, Gierer C, Ibe W, Meyer J, Darius H. 2000. NO reduces PMN
adhesion to human vascular endothelial cells due to downregula-
tion of ICAM-1 mRNA and surface expression. Thromb Res 97:113–
123.
J. Med. Virol. DOI 10.1002/jmv
242
Liu CC, Huang KJ, Lin YS, Yeh TM, Liu HS, Lei HY. 2002. Transient
CD4/CD8 ratio inversion and aberrant immune activation during
dengue virus infection. J Med Virol 68:241–252.
Lobigs M, Mullbacher A, Regner M. 2003. MHC class I upregulation by
flavivises: Immune interaction with unknown advantage to host or
pathogen. Immunol Cell Biol 81:217–223.
Luria SE, Darnell JE, Baltimore D, Campbell A. 1978. Titration of
viruses. In: Luria SE, Darnell JE, editors. General virology, 3rd
edition. New York: John Wiley and Sons. pp 21–32.
Martinez-Mier G, Toledo-Pereya LH, Ward PA. 2001. Adhesion
molecules and hemorrhagic shock. J Trauma 51:408–415.
Marzi A, Gramberg T, Simmons G, Möller P, Rennekamp AJ,
Krumbiegel M, Geier M, Eisemann J, Turza N, Saunier B,
Steinkasserer A, Becker S, Bates S, Hofmann H, Pöhlmann S.
2004. DC-SIGN and DC-SIGNR interact with the glycoprotein of
Marburg virus and the S protein of the severe acute respiratory
syndrome coronavirus. J Virol 78:12090–12095.
Mason JC, Yarwood H, Sugars K, Haskard DO. 1997. Human umbilical
vein and dermal microvascular endothelial cells show heterogene-
ity in response to PKC activation. Am J Physiol 273:C1233–C1240.
Mohan B, Patwari AK, Anand VK. 2000. Hepatic dysfunction in
childhood dengue infection. J Trop Pediatr 46:40–43.
Momburg F, Mullbacher A, Lobigs M. 2001. Modulation of transporter
associated with antigen processing (TAP)-mediated peptide import
into the endoplasmatic reticulum by flavivirus infection. J Virol
75:5663–5671.
Murakami S, Morioka T, Nakagawa Y, Suzuki Y, Arakawa M, Oite T.
2001. Expression of adhesion molecules by cultures human
glomerular endothelial cells in response to cytokines: Comparison
to human umbilical vein and dermal microvascular endothelial
cells. Microvasc Res 62:383–391.
Navarro-Sanchez E, Altmayer R, Amara A, Schwartz O, Fieschi F,
Virelizier JL, Arenzana-Seidedos F, Despres P. 2003. Dendritic-
cell-specific ICAM3-grabbing non-integrin is essential for the
productive infection of human dendritic cells by mosquito-cell-
derived dengue viruses. EMBO Rep 4:723–728.
Pancharoen C, Rungsarannont A, Thisyakorn U. 2002. Hepatic
dysfunction in dengue patients with various severity. J Med Assoc
Thai 85:S298–S301.
Pöhlmann S, Soilleux EJ, Baribaud F, Leslie GJ, Morris LS, Trowsdale
J, Lee B, Coleman N, Doms RW. 2001. DC-SIGNR, a DC-SIGN
homologue expressed in endothelial cells, binds to human and
simian immunodeficiency viruses and activates infection in trans.
Proc Natl Acad Sci 98:2670–2675.
Pöhlmann S, Zhang J, Baribaud F, Chen Z, Leslie GJ, Lin G, Granelli-
Piperno A, Doms RW, McKeating JA. 2003. Hepatitis C virus
glycoproteins interact with DC-SIGN and DC-SIGNR. J Virol 77:
4070–4080.
Raghupathy R, Chaturvedi UC, Al-Sayer H, Elbishbishi EA, Agarway
R, Nagar R, Kapoor S, Mistra A, Mathur A, Nusrat H, Azizieh F,
Khan MAY, Mustafa AS. 1998. Elevated levels of IL-8 in dengue
hemorrhagic fever. J Med Virol 56:280–285.
Reinhart K, Bayer O, Brunkhorst F, Meisner M. 2002. Markers of
endothelial damage in organ dysfunction and sepsis. Crit Care Med
30:S302–S312.
Rico-Hesse R, Harrison LM, Sala RA, Tovar D, Nisalak A, Ramos C,
Boshell de Mesa MT, Nogueira RMR, Travassos da Rosa A. 1997.
Origins of Dengue type 2 viruses associated with increased patho-
genicity in the Americas. Virology 230:244–251.
Sahaphong S, Riengrojpitak S, Bhamarapravati N, Chirachariyavej T.
1980. Electron microscopic study of the vascular endothelial cell in
dengue hemorrhagic fever. Southeast Asian J Trop Med Public
Health 11:194–204.
