Invasion of endothelial cells and arthritogenic

Microbiology (2014), 160, 537–546
Editor’s Choice
DOI 10.1099/mic.0.069948-0
Invasion of endothelial cells and arthritogenic
potential of endocarditis-associated
Corynebacterium diphtheriae
Renata Stavracakis Peixoto,1,23 Gabriela Andrade Pereira,1,23
Louisy Sanches dos Santos,1 Cláudio Marcos Rocha-de-Souza,1
Débora Leandro Rama Gomes,3 Cintia Silva dos Santos,1,2
Lucia Maria Correa Werneck,4 Alexandre Alves de Souza de Oliveira Dias,4
Raphael Hirata Jr,1 Prescilla Emy Nagao5 and Ana Luı́za Mattos-Guaraldi1
1
Correspondence
Ana Luı́za Mattos-Guaraldi
[email protected]
Laboratory of Diphtheria and Corynebacteria of Clinical Relevance (LDCIC),
Faculty of Medical Sciences, Rio de Janeiro State University (UERJ), Rio de Janeiro, RJ, Brazil
2
Department of Medical Microbiology, Institute of Microbiology,
Rio de Janeiro Federal University (IMPPG/UFRJ), Rio de Janeiro, RJ, Brazil
3
Faculty of Pharmacy, Federal Institute of Education, Science and Technology of Rio de Janeiro
(IFRJ), Rio de Janeiro, RJ, Brazil
4
National Institute for Quality Control in Health (INCQS), Fundação Oswaldo Cruz (FIOCRUZ),
Rio de Janeiro, RJ, Brazil
5
Biology Institute Roberto Alcântara Gomes, Rio de Janeiro State University (UERJ),
Rio de Janeiro, RJ, Brazil
Received 27 May 2013
Accepted 13 December 2013
Although infection by Corynebacterium diphtheriae is a model of extracellular mucosal
pathogenesis, different clones have been also associated with invasive infections such as sepsis,
endocarditis, septic arthritis and osteomyelitis. The mechanisms that promote C. diphtheriae
infection and haematogenic dissemination need further investigation. In this study we evaluated
the association and invasion mechanisms with human umbilical vein endothelial cells (HUVECs)
and experimental arthritis in mice of endocarditis-associated strains and control non-invasive
strains. C. diphtheriae strains were able to adhere to and invade HUVECs at different levels. The
endocarditis-associated strains displayed an aggregative adherence pattern and a higher number
of internalized viable cells in HUVECs. Transmission electron microscopy (TEM) analysis revealed
intracellular bacteria free in the cytoplasm and/or contained in a host-membrane-confined
compartment as single micro-organisms. Data showed bacterial internalization dependent on
microfilament and microtubule stability and involvement of protein phosphorylation in the HUVEC
signalling pathway. A high number of affected joints and high arthritis index in addition to the
histopathological features indicated a strain-dependent ability of C. diphtheriae to cause severe
polyarthritis. A correlation between the arthritis index and increased systemic levels of IL-6 and
TNF-a was observed for endocarditis-associated strains. In conclusion, higher incidence of
potential mechanisms by which C. diphtheriae may access the bloodstream through the
endothelial barrier and stimulate the production of pro-inflammatory cytokines such as IL-6 and
TNF-a, in addition to the ability to affect the joints and induce arthritis through haematogenic
spread are thought to be related to the pathogenesis of endocarditis-associated strains.
INTRODUCTION
3These authors contributed equally to this paper.
Abbreviations: FAS, FITC-labelled phalloidin assay; HUVEC, human
umbilical vein endothelial cell; i.v., intravenously; SW, Swiss Webster;
TEM, transmission electron microscopy.
069948 G 2014 SGM
Corynebacterium diphtheriae is the causative agent of diphtheria, a toxemic localized infection of the respiratory tract.
Furthermore, C. diphtheriae is not only the aetiological
agent of diphtheria, but can cause other infections as
well (Ott et al., 2010). Systemic infections caused by
Downloaded from www.microbiologyresearch.org by
IP: 78.47.27.170
On: Thu, 29 Sep 2016 18:49:00
Printed in Great Britain
537
R. S. Peixoto and others
C. diphtheriae (Patey et al., 1997; Mattos-Guaraldi &
Formiga, 1998) suggest that diphtheria bacilli are able
to penetrate respiratory epithelial cells and gain access to
deeper tissues (Mattos-Guaraldi & Formiga, 1998; Hirata
et al., 2004). Bacteraemia and endocarditis caused by both
non-toxigenic and toxigenic C. diphtheriae strains have
been reported with increased frequency (Hirata et al., 2004;
Bertuccini et al., 2004). Endocarditis lethality due to nontoxigenic C. diphtheriae has been shown to be 64 % and
19 % for C. diphtheriae subsp. mitis and C. diphtheriae
subsp. gravis, respectively (Bertuccini et al., 2004). Infective
endocarditis due to C. diphtheriae may occur (37 %) in
patients with normal cardiac valves or without any known
risk factors and on occasion may be an aggressive disease.
Non-toxigenic strains seem to have a preference for leftside valves and tend to form large valvular vegetations,
septic emboli and aneurysms (Belko et al., 2000). Septic
arthritis caused by non-toxigenic C. diphtheriae has also
been described (Guran et al., 1979; Appelbaum & Dossett,
1982; Tiley et al., 1993; Damade et al., 1993; Barakett et al.,
1993; Patey et al., 1997; Puliti et al., 2006; Hirata et al.,
2008). In a study conducted by Tiley et al. (1993), 50 %
of patients developed septic arthritis as a complication of
endocarditis due to non-toxigenic C. diphtheriae subsp.
gravis. Osteoarticular involvement due to non-toxigenic
C. diphtheriae subsp. mitis infection was also observed in
between 27.5–40 .0% of patients (Tiley et al., 1993).
Internalization or translocation through endothelial cells
results in invasion of the vascular system by human pathogens (Mishra et al., 2005). The role of endothelial cells
in the pathogenesis of invasive and localized C. diphtheriae
infections has not been documented until now. The main
objective of the present study was to investigate the in vitro
association and invasion mechanisms of C. diphtheria
with human umbilical vein endothelial cells (HUVECs) in
addition to the arthritogenic potential observed using conventional Swiss Webster (SW) mice intravenously (i.v.)
injected with C. diphtheriae subsp. mitis endocarditisassociated (HC01 and HC02) invasive strains and noninvasive (non-toxigenic type ATCC 27010 and homologous
toxigenic ATCC 27012) strains.
