www.fems-microbiology.org

FEMS Microbiology Letters 249 (2005) 95–103

Helicobacter pylori induce neutrophil transendothelialmigration: Role of the bacterial HP-NAP

Mikael Brisslert a,b,*, Karin Enarsson a, Samuel Lundin a, Anna Karlsson b,Johannes G. Kusters c, Ann-Mari Svennerholm a, Steffen Backert d,

Marianne Quiding-Jarbrink a

a Department of Medical Microbiology and Immunology, Goteborg University, Swedenb Department of Rheumatology and Inflammation Research, The Sahlgrenska Academy, Goteborg University, Guldhedsgatan 10A,

413 45 Goteborg, Swedenc Department of Gastroenterology and Hepatology, Erasmus MC-University Medical Center, Rotterdam, The Netherlands

d Department of Medical Microbiology, Otto von Guericke University, Magdeburg, Germany

Received 12 April 2005; received in revised form 2 June 2005; accepted 2 June 2005

First published online 22 June 2005

Edited by A.H.M. van Vliet

Abstract

Continuous recruitment of neutrophils into the inflamed gastric mucosal tissue is a hallmark of Helicobacter pylori infection in

humans. In this study, we examined the ability ofH. pylori to induce transendothelial migration of neutrophils using a transwell system

consisting of a cultured monolayer of human endothelial cells as barrier between two chambers. We showed for the first time that live

H. pylori, but not formalin-killed bacteria, induced a significantly increased transendothelial migration of neutrophils. H. pylori con-

ditioned culture medium also induced significantly increased transendothelial migration, whereas heat-inactivated culture filtrates had

no effect, suggesting that the chemotactic factor was proteinaceous. Depletion of H. pylori-neutrophil activating protein (HP-NAP)

from the culture filtrates resulted in significant reduction of the transmigration. Culture filtrates from isogenic HP-NAP deficient

mutant bacteria also induced significantly less neutrophil migration than culture filtrates obtained from wild-type bacteria.

HP-NAP did not induce endothelial cell activation, suggesting that HP-NAP acts directly on the neutrophils. In conclusion, our results

demonstrate that secreted HP-NAP is one of the factors resulting inH. pylori induced neutrophil transendothelial migration. We pro-

pose that HP-NAP contributes to the continuous recruitment of neutrophils to the gastric mucosa of H. pylori infected individuals.

� 2005 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved.

Keywords: HP-NAP; Helicobacter pylori; Transmigration; Endothelial cells; Inflammation

1. Introduction

Infection with Helicobacter pylori gives rise to active

chronic gastritis in virtually all infected subject. In some

individuals, the infection eventually leads to the devel-

0378-1097/$22.00 � 2005 Federation of European Microbiological Societies

doi:10.1016/j.femsle.2005.06.008

* Corresponding author. Tel.: +46 0 31 3424021; fax: +46 0 31

823925.

E-mail address: mikael.brisslert@rheuma.gu.se (M. Brisslert).

opment of ulcer disease or gastric adenocarcinoma. Sev-

eral bacterial and host factors seem to play important

roles in determining the severity of the inflammation

and the outcome of the infection [1,2]. Continuous

recruitment of neutrophils into the inflamed gastricmucosal tissue is a hallmark in human H. pylori

infections. A strong correlation exists between gastric

neutrophil infiltration, mucosal damage, and develop-

ment of duodenal ulcer (DU) disease in H. pylori

. Published by Elsevier B.V. All rights reserved.

Table 1

Characterization of H. pylori strains used in the study

Strain cag PAIa VacAb Origin

Hel 312 + s1/m1c ASd

Hel333 � s2/m2 DUe

96 M. Brisslert et al. / FEMS Microbiology Letters 249 (2005) 95–103

infections [3,4]. There is also strong evidence that H. py-

lori is able to activate neutrophils; in some studies, the

strains isolated from DU patients were shown to be

more potent activators of neutrophils than other isolates

[3,5–9]. One of the bacterial factors shown to mediate

neutrophil activation is the H. pylori neutrophil activat-ing protein (HP-NAP) [5]. Evans et al. [5,9] first charac-

terized the HP-NAP protein and showed that it

stimulated the release of reactive oxygen species by hu-

man neutrophils. HP-NAP is dodecameric and is able

to bind iron. It has sequence similarities with iron-bind-

ing proteins in other bacterial species [10,11]. HP-NAP

is a cytosolic protein expressed by virtually all H. pylori

isolates [12]. It is released into the supernatant by autol-ysis, upon which a fraction of HP-NAP is then bound to

carbohydrates on the external surface of the outer mem-

brane [5,13,14].