Salcedo R, Resau JH, Halverson D, Hudson EA, Dambach M, Powell D,
Wasserman K, Oppenheim JJ. 2000. Differential expression and
responsiveness of chemokine receptors (CXCR1-3) by human
J. Med. Virol. DOI 10.1002/jmv
Peyrefitte et al.
microvascular endothelial cells and umbilical vein endothelial
cells. FASEB J 14:2055–2064.
Salmon P, Oberholzer J, Occhiodoro T, Morel P, Lou J, Trono D. 2000.
Reversible immortalization of human primary cells by lentivector-
mediated transfer of specific genes. Mol Ther 2:404–414.
Shen J, T-TO SS, Schrieber L, King NJC. 1997. Early E-Selectin,
V-CAM-1 and late major histocompatibility complex antigen
induction on human endothelial cells by flavivirus and comodula-
tion of adhesion molecule expression by immune cytokines. J Virol
71:9323–9332.
Simmons G, Reeves JD, Grogan CG, Vandenberghe LH, Baribaud F,
Whitbeck JC, Burke E, Buchmeier MJ, Soilleux EJ, Riley JL, Doms
RW, Bates P, Pöhlmann S. 2003. DC-SIGN and DC-SIGNR bind
Ebola glycoproteins and enhance infection of macrophages and
endothelial cells. Virology 305:115–123.
Srichaikul T, Nimmannitya S. 2000. Haematology in dengue and
dengue haemorrhagic fever. Baillieres Best Pract Res Clin Haematol
13:261–276.
Suharti C, van Gorp ECM, Dolmans WMV, Setiati TE, Hack CE,
Djokomoeljanto RJ, van der Meer JWM. 2003. Cytokine patterns
during dengue shock syndrome. Eur Cytokine Netw 14:172–177.
Sundstrom JB, McMullan LK, Spiropoulou CF, Hooper WC, Ansari AA,
Peters CJ, Rollin PE. 2001. Hantavirus induces expression of
Rantes and IP-10 without causing increased permeability in human
lung microvascular endothelial cells. J Virol 75:6070–6085.
Talavera D, Castillo AM, Domingez MC, Gutierrez AE, Meza I. 2004.
IL-8 release, tigh junction and cytosqueleton dynamic reorganiza-
tion conducive to permeability increase are induced by dengue virus
infection of microvascular endothelial monolayers. J Gen Virol 85:
1801–1813.
Tassaneetrithep B, Burguess TH, Granelli-Piperno A, Trumpfheller C,
Finke J, Wellington S, Eller MA, Pattanapanyasat K, Sarasombath
S, Birx DL, Steinman RM, Schlesinger S, Marovich MA. 2003. DC-
SIGN (CD209) mediates dengue virus infection of human dendritic
cells. J Exp Med 197:823–829.
Thayan R, Morita K, Vijayamalar B, Zainah S, Chew TK, Oda K,
Sinniah M, Igarashi A. 1997. Comparison of sequences of E/NS1
gene junction of dengue type 3 virus following culture subpassage in
C6/36 cells to study the possible occurrence of mutations. Southeast
Asian J Trop Med Public Health 28:380–386.
Tolou H, Couissinier-Paris P, Mercier V, Pisano MR, de Lamballerie X,
de Micco P, Durand JP. 2000. Complete genomic sequence of a
dengue type 2 virus from the French West Indies. Biochem Biophys
Res Commun 277:89–92.
Vermot-Desroches C, Marchand D, Roy C, Wijdenes J. 1995. Hetero-
geneity of antigen expression among human umbilical cord
vascular endothelial cells: Identification of cell subsets by co-
expression of haemopoietic antigens. Immunol Lett 48:1–9.
Wahid SF, Sanusi S, Zawawi MM, Ali RA. 2000. A comparison of the
pattern of liver involvement in dengue hemorrhagic fever with
classic dengue fever. Southeast Asian J Trop Med Public Health
31:259–263.
Warke RV, Xhaja K, Martin KJ, Fournier MF, Shaw SK, Brizuela N, de
Bosch N, Lapointe D, Ennis FA, Rothman AL, Bosch I. 2003.
Dengue virus induces novel changes in gene expression of human
umbilical vein endothelial cells. J Virol 7:11822–11832.
Warrilow D, Northill JA, Pyke A, Smith GA. 2002. Single rapid TaqMan
fluorogenic probe based PCR assay that detects all four dengue
serotypes. J Med Virol 66:524–528.
Wei HY, Jiang LF, Fang DY, Guo HY. 2003. Dengue virus type 2 infects
human endothelial cells through binding of the viral envelope
glycoprotein to cell surface polypeptides. J Gen Virol 84:3095–3098.
Xu Y, Swerlick RA, Sepp N, Bosse D, Ades EW, Lawley TJ. 1994.
Characterization of expression and modulation of cell adhesion
molecules on an immortalized human dermal microvascular
endothelial cell line (HMEC-1). J Invest Dermatol 102:833–837.