METHODS
Bacterial strains and growth conditions. Clinical and micro-
biological characteristics of partially studied endocarditis-associated
(HC01 and HC02) C. diphtheriae subsp. mitis blood isolates evaluated
in this study are displayed in Table 1: non-toxigenic ATCC 27010
(C7s (2) tox-; NCTC 11397) type strain and the homologous toxigenic ATCC 27012 strain from the American Type Culture Collection
were also included in the study. The non-adherent and non-virulent
DH5-a Escherichia coli strain was used as negative control in a cytoskeleton mobilization test. Stock cultures in 10 % skimmed milk with
25 % glycerol added were maintained at 270 uC and recovered as
required by cultivation in trypticase soy broth (TSB; Difco; Hirata
et al., 2002; Santos et al., 2005).
Human umbilical vein endothelial cell culture. Primary human
umbilical vein endothelial cells (HUVECs) were obtained by treatment
538
of umbilical veins with 0.1 % collagenase IV solution (Sigma) as previously described (Menzies & Kourteva, 2000; Cossart & Sansonetti,
2004). HUVECs were maintained in 199 medium (M199)/HEPES
(Sigma), supplemented with 20 % FCS, 2 mM glutamine, 50 mg gentamicin ml21 and 250 mg amphotericin ml21 at 37 uC in a humidified
5 % CO2 atmosphere until monolayers reached confluence. Cells were
cultured in 24-well tissue culture plates and only cells collected after the
first or second passage were assayed. Confluent cultures were treated
with 0.025 % trypsin/0.2 % EDTA solution prepared in Dulbecco’s
mineral salt solution (PBSS), rinsed in serum-depleted culture medium, and used for the experiments.
Bacterial adherence and intracellular viability assays. Aliquots
(500 ml) of bacterial suspensions in a concentration of 108 c.f.u. ml21
were used to infect HUVEC monolayers grown to about 95 % confluence on 24-well tissue culture plates (approximate m.o.i. of 100 : 1).
After incubation periods of 0, 30, 60, 120, 360 min and 24 h, infected
HUVECs were washed six times with PBSS, lysed with 0.1 % Triton
X-100 (Sigma) in PBSS (PBSS-T), diluted and cultured on trypticase
soy agar (TSA; Difco) at 37 uC, 24 or 48 h for bacterial viable counting (c.f.u. ml21). Bacterial inoculum corresponded to the number of
viable bacterial cells in supernatant plus the number of viable bacteria
associated with HUVEC monolayers (intracellular plus extracellular;
Hirata et al., 2002). The total number of cell-associated bacteria
(intracellular plus extracellular) was expressed as the percentage of the
inoculum recovered from monolayers after 3 h incubation.
For bright-field microscopic analysis, bacterial infection was
performed in semi-confluent HUVEC monolayers grown on circular
coverslips (13 mm diameter). Following a 180 min incubation, infected cells were washed three times with PBSS, fixed with methanol
and stained by Giemsa (Hirata et al., 2004).
The ability of C. diphtheriae to enter HUVECs was assessed by the
gentamicin protection assay as previously described (Hirata et al.,
2002). Briefly, cells were gently washed five times with PBSS and
subsequently incubated with 500 ml M199 medium containing
gentamicin (150 mg ml21) and incubated for 1 h at 37 uC. The wells
were then washed three times with PBSS. Cell lysis was carried out as
described above and the lysates (100 ml) from each well were diluted
and plated onto TSA-agar for determination of viable intracellular
bacteria. The number of intracellular bacteria was determined by
viable counts after lysis of gentamicin-pretreated HUVEC monolayers
with 0.5 ml 0.1 % Triton X-100 in PBS. The percentage of intracellular bacteria was deduced from HUVEC-associated bacteria. All
assays were performed in triplicate and repeated at least three times.
FITC-labelled phalloidin assay (FAS test). Infected semiconfluent
HUVECs monolayers (3 h) were rinsed six times with PBS, fixed with
3 % formaldehyde for 15 min, and then treated with PBSS-T for
10 min. Monolayers were washed three times with PBS and treated
with 5 mg fluorescein conjugated phalloidin ml21 (Sigma) for 30 min.
FAS test results were considered positive when foci of intense
fluorescence corresponded to areas of bacterial adherence observed
under phase-contrast microscopy (Hirata et al., 2002). The control
strains E. coli E2348/69 (FAS-positive) and E. coli DH5-a (FASnegative) were included in this study. The experiments were performed in triplicate.
Treatment of HUVECs with cytochalasin E, colchicine and
genistein. Final concentrations of inhibitors and solvents were used
as follows: 5 mM cytochalasin E in DMSO; 2 mg colchicine in 70 %
ethanol; 100 mM genistein in DMSO. Monolayers were pre-incubated
with each inhibitor at 37 uC for 30 min prior to bacterial infection as
described above. All chemicals were purchased from Sigma (von
Hunolstein et al., 2003). None of the inhibitors tested affected either
bacterial viability or growth (as assessed by standard counts of viable
Downloaded from www.microbiologyresearch.org by
IP: 78.47.27.170
On: Thu, 29 Sep 2016 18:49:00
Microbiology 160
http://mic.sgmjournals.org
Table 1. Clinical and microbiological aspects of endocarditis-associated Corynebacterium diphtheriae subsp. mitis strains evaluated in this study.
Both strains exhibited an aggregative adherence pattern in HUVEC, similar to those observed in human endothelial (HEp-2) cells (Hirata et al., 2008).
agglutination of human B erythrocytes; A%, percentage of viable adherent bacteria; I%, percentage of intracellular viability.
Strain
Clinical aspects
Gender/
age (y)
HC01
F/14
HC02
F/9
Arthritis/other
complications
ND/Peripheral
emboli
of legs and arms;
microaneurism
with brain
haemorrhage
Arthritis/Myositis;
valvar abscess;
acute renal failure
ND,
not determined; F, female; HA,
Microbiological aspects
Prior
heart
defect
Outcome
Tox
gene*
Biofilm
formation
on abiotic
surfaceD
HA
HEp-2
cells (3 h
incubation)
HUVEC cells
(3 h incubation)
A%
I%
A%
I%
Arthritis
index
Infected
mice with
arthritis
clinical
signs (%)
IL-6/
TNF-a
plasma
levels
(pg ml”1)
51.80
(P,0.05)
1.79
(P,0.05)
3.5
(P50.0007)
86.66
134.4 /
115.6
(P,0.001)
35.40
0.06
1.5
50.00
75.97 /
79.75
(P,0.001)
No
Death
No
Yes
Yes
51.25
2.59
No
Valve
replacement
No
Yes
Yes
ND
ND
539
C. diphtheriae invasive properties
*Detected by Multiplex PCR (Pimenta et al., 2008).