While the mechanisms behind the continuous recruit-

ment of neutrophils to H. pylori infected tissues in vivo

remain to be fully understood, endothelial cell function

and activation appear to be key components in this pro-

cess. Endothelial cells are involved in a wide range ofphysiological reactions including blood coagulation,

blood pressure regulation, and inflammatory responses.

One important function is to direct circulating leuko-

cytes to sites of inflammation via a series of steps that

are independent of the stimulatory agent [15]. During

the inflammatory process, or upon in vitro stimulation

with TNF-a, IL-1 or bacterial products, endothelial cellsrespond rapidly (within a few hours) by upregulation ofadhesion molecules such as VCAM-1, ICAM-1, and E-

selectin as well as production of proinflammatory cyto-

kines and chemokines [16], all of which contribute to

leukocyte recruitment.

We have shown in a previous study that endothelial

cells are strongly activated by H. pylori, resulting in in-

creased production of neutrophil recruiting chemokines

and expression of adhesion molecules [17]. To furthercharacterize this response, we investigated the effect of

H. pylori on the migration of neutrophils through cul-

tured endothelial monolayers. We showed that live bac-

teria, secreted bacterial proteins containing HP-NAP,

and purified HP-NAP all resulted in an increased trans-

endothelial migration of neutrophils. In contrast, cul-

ture filtrates depleted of HP-NAP, as well as culture

filtrates produced by isogenic HP-NAP deficient mu-tants were less effective in inducing neutrophil transen-

dothelial migration.

P1 + s1/m2 NUDf

P12 + s1/m1 DU

SS1 + s2/m2 DU, mouse adapted

a cag PAI, cytotoxin-associated gene pathogenicity island.b VacA, vacuolating toxin A.c s, sequence region; m, middle region.d AS, asymptomatic carrier.e DU, duodenal ulcer patient.f NUD, non-ulcer dyspepsia patient.

2. Materials and methods

2.1. H. pylori strains

H. pylori strains were isolated from the duodenum of

adults suffering from DU (strain Hel 333, P12, Table 1),

from the stomach of adult asymptomatic subject (strain

Hel 312, Table 1) or from a patient suffering from a non-

ulcer dyspepsia (strain P1, Table 1), as previously de-

scribed [3,18]. Isolates were stored in a freeze-drying

medium (tryptic soy broth with 15% glycerol) at

�80 �C. The mouse adapted strain SS1 was originallyprovided by Dr. Adrian Lee, Sydney, Australia [19].

Characteristics of these strains [12,17] are summarized

in Table 1.

2.2. Generation of isogenic HP-NAP deficient mutants

Isogenic P1D nap and P12D nap knock-out mutants

were constructed by insertion of a chloramphenicolresistance gene cassette (CmR, 1 kb fragment of plasmid

pTnMax1) or kanamycin resistance gene cassette (Aph-

3, 1.4 kb fragment of plasmid pILL-600) according to a

standard protocol [18]. For this purpose, we amplified a

fragment of 1.8 kb containing the nap gene and flanking

regions with the following primers: 179F 5 0-TTT-

TTGAAGGGCCAATCTTAGAA and 179R 5 0-AGC-

GACGAAGGGTTTTTTG. This product was ligatedinto pCR2.1 vector (Invitrogen, Karlsruhe, Germany)

and the correct integration of the CmR or Aph-3 cas-

settes in the nap gene was checked by standard PCR.

Both mutants were created with terminatorless cassettes,

and are non-polar.

2.3. H. pylori culture conditions

H. pylori were cultured routinely on Columbia-iso

agar plates at 37 �C in a microaerophilic milieu (3%

O2, 5% CO2 and 92% N2). After four days of incubation,

the bacteria were resuspended in 3–5 ml of PBS. The

optical density (OD) of the suspension was determined

(Shimadzu UV-1201, Lambda Polynom, Stockholm,

Sweden), and adjusted to a final OD600nm = 1.0 in

PBS, corresponding to approximately 5 · 109 H. pylori

bacteria ml�1 (5 · 109 CFUs ml�1). In some experi-

ments, PBS resuspended bacteria inactivated by treat-

ment with 0.01 M Formaldehyde at 37 �C for 2 h and

then at room temperature over night, were used.

M. Brisslert et al. / FEMS Microbiology Letters 249 (2005) 95–103 97

2.4. Preparation of H. pylori secreted proteins (culture

filtrate)

H. pylori strains were grown and resuspended as de-

scribed above. Bacterial suspension (2.5 ml) was added

to 48 ml of M200 medium (Cascade Biologics Inc. OR,USA) containing Brucella broth (Difco Laboratories,

Detroit, MA, USA), 5% new born calf serum (Bio-

chrome AG, Berlin Germany), 10lg ml�1 Vancomycin,

5 lg ml�1 Trimothoprim, and 20 U ml�1 Polymyxin B,

and the mixture was then incubated for 6.5 h at 37 �Cin 5% CO2. The time point was chosen based on our pre-

vious study showing that human umbilical vein endothe-

lial cells were strongly stimulated by secreted H. pylori

proteins produced by a 6.5 h culture [17]. The cultures

were sterile filtered, and one aliquot of each culture

filtrate was heat-inactivated at 85 �C for 35 min. All

filtrates were then stored at �80 �C until use.

2.5. Depletion of HP-NAP from H. pylori culture filtrate

The culture filtrate was depleted of HP-NAP usingmagnetic beads (Dynal, Norway) coupled to IgG1

monoclonal antibody (mAb) against HP-NAP (NAP

6:1), the preparation of which has been described previ-

ously [12]. Sixty milligrams of beads coated with sheep

antibodies to mouse IgG1 were incubated with 120 lgIgG1 mAb NAP 6:1 for 4 h, washed on a magnet, resus-

pended in 2 ml of the culture filtrate and incubated for

4 h at 4 �C on a waddle table. The beads were removedwith a magnet and the HP-NAP depleted culture filtrate

was stored at �80 �C until use. Beads incubated with an

IgG1 mAb reacting with cholera toxin were used as con-

trols for unspecific binding. HP-NAP depletion from

culture filtrates was confirmed by Western blot analysis

using the same NAP-mAb. In brief, samples were sepa-

rated on a 18% Tris–Glycine SDS–PAGE, and blotted

onto a PVDF membrane using standard procedures.After blocking in 2% milk powder diluted in blotting

buffer (25 mM Tris–HCl, pH 7.4, 150 mM NaCl, 0.1%

Tween 20), the membrane was probed with monoclonal

mouse anti-NAP [12] diluted 1:5 in blotting buffer over

night at 4 �C. Thereafter the membrane was washed

and mouse immunoglobulins were detected with biotin-

ylated rabbit anti-mouse Ig (1:2000, DAKO, E0413),

followed by a combination of HRP-conjugated donkeyanti-rabbit Ig (1:2000, Amersham, NA9340) and Strep

ABComplex/HRP (DAKO). Peroxidase activity was

subsequently detected by ECL chemiluminescense

(Amersham) according to the manufacturer�s instruc-

tions and exposure for 15 min on Fuji NP100 film

(Fig. 3B).

The distribution of immunogenic H. pylori proteins

in the culture filtrates was examined by Western blottingusing an H. pylori-reactive polyclonal mouse serum pool

(data not shown). With the exception of HP-NAP, the

relative distribution of other filtrate proteins remained

unchanged during incubation with the beads.

2.6. Production of rNAP

rNAP was produced in E. coli as described and pro-vided by Dr. Susanne Nystrom [12]. Briefly, E. coli cells

were transformed with an expression vector containing

the H. pylori nap gene under the control of the T7 pro-

moter. After disruption of the cells, the soluble proteins

were applied to a Q-Sepharose column (Pharmacia Bio-

Tech, Uppsala, Sweden) in 25 mM Tris–HCl and

50 mM NaCl (pH 8.0). rNAP was found in the break-

through. The fractions containing rNAP were concen-trated and then purified by gel filtration through a

Superdex 200 column (Pharmacia). rNAP eluted as a

single peak. Purity of the protein was confirmed by

SDS–PAGE and Coomassie staining as well as immuno-

blotting, using mAbs specific for rNAP [20].