DPolystyrene (negatively charged).
SW mice i.v. infection
(3 days after infection)
Downloaded from www.microbiologyresearch.org by
IP: 78.47.27.170
On: Thu, 29 Sep 2016 18:49:00
R. S. Peixoto and others
bacteria), or affected HUVEC viability (as assessed by Trypan blue
staining) or morphology.
Transmission electron microscopy (TEM). Infected monolayers
were rinsed and fixed with 0.1 M cacodylate buffer pH 7.2 containing
2 % glutaraldehyde, 4 % paraformaldehyde, 5 mM CaCl2 for 1 h at
4 uC. Monolayers were post-fixed with 1 % OsO4, 5 mM CaCl2 and
0.8 % K4 [Fe (CN6)] in cacodylate buffer for 1 h, 22 uC, and dehydrated through a graded ethanol series. They were then embedded in
Epon, thin sectioned, and examined with an EM 906 Zeiss transmission electron microscope (Zeiss; Hirata et al., 2002).
Experimental model of bacterial infectious arthritis in mice.
Statistical analysis. All assays were performed in triplicate and
repeated at least three times. Results are presented as the mean±SD.
One way ANOVA analysis with Tukey test was used to compare
means of columns, with P-value ,0.05 considered significant (dos
Santos et al., 2010).
RESULTS
Adherence and entry of C. diphtheriae strains into
HUVECs
Conventional healthy SW mice, which were sex-matched and weighed
18–22 g, from CAECAL-FIOCRUZ were anaesthetized with intraperitoneal injection of ketamin (50 mg kg21). The study was performed
in compliance with guidelines outlined in the International Guiding
Principles for Biomedical Research Involving Animals as issued by
the Council for the International Organizations of Medical Sciences
and with the Brazilian government’s ethical guidelines for research
involving animals (Fiocruz Ethic Committee for Animal Experiments –
CEUA/FIOCRUZ – L-034/07). For clinical evaluation of arthritis
and histological assessment, induced-septic arthritis experiments
were carried out in triplicate based on methods previously described
(Puliti et al., 2000). Mice injected i.v. with 46108 ml21 of tox+ or toxC. diphtheriae strains were evaluated daily for up to 1 month for signs
of arthritis, in particular for the presence of joint inflammation, and
scored for arthritis severity (macroscopic score). Arthritis was defined
as visible erythema and/or swelling of at least one joint. To evaluate
the intensity of arthritis, the following clinical scoring (arthritic index)
was used for each limb: 1 point, mild swelling and erythema; 2 points,
moderate swelling and erythema; 3 points, marked swelling, erythema,
and/or ankylosis. Thus, a mouse could have a maximum score of 12.
The arthritis index (mean±SD) was constructed by dividing the total
score (cumulative value of all paws) by the number of animals in each
experimental group.
Percentages of both viable cell-associated bacteria (extracellular and intracellular) and intracellular bacteria at
different periods of incubation (non-toxigenic ATCC 27010
type strain and endocarditis-associated HC01 strain) are
displayed in Table 2. Data indicated that C. diphtheriae
isolates were able to adhere to and survive within HUVECs
in varied periods of incubation and at different levels. Viable
associated and internalized bacteria were detected immediately post-infection of monolayers with the HC01 strain and
ATCC 27010 strain, respectively. The non-toxigenic ATCC
27010 strain was the only one capable of survival (0.45 %)
within the intracellular compartment 24 h post-infection.
Histopathological procedures to evaluate the features of the disease
were based on methods previously described (Dias et al., 2011).
Briefly, mice inoculated i.v. with non-toxigenic C. diphtheriae were
sacrificed after 9 days. Control mice groups were injected with saline.
Arthritic paws (one per mouse) were removed aseptically, fixed with
10 %, v/v, formalin for 24 h and then decalcified in 10 % EDTA in
phosphate buffer 100 mM for 7 days. Subsequently, the specimens
were dehydrated, embedded in paraffin, sectioned at 5–7 mm and
stained with haematoxylin and eosin. Joints were examined for
synovitis (defined as synovial membrane thickness of more than two
cell layers), extent of infiltrate (presence of inflammatory cells in the
subcutaneous and/or periarticular tissues), cartilage damage, bone
erosion and loss of joint architecture. Arthritis severity was classified
as mild (mild synovial hypertrophy and minimal infiltrate), moderate
(moderate synovial hypertrophy, presence of infiltrate, minimal
exudate, integrity of joint architecture) and severe (marked synovial
hypertrophy, presence of massive infiltrate/exudate, cartilage and
bone erosion and disrupted joint architecture) (Dias et al., 2011).
After reaching the highest level of interaction, adherence
seemed independent of toxin production, as evidenced by
comparison of the adherence rates of the non-toxigenic
ATCC 27010 type strain (30.80 %) and the toxigenic
homologous strain (22.60 %; P.0.05). The invasion rates
of the non-toxigenic ATCC 27010 type strain (0.92 %) and
the toxigenic homologous ATCC 27012 strain (1.33 %;
P50.0012) 3 h post-infection suggested an influence of tox
gene in the internalization process by HUVECs. However,
the non-toxigenic HC01 endocarditis-associated isolate
demonstrated the highest number of viable internalized
bacteria followed by the toxigenic ATCC 27012 strain
(1.79 % and 1.33 %, respectively; P50.0009). Intracellular
viability was lower (0.06 %) for the HC02 blood/endocarditis isolate.
Serum sample preparation and evaluation of cytokine concentration. Blood samples from C. diphtheriae-infected mice were
obtained by retro-orbital sinus bleeding at different time points (day
0, 1, 3, 15 and 20). They were left for 30 min at room temperature
to clot and then were centrifuged at 2000 g for 10 min. Sera were
collected and stored at 280 uC until analysed. The levels of IL-6 and
TNF-a in the serum samples were determined by using specific ELISA
kits according to the manufacturer’s instructions (R&D Systems).
The results were expressed as the concentration of cytokine as pg
(ml serum)21, as extrapolated from a standard curve constructed
using recombinant cytokine (Dias et al., 2011).
540
All C. diphtheriae strains showed a maximum level of
ability to enter and survive within HUVECs at 3 h postinfection. A higher ability to enter and survive within endothelial cells was observed for the endocarditis-associated
HC01 strain when compared to the ATCC 27010 type
strain (P,0.05). Interaction with HUVECs was higher for
the non-toxigenic HC01 (51.80 %) and HC02 (35.40 %)
blood/endocarditis isolates.
The non-toxigenic HC01 blood/endocarditis and ATCC
27010 strains demonstrated the highest mean level of
adherent (31.16 % and 16.80 %, respectively) and internalized (0.58 % and 0.40 %, respectively; P50.0003) viable
bacteria (Table 2).
Since the maximum interaction occurred at about 3 h
post-infection for the C. diphtheriae strains tested, this
incubation period was chosen for further investigation into
the mechanisms of bacterial adherence and internalization
within endothelial cells (Table 1; Figs 1–4).
Downloaded from www.microbiologyresearch.org by
IP: 78.47.27.170
On: Thu, 29 Sep 2016 18:49:00
Microbiology 160
C. diphtheriae invasive properties
Table 2. Percentage of viable bacterial cells associated to (A%) and internalized by (I%) HUVECs at different periods post-infection
of non-toxigenic Corynebacterium diphtheriae strains. Results are presented as the mean±SD.
Infection period (h)
C. diphtheriae strains
Control
Endocarditis-associated
ATCC 27010
HC01
A%
0
0.5
1
2
3
24
Mean value
0
7.70±0.50
17.77±0.50
28.70±1.00
30.80±13.74*
15.90±0.60
16.80
I%
A%
I%
0.11±0.00
0.15±0.03
0.34±0.08
0.41±0.04
0.92±0.01*
0.45±0.09
0.40
0.60±0.59
23.30±0.82
34.30±9.02
44.40±6.85
51.80±9.55*
32.60±1.26
31.16
0
0.37±0.50
0.44±0.02
0.90±0.15
1.79±0.01*
0
0.58
*Comparative analysis between data presenting statistically significant differences with P,0.05.
Microscopic examination of infected endothelial
cells
Light micrographs demonstrating the aggregative adherence
pattern in HUVECs, characterized by clumps of bacteria
with a ‘stacked-brick’ appearance of endocarditis-associated
C. diphtheriae subsp. mitis HC01 and HC02 blood isolates
(3 h incubation) are displayed in Fig. 1a and 1b, respectively.
TEM assays illustrated the close association between bacteria
and endothelial cell surfaces for both non-toxigenic ATCC
27010 and endocarditis-associated HC01 C. diphtheriae
strains after 3 h of incubation (Fig. 1c, d). Micro-organisms
were found attached by focal linkages to HUVEC membranes (Fig. 1d). Membrane invagination of the host cell at
the site of bacterial attachment is shown in Figs. 1e–h.
Fibrillar-like structures were visible on bacterial surfaces
during the internalization process (Fig. 1f–h). Internalized
bacteria were also found in a host-membrane-confined
compartment as a single micro-organism (Fig. 1g, h) and/or
free in the cytoplasm (Fig. 1i, j).
FAS test
Fluorescence and phase-contrast microscopy showed positive FAS tests of invasive C. diphtheriae subsp. mitis (HC01
and HC02) strains isolated from the blood of patients with
endocarditis, the toxigenic C. diphtheriae strain ATCC 27012
and the ATCC 27010 non-toxigenic type strain. Cytoskeletal
changes were observed directly beneath the adherent bacteria with accumulation of polymerized actin. Foci of intense
fluorescence corresponded to areas of bacterial adherence
observed under phase-contrast microscopy (Fig. 2).
The effects of eukaryotic inhibitors on the
C. diphtheriae internalization process
Analysis of the effect of cytochalasin E (a known inhibitor of eukaryotic microfilament formation) showed a
http://mic.sgmjournals.org
complete inhibition of bacterial internalization for all four
C. diphtheriae (HC01, HC02, ATCC 27010 and ATCC
27012) strains tested. To investigate whether the invasion
process involved actin polymerization via a phosphotyrosine-signalling cascade, bacterial interaction with HUVECs
was evaluated in the presence of genistein, a tyrosine kinase
inhibitor. This showed a reduction of bacterial internalization to ¡0.10 % (P,0.05) for all strains tested. Pretreatment of HUVECs with the microtubule-disrupting drug
colchicine also caused a significant reduction of internalization to ¡0.15 % (P,0.05) of all C. diphtheriae strains
tested.
Arthritogenicity and histopathological findings
SW mice were infected with invasive (HC01, HC02) or
non-invasive (ATCC 27010 and ATCC 27012) C. diphtheriae strains. Viable bacteria were recovered from joints,
blood, kidneys, liver, and spleen but not from the heart
and lungs of mice. Data displayed in Fig. 3a, b showed
that non-toxigenic C. diphtheriae strains were capable of
inducing arthritis through haematogenic spread at different levels. The invasive (HC01, HC02) strains were arthritogenic at a dose of 46108 c.f.u. per mouse. The highest
percentages (85.42 %) of clinical signs of arthritis in affected
mice (arthritis index) were observed for the endocarditisassociated HC01 (85.42 %) and HC02 (50 %) strains (Fig.
3a, b; Table 1). Arthritis index, indicating the clinical
severity of septic arthritis caused by C. diphtheriae, was
graded on a scale of 0–3 for each paw, according to changes
in erythema and swelling (0, no change; 1, mild swelling
and/or erythema; 2, moderate swelling and erythema; 3,
marked swelling, erythema and/or ankylosis; Fig. 3b). Clinical signs of arthritis were evident, starting 3 days after i.v.
injection. The most frequently affected joints were ankle
and wrist. To confirm clinical signs of arthritis, histopathological studies of joints of invasive non-toxigenic
Downloaded from www.microbiologyresearch.org by
IP: 78.47.27.170
On: Thu, 29 Sep 2016 18:49:00
541
R. S. Peixoto and others
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
(j)
(a)
(b)
(c)
(d)
(e)
(f)
Fig. 2. Fluorescence and phase-contrast microscopy shows
positive FAS tests of C. diphtheriae subsp. mitis HC01 blood
isolate (a, b), ‘non-endocarditis’ toxigenic ATCC 27012 (c, d) and
non-toxigenic ATCC 27010 type strain (e, f). The experiments
were done in triplicate. Arrows indicate the positive reactions to
the FAS test and the presence of adherent bacteria in phasecontrast micrographs. Details of fluorescent spots were magnified
by ¾2 in the right inferior corner.
to bone tissue and synoviocyte hypertrophy (data not
shown).