2.7. Isolation of neutrophils

Human neutrophils were isolated from buffy coats, or

alternatively, from heparinised whole blood obtained

from healthy blood donors. After dextran sedimentation

at 1g, hypotonic lysis of the remaining erythrocytes, and

subsequent centrifugation in a Ficoll-Paque gradient

[21], the neutrophils were washed twice and resuspended

to a final volume of 1 · 107 ml�1 in Krebs–Ringer phos-

phate buffer (KRG) containing glucose (10 mM), Ca2+

(1 mM), and Mg2+ (1.5 mM), and kept on ice up to

1.5 h until use.

2.8. Sub-culturing and stimulation of endothelial cells

Human umbilical vein endothelial cells (HUVEC),

passage 1, were purchased from Cascade Biologics Inc.

Cells were grown in 75 cm2 plastic bottles (Nunc, Ros-kilde, Denmark) in M200 medium supplemented with

low serum growth supplement, (LSGS) and penicillin–

streptomycin–amphotericin (PSA) (all from Cascade

Biologics Inc.). Medium was changed every 48 h during

culturing at 37 �C, 5% CO2. The medium used in this

study contained routinely less than 0.03 EU ml�1 endo-

toxin as confirmed by Limulus test.

Confluent monolayers were trypsinized using0.25 mg ml�1 trypsin/EDTA (Cascade Biologics Inc.)

and split at a ratio of 1:3. At passage 4, the cells were

frozen in fetal calf serum containing 10% DMSO

(Sigma) at a density of 106 cells ml�1. HUVEC within

passage 4–7 were used throughout the experiments [22]

to minimize spontaneous morphological or functional

changes as a result of continuous passaging.

HUVEC were grown for four days, after which themedium was replaced with fresh antibiotic-free

medium containing 10 lg ml�1 rHP-NAP or 5 · 108 live

Fig. 1. FACS-analysis of migrating neutrophils. 105 cells were

collected after transendothelial migration against culture medium

(A), 10�8 M fMLF (B) or 5 · 108 live H. pylori bacteria (C) and

enumerated by flow cytometry. Axes show forward (FSC-H) and side

scatter (SSC-H) characteristics. In all figures, TrueCount� beads can

be seen as indicated by arrow and text in (B). Following encounter

with bacteria (C), neutrophils increase their side scatter signal. The

figure shows one representative example of all FACS plots analyzed in

this study.

98 M. Brisslert et al. / FEMS Microbiology Letters 249 (2005) 95–103

wild-type P1 or P1Dnap. After 6.5 h incubation, the cul-

ture medium was collected and HUVEC were detached

from the plate using trypsin. The endothelial expression

of ICAM-1, E-selectin and VCAM-1 was determined by

flow cytometry, and the IL-8 concentration in the cul-

ture medium analyzed by ELISA, as previously de-scribed [17].

2.9. Transendothelial migration of neutrophils

HUVEC (3 · 105 cells) were added to collagen-coated

nets of Transwell chambers (Costar, Badhoevedorp, The

Netherlands) with a growth area of 4.7 cm2, and allowed

to grow to confluence. When HUVEC reached a conflu-ent monolayer after 24 h as confirmed bymicroscopy, the

medium was changed and incubation continued for an-

other hour at 37 �C. This creates a system with two sep-

arate chambers, where transport through the

endothelial monolayer is the only means of communica-

tion between the chambers, thereby allowing a chemotac-

tic gradient of the substances added to the lower chamber

to be established. Stimuli were placed in the lower cham-ber in 1 ml of M200 medium, and purified neutrophils

(106 cells in 100 ll KRG medium) were added in the

upper chamber to a final volume of 1 ml. The addition

of bacteria was carried out 1 h before the addition of neu-

trophils to allow a gradient of secreted products to be

established, whereas all other stimulatory agents were

added at the same time as the neutrophils. The chemo-

attractant formyl-methionyl-leucyl-phenylalanine (fMLF)was used at a concentration of 10�8 M as a positive con-

trol in all experiments and was added in a volume of

100 ll. All stimuli (including bacteria) used were present

in the lower chamber throughout the experiments.