Cytokine production during C. diphtheriae
infection
Fig. 1. Micrographs demonstrating the aggregative adherence
pattern in HUVECs, characterized by clumps of bacteria with
a ‘stacked-brick’ appearance of endocarditis-associated C.
diphtheriae subsp. mitis HC01 (a) and HC02 (b) blood isolates
(3 h incubation). Magnification, ¾1000. TEM demonstrating
intimate contact (c) and focal points of adherence (d) of C.
diphtheriae with the host cell membrane; (e, f) the HC01 strain
internalization process in HUVECs after 3 h incubation; (g, h)
electron-dense material surrounding adherent bacteria and formation of a cup-shaped process on the endothelial cell membrane
underlying adherent bacteria. Internalized bacterial cells are
located in a loose vacuole. (i, j) Micro-organisms appear to be
free within partially degraded cytoplasm. Areas of interest are
marked with black arrows. Scale bars represent 500 nm in (c), (e),
(g) and (i), and 200nm in (f), (h) and (j).
C. diphtheriae-infected mice were performed. Three affected
paws were examined in each mice group after 9 days of
infection. The most severe histopathological features of
arthritis in mice were caused by the HC01 strain, with
subcutaneous oedema, inflammatory infiltrate and damage
542
Cytokines appear to play a central role in the development
of articular lesions in experimental models of septic
arthritis. To assess the involvement of cytokines in C.
diphtheriae-induced arthritis, systemic levels of production
of IL-6 and TNF-a were investigated at different time
points after infection using different strains. Time-course
experiments showed that there was a positive correlation
between IL-6 and TNF-a release and time after infection
(Fig. 4). A significant (P,0.001) increase of TNF-a (up
to 116 pg ml21) and IL-6 (up to 134 pg ml21) production
was evident on day 3 after infection. A progressive decrease
was observed from these peaks until day 20. High systemic
levels of IL-6 were found in the mice groups infected with
the HC01 and HC02 invasive strains in comparison to
the non-toxigenic ATCC 27010 and toxigenic ATCC 27012
strains after infection. The toxigenic ATCC27012 strain
showed the lowest capacity to induce IL-6 and TNF-a after
infection. A correlation between the arthritis index (Fig.
3b) and IL-6 and TNF-a production was also observed
for all invasive C. diphtheriae strains (Fig. 4), where the
HC01 strain had the highest arthritogenic potential and the
ability to induce cytokine release after mice infection.
Downloaded from www.microbiologyresearch.org by
IP: 78.47.27.170
On: Thu, 29 Sep 2016 18:49:00
Microbiology 160
C. diphtheriae invasive properties
140
80
120
60
100
40
TNF-α (pg ml–1)
Affected mice (%)
(a) 100
20
0
HC01
HC02
27010
80
60
27012
40
(b)
4
20
Arthritis index
3
0
0
2
1
3
10
15
Time after infection (days)
20
160
140
1
0
3
6
9
12
15
20
Time post-infection (days)
25
30
Fig. 3. Experimental model of bacterial infectious arthritis in mice.
(a) Percentage of C. diphtheriae-infected SW mice exhibiting
clinical signs of arthritis. Mice were challenged i.v. with 4¾108
c.f.u. of endocarditis-associated HC01, HC02 strains and of
ATCC 27010 and ATCC 27012 control strains. Results are
means±SD of three separate experiments; n, 15 mice per group.
(b) Severity of septic arthritis in SW mice infected with 4¾108
c.f.u. of different C. diphtheriae strains per mouse: the toxigenic
(#) ATCC 27012 strain, the non-toxigenic homologous (e)
ATCC 27010 strain and the endocarditis-associated ($) HC01
and (&) HC02 strains. Clinical severity of arthritis was graded on
a scale of 0–3 for each paw, according to changes in erythema
and swelling (0, no change; 1, mild swelling and/or erythema; 2,
moderate swelling and erythema; 3, marked swelling, erythema
and/or ankylosis. The arthritis index was constructed by dividing
the total score (cumulative value of all paws) by the number of
animals in each experimental group. Results are means±SD of
three separate experiments. P,0.05 Tukey test (one-way ANOVA)
with Graphpad Prism; n, 15 mice per group.
DISCUSSION
In addition to diphtheria, C. diphtheriae has been also associated with invasive infections such as sepsis and endocarditis (Mattos-Guaraldi et al., 2000). Bacteraemia followed by
attachment of bacteria to an endocardial surface precedes
the establishment of endocarditis (Plotkowski et al., 1994;
Mishra et al., 2005). The fact that C. diphtheriae endocarditis
http://mic.sgmjournals.org
IL6 (pg ml–1)
120
100
80
60
40
20
0
0
1
3
10
15
Time after infection (days)
20
Fig. 4. Induction of IL-6 and TNFa in SW mice exhibiting clinical
signs of arthritis due to C. diphtheriae endocarditis-associated (&)
HC01, ($) HC02 strains and control (¤) 27010 and (e) 27012
strains. Cytokine levels were measured by ELISA and calculated
as the means±SD of three independent experiments; values are
in pg (ml serum)”1. P,0.05 Tukey test (one-way ANOVA) with
Graphpad Prism; n: 5 mice per group per time point. SD values,
which were usually ,10 % have been omitted.
involves normal heart valves in many instances underscores
the importance of using viable undamaged endothelial cell
surfaces as the substrate in experimental models (Ogawa
et al., 1985). For this reason, confluent monolayers of
HUVECs were used in tissue culture as an analogue of the
host surface encountered in vivo. This is the first report on
the adherence and intracellular survival within endothelial cells by C. diphtheriae strains. Presently, viable bacteria
were found on the surface of endothelial cells for all strains
tested. C. diphtheriae strains displayed various degrees of
Downloaded from www.microbiologyresearch.org by
IP: 78.47.27.170
On: Thu, 29 Sep 2016 18:49:00
543
R. S. Peixoto and others
attachment and internalization to human endothelial cells
indicating differences in the ability to cause invasive bloodborn infection. In fact, the HC01 and HC02 blood culture
isolates which, by definition, have caused invasive bloodborne infection and therefore have had the potential to
infect endothelial surfaces, consistently adhered most efficiently to HUVECs, with similar adherence patterns observed to HEp-2 cells (Hirata et al., 2008). Data suggest that
potential adherence to normal endothelium may help to
elucidate the ability of C. diphtheriae to cause endocarditis,
as for other Gram-positive pathogens, including Staphylococcus aureus, Streptococcus agalactiae, Streptococcus sanguis
and Listeria monocytogenes (Vercellotti et al., 1984; Gibson
et al., 1993; Drevets, 1998; Matussek et al., 2005).
Invasion is a common strategy of pathogens to establish a
privileged niche in the host. Earlier investigations showed
the ability of both toxigenic and non-toxigenic clinical
C. diphtheriae isolates to enter and survive within cultured respiratory epithelial cells (Plotkowski et al., 1994;
Bertuccini et al., 2004). The ability of C. diphtheriae to
survive intracellularly would benefit the bacterium, as it
would be protected against host immune control mechanisms and many antimicrobial agents. It is likely that internalization and intracellular survival of C. diphtheriae is
involved in throat colonization, thus contributing to the
failure of bacterial eradication and asymptomatic carriage.
Several bacterial species are internalized by endothelial
cells, even though these cells are not normally considered
phagocytic (Finlay & Falkow, 1997). In the present study, it
was demonstrated that C. diphtheriae also has the ability
to enter and survive within endothelial cells at a variety of
levels. Although some C. diphtheriae strains seem to have
a preference for attaching to vascular endothelium, microorganisms (HC01 and HC02 strains) that have caused
invasive blood-born infection also differed in their potential to infect HUVECs: the HC01 strain was found to be
more adherent and invasive to HUVECs than the HC02
strain. This fact may help to explain the aggressive nature
of the disease leading to death or valve replacement in
patients with endocarditis despite antimicrobial therapy in
many instances, including in cases related with HC01 and
HC02 strains, respectively (Mattos-Guaraldi & Formiga,
1998; Belko et al., 2000).
This current study showed that non-toxigenic strains were
highly adherent to and invasive into HUVECs. Furthermore, live intracellular bacterial cells detectable 24 h
post-infection were observed only for the non-toxigenic
ATCC 27010 type strain. These data together strengthen
clinical findings showing invasive infections caused by
non-toxigenic C. diphtheriae strains (Patey et al., 1997; von
Hunolstein et al., 2003).
Positive FAS test results indicated the involvement of host
cell cytoskeletal rearrangements in C. diphtheriae uptake
by endothelial cells. The internalization of C. diphtheriae
by HUVECs was found to be an active process that can be
blocked by the phagocytic inhibitor cytochalasin E. Actin
544
polymerization along with phospho-tyrosine signalling
events seems to be a mechanism required during the
C. diphtheriae internalization process into HUVECs, as
previously observed with human respiratory epithelial cells
(Plotkowski et al., 1994; Bertuccini et al., 2004). Inhibition
assays with colchicines suggested that there are either
two pathways (microtubule dependent and microfilament
dependent) or one complex pathway involving polymerization of microtubules and/or microfilaments in C.
diphtheriae uptake by endothelial cells.
The mechanisms underlying bacterial entry, phagosome
maturation, and dissemination reveal common strategies
as well as unique tactics evolved by individual species to
establish infection. Invasive bacteria actively induce their
own uptake by phagocytosis in normally non-phagocytic
cells and then either establish a protected niche within
which they survive and replicate, or disseminate from cell
to cell by means of an actin-based motility process (Cossart
& Sansonetti, 2004). The C. diphtheriae internalization
process in HUVECs appeared to proceed in a sequential
fashion as it is observed with professional phagocytes.
Bacterial adherence to the endothelial cell was followed
by the appearance of endothelial appendages which elongated and enclosed the bacteria into apparent phagosomes.
Previous studies showed a similar sequence of ultrastructural events for endothelial phagocytosis of Staphylococcus
aureus (Menzies & Kourteva, 2000). Like Staphylococcus
aureus (Menzies & Kourteva, 2000; Matussek et al., 2005),
ultrastructural analysis of C. diphtheriae-infected HUVECs
demonstrated intracellular bacteria free in the cytoplasm
or within loose vacuoles. Cytoplasmic vacuolization and
degradation suggest death of infected endothelial cells that
would probably provide a route for the dissemination of
C. diphtheriae to distant sites where metastatic infections
could be established. A previous report showed the ability of
these bacteria to cause apoptosis in human U-937 macrophages (dos Santos et al., 2010). Studies are in progress
to elucidate the molecular mechanisms of C. diphtheriaeinduced cell death in endothelial cells. Independently of
toxin production, C. diphtheriae strains may have diverged
regarding mechanisms of interaction with the endothelial
barrier, possibly developing different pathways of entry.
Differences in the ability of C. diphtheriae strains to adhere
to HUVECs are suggestive of specific mechanisms for the
binding of the pathogen to endothelial surfaces.
Active uptake of bacteria by endothelial cells may not only
assist in C. diphtheriae virulence in endovascular infections,
independently of toxin production, but can be important
to establish metastatic infection, like polyarthritis.
In an attempt to better understand endocarditis-associated
C. diphtheriae subsp. mitis pathogenesis, a bacterial arthritis
infection experimental model which showed similarities
with human disease (Goldenberg & Reed, 1985) was prepared using conventional SW mice injected i.v. with strains
isolated from animals and humans. The current study using
this experimental mouse model showed the ability of
Downloaded from www.microbiologyresearch.org by
IP: 78.47.27.170
On: Thu, 29 Sep 2016 18:49:00
Microbiology 160
C. diphtheriae invasive properties
invasive non-toxigenic C. diphtheriae strains to affect the
joints and to induce arthritis through haematogenic spread.
The joint histopathological analysis of the C. diphtheriaeinfected mice (subcutaneous oedema, inflammatory infiltrate, damage to bone tissue and synoviocyte hypertrophy)
strongly demonstrated the arthritogenic potential of the
invasive non-toxigenic C. diphtheriae. Nevertheless, no signs
of arthritis were observed for the toxigenic C. diphtheriae
ATCC27012, showing that the diphtheria toxin might not
be crucial in arthritogenic pathogenesis. These data are in
accordance with recently published studies also showing
the arthritogenic potential of non-toxigenic C. diphtheriae
strains. However, those strains were isolated from patients
with pharyngitis (Puliti et al., 2006). To the best of our
knowledge, this is the first report showing not only the
ability of non-toxigenic C. diphtheriae strains isolated
from patients with endocarditis and arthritis to invade
and survive into endothelial cells, but to cause arthritis
in a mouse model. Interestingly, osteoarthritis has been
described as an important clinical feature in 27–50 % of
patients with C. diphtheriae systemic infection (Patey et al.,
1997).
septic arthritis in SW conventional mice could be a useful
tool in studies on the pathogenicity and characterization of
virulence factors other than diphtheria toxin.
The non-toxigenic C. diphtheriae strains were different in
the incidence and the severity of arthritis. The appearance
of arthritis lesions is undoubtedly a multifactorial process
involving bacterial virulence factors and the host immune
system (Puliti et al., 2000). For instance, the role of proinflammatory cytokines (IL-6, IL1-b and TNF-a) in the
arthritis pathogenesis has been well documented in both
human and animal models (Feldmann et al., 1996). A
direct correlation between arthritis severity and levels of
IL-6 and TNF-a has been demonstrated in a mouse model
for Corynebacterium ulcerans, group B Streptococcus agalactiae and Staphylococcus aureus (Tissi et al., 1990; Bremell
et al., 1991; Dias et al., 2011). Here, this study has also
provided evidence that all non-toxigenic C. diphtheriae
strains induce IL-6 and TNF-a secretion. In addition,
arthritis severity observed in infected mice was directly
associated with IL-6 levels, as observed in C. ulcerans
(Dias et al., 2011). In fact, the more severe arthritis index
observed in mice infected with non-toxigenic C. diphtheriae HC01 strain is associated with highest levels of IL-6 and
TNF-a. HC01 and HC02 strains also induced high IL-6 and
TNF-a release by HUVEC during adherence and invasion
(data not shown).
nalization of non-toxigenic Corynebacterium diphtheriae by cultured
human respiratory epithelial cells. Microb Pathog 37, 111–118.
In conclusion, C. diphtheriae isolates behave differently
when they meet host surfaces, with regard to the cells they
efficiently infect and the kind of inflammatory response they
trigger. Adherence to and internalization by endothelial cells
may be critical steps during blood-barrier disruption and
systemic dissemination of both non-toxigenic and toxigenic
C. diphtheriae strains. In a similar way to Staphylococcus
aureus (Vercellotti et al., 1984), the ability of C. diphtheriae
to invade and survive within endothelial cells may also
contribute to its propensity to cause persistent endovascular
infection with endothelial destruction and metastatic
infections. Furthermore, a model of C. diphtheriae-induced
http://mic.sgmjournals.org
ACKNOWLEDGEMENTS
This work was supported by CAPES, FAPERJ, CNPq, SR-2/UERJ and
the ProgramaNacional de Pós-Doutorado (PAPD –FAPERJ/CAPES).
No competing financial interests exist.
REFERENCES
Appelbaum, P. C. & Dossett, J. H. (1982). Arthritis caused by
unidentified gram-positive rods in a child with acute lymphocytic
leukemia. J Clin Microbiol 15, 525–527.
Barakett, V., Morel, G., Lesage, D. & Petit, J. C. (1993). Septic arthritis
due to a nontoxigenic strain of Corynebacterium diphtheriae subspecies mitis. Clin Infect Dis 17, 520–521.
Belko, J., Wessel, D. L. & Malley, R. (2000). Endocarditis caused by
Corynebacterium diphtheriae: case report and review of the literature.
Pediatr Infect Dis J 19, 159–163.
Bertuccini, L., Baldassarri, L. & von Hunolstein, C. (2004). Inter-
Bremell, T., Lange, S., Yacoub, A., Rydén, C. & Tarkowski, A. (1991).
Experimental Staphylococcus aureus arthritis in mice. Infect Immun
59, 2615–2623.
Cossart, P. & Sansonetti, P. J. (2004). Bacterial invasion: the
paradigms of enteroinvasive pathogens. Science 304, 242–248.
Damade, R., Pouchot, J., Delacroix, I., Boussougant, Y. & Vinceneux,
P. (1993). Septic arthritis due to Corynebacterium diphtheriae. Clin
Infect Dis 16, 446–447.
Dias, A. A., Silva, F. C., Jr, Santos, L. S., Ribeiro-Carvalho, M. M.,
Sabbadini, P. S., Santos, C. S., Filardy, A. A., Myioshi, A., Azevedo,
V. A. & other authors (2011). Strain-dependent arthritogenic poten-
tial of the zoonotic pathogen Corynebacterium ulcerans. Vet Microbiol
153, 323–331.
dos Santos, C. S., dos Santos, L. S., de Souza, M. C., dos Santos
Dourado, F., de Souza de Oliveira Dias, A. A., Sabbadini, P. S.,
Pereira, G. A., Cabral, M. C., Hirata, R., Jr & de Mattos-Guaraldi, A. L.
(2010). Non-opsonic phagocytosis of homologous non-toxigenic
and toxigenic Corynebacterium diphtheriae strains by human U-937
macrophages. Microbiol Immunol 54, 1–10.
Drevets, D. A. (1998). Listeria monocytogenes virulence factors that
stimulate endothelial cells. Infect Immun 66, 232–238.
Feldmann, M., Brennan, F. M. & Maini, R. N. (1996). Role of cytokines
in rheumatoid arthritis. Annu Rev Immunol 14, 397–440.
Finlay, B. B. & Falkow, S. (1997). Common themes in microbial
pathogenicity revisited. Microbiol Mol Biol Rev 61, 136–169.
Gibson, R. L., Lee, M. K., Soderland, C., Chi, E. Y. & Rubens,
C. E. (1993). Group B streptococci invade endothelial cells: type III
capsular polysaccharide attenuates invasion. Infect Immun 61, 478–485.
Goldenberg, D. L. & Reed, J. I. (1985). Bacterial arthritis. N Engl J
Med 312, 764–771.
Guran, P., Mollaret, H. H., Chatelain, R., Gropman, M., Prigent, F. &
Béal, G. (1979). [Septic arthritis due to a nontoxigenic diphtheria
bacillus]. Arch Fr Pediatr 36, 926–929 (in French).
Hirata, R., Jr, Napoleão, F., Monteiro-Leal, L. H., Andrade, A. F. B.,
Nagao, P. E., Formiga, L. C. D., Fonseca, L. S. & Mattos-Guaraldi,
Downloaded from www.microbiologyresearch.org by
IP: 78.47.27.170
On: Thu, 29 Sep 2016 18:49:00
545
R. S. Peixoto and others
A. L. (2002). Intracellular viability of toxigenic Corynebacterium
diphtheriae strains in HEp-2 cells. FEMS Microbiol Lett 215, 115–119.
Pimenta, F. P., Hirata, R. Jr, Rosa, A. C. , Milagres, L. G. & MattosGuaraldi, A. L. (2008). A multiplex PCR assay for simultaneous
Hirata, R., Jr, Souza, S. M. S., Rocha-de-Souza, C. M., Andrade, A. F. B.,
Monteiro-Leal, L. H., Formiga, L. C. D. & Mattos-Guaraldi, A. L. (2004).
detection of Corynebacterium diphtheriae and differentiation between
non-toxigenic and toxigenic isolates. J Med Microbiol 57, 1438–1439.
Patterns of adherence to HEp-2 cells and actin polymerisation by
toxigenic Corynebacterium diphtheriae strains. Microb Pathog 36, 125–130.
Plotkowski, M. C., Saliba, A. M., Pereira, S. H. M., Cervante, M. P. &
Bajolet-Laudinat, O. (1994). Pseudomonas aeruginosa selective
Hirata, R., Jr, Pereira, G. A., Filardy, A. A., Gomes, D. L., Damasco,
P. V., Rosa, A. C., Nagao, P. E., Pimenta, F. P. & Mattos-Guaraldi,
A. L. (2008). Potential pathogenic role of aggregative-adhering
adherence to and entry into human endothelial cells. Infect Immun
62, 5456–5463.
Corynebacterium diphtheriae of different clonal groups in endocarditis. Braz J Med Biol Res 41, 986–991.
Puliti, M., Nizet, V., von Hunolstein, C., Bistoni, F., Mosci, P.,
Orefici, G. & Tissi, L. (2000). Severity of group B streptococcal
arthritis is correlated with b-hemolysin expression. J Infect Dis 182,
Mattos-Guaraldi, A. L. & Formiga, L. C. D. (1998). Bacteriological
824–832.
properties of a sucrose-fermenting Corynebacterium diphtheriae strain
isolated from a case of endocarditis. Curr Microbiol 37, 156–158.
Puliti, M., von Hunolstein, C., Marangi, M., Bistoni, F. & Tissi, L.
(2006). Experimental model of infection with non-toxigenic strains
Mattos-Guaraldi, A. L., Duarte Formiga, L. C. & Pereira, G. A. (2000).
of Corynebacterium diphtheriae and development of septic arthritis.
J Med Microbiol 55, 229–235.
Cell surface components and adhesion in Corynebacterium diphtheriae. Microbes Infect 2, 1507–1512.
Matussek, A., Strindhall, J., Stark, L., Rohde, M., Geffers, R., Buer, J.,
Kihlström, E., Lindgren, P. E. & Löfgren, S. (2005). Infection of
human endothelial cells with Staphylococcus aureus induces transcription of genes encoding an innate immunity response. Scand J
Immunol 61, 536–544.
Menzies, B. E. & Kourteva, I. (2000). Staphylococcus aureus alpha-
toxin induces apoptosis in endothelial cells. FEMS Immunol Med
Microbiol 29, 39–45.
Mishra, B., Dignan, R. J., Hughes, C. F. & Hendel, N. (2005).
Santos, G. S., Lione, V. O. F., Costa E Silva Filho, F. & Nagao, P. E.
(2005). Signal transduction in human endothelial cells induced
by their interaction with group B Streptococci. Int J Mol Med 15,
859–863.
Tiley, S. M., Kociuba, K. R., Heron, L. G. & Munro, R. (1993). Infective
endocarditis due to nontoxigenic Corynebacterium diphtheriae: report
of seven cases and review. Clin Infect Dis 16, 271–275.
Tissi, L., Marconi, P., Mosci, P., Merletti, L., Cornacchione, P., Rosati,
E., Recchia, S., von Hunolstein, C. & Orefici, G. (1990). Experimental
Corynebacterium diphtheriae endocarditis–surgery for some but not
all! Asian Cardiovasc Thorac Ann 13, 119–126.
model of type IV Streptococcus agalactiae (group B streptococcus)
infection in mice with early development of septic arthritis. Infect
Immun 58, 3093–3100.
Ogawa, S. K., Yurberg, E. R., Hatcher, V. B., Levitt, M. A. & Lowy, F. D.
(1985). Bacterial adherence to human endothelial cells in vitro. Infect
Vercellotti, G. M., Lussenhop, D., Peterson, P. K., Furcht, L. T.,
McCarthy, J. B., Jacob, H. S. & Moldow, C. F. (1984). Bacterial
Immun 50, 218–224.
Ott, L., Höller, M., Rheinlaender, J., Schäffer, T. E., Hensel, M. &
Burkovski, A. (2010). Strain-specific differences in pili formation and
adherence to fibronectin and endothelial cells: a possible mechanism
for bacterial tissue tropism. J Lab Clin Med 103, 34–43.
the interaction of Corynebacterium diphtheriae with host cells. BMC
Microbiol 10, 257.
von Hunolstein, C., Alfarone, G., Scopetti, F., Pataracchia, M.,
La Valle, R., Franchi, F., Pacciani, L., Manera, A., Giammanco, A. &
other authors (2003). Molecular epidemiology and characteristics of
Patey, O., Bimet, F., Riegel, P., Halioua, B., Emond, J. P., Estrangin,
E., Dellion, S., Alonso, J. M., Kiredjian, M. & other authors (1997).
Corynebacterium diphtheriae and Corynebacterium ulcerans strains
isolated in Italy during the 1990s. J Med Microbiol 52, 181–188.
Clinical and molecular study of Corynebacterium diphtheriae systemic
infections in France. J Clin Microbiol 35, 441–445.
Edited by: T. Mitchell
546
Downloaded from www.microbiologyresearch.org by
IP: 78.47.27.170
On: Thu, 29 Sep 2016 18:49:00
Microbiology 160