After 90 min (optimal duration for reducing sponta-

neous migration and yet achieving maximum migration

of attracted cells) of transendothelial migration, the neu-

trophil containing medium in the lower chamber was re-moved and collected in polypropylene plastic tubes on

ice. Meanwhile, 2 ml of 0.25% Trypsin/EDTA solution

was added in order to detach neutrophils that were

adhering to the plastic well or to the bottom of the filter.

The cells harvested were sedimented, transferred to a 96-

well polypropylene plate and stained for the expression

of the complement receptor 3 (CR3) by adding

10 lg ml�1 of the PE-conjugated anti-CR3 antibody(BD, Erembodegem, Belgium) in a total volume of

100 ll PBS containing 1.46 g l�1 EDTA, 2.5 g l�1 bovine

serum albumin, 0.2 g l�1 NaN3 and 2% AB+ human ser-

um (FACS-wash), and incubated on ice for 35 min.

After washing the pellet twice in FACS-wash buffer

the transmigrated cells were fixed in Cell Fix� (BD)

and enumerated by flow cytometry (Facs-Calibur, BD)

using True Count� beads (BD). Neutrophils were iden-tified by their forward and side scatter characteristics, as

illustrated in Fig. 1. Since these parameters vary with the

agent used for stimulation, expression of CR3, which is

upregulated during transendothelial migration, was used

to confirm that gated events were indeed neutrophils.

2.10. Statistical analysis

Statistical evaluations were performed using the

Mann–Whitney test and the Prism software.

M. Brisslert et al. / FEMS Microbiology Letters 249 (2005) 95–103 99

3. Results

3.1. Live H. pylori and culture filtrates induce neutrophil

transendothelial migration

Neutrophil transendothelial migration towardsH. py-

lori bacteria was analyzed in the Transwell system and

quantified using flow cytometry. Five well-characterised

H-pylori strains (Hel 312, Hel 333, SS1, P1 and P12;

Table 1) were used in our experiments and gave similar

results. One representative FACS-dot plot is shown in

Fig. 1. Titration experiments demonstrated a dose

dependent migration of neutrophils towards H. pylori,

with 5 · 108 of bacteria giving a strong and consistent re-sponse (data not shown). This amount of bacteria was

used in all further experiments. H. pylori strain Hel 312

carrying the cag pathogenicity island (cag PAI+) and

vacA s1/m1 allele induced transendothelial migration of

significantly (p < 0.01) more neutrophils than the

unstimulated control (approximately 25% of the added

neutrophils compared to about 5% in the control) (Fig.

2). The H. pylori strain Hel 333 (cag PAI�, vacA s2/m2allele) also induced a substantially increased neutrophil

transendothelial migration, where approximately 70%

of the total added neutrophils had migrated (p < 0.01,

data not shown). In contrast, the same number of forma-

lin-inactivated Hel 312 induced significantly less

(p < 0.05) migration compared to live Hel 312 (Fig. 2).

Since live but not dead bacteria could induce neutro-

phil migration, we sought to test if the effect was caused

Fig. 2. Induction of neutrophil transendothelial migration by culture

filtrates from H. pylori. 5 · 108 live or formalin killed H. pylori

bacteria, culture filtrate, or heat-inactivated culture filtrate, were added

to the lower wells of Transwell cultures and 106 neutrophils to the

upper wells. fMLF (10�8 M) was used as a positive control. Bars

indicate geometric mean + SEM of percentage of transmigrated

neutrophils. Statistical evaluation between negative control and

stimulated samples was performed by the Mann–Whitney test.

***p < 0.001, **p < 0.01, and *p < 0.05. Data are obtained in tripli-

cates from at least three independent experiments.

by secretion of neutrophil chemotactic factors by bacte-

ria into the culture filtrates. We chose to use strain Hel

312, since Hel 333 grows poorly and since the presence

of the CagPAI did not seem to influence induction of

migration. In addition, H. pylori induced chemokine

and adhesion molecule upregulation on endothelial cellsis independent of the CagPAI (16). Indeed, when culture

filtrate from Hel 312 bacteria was used as stimulatory

agent, we could detect a significant level of approxi-

mately 20% neutrophil transendothelial migration

(p < 0.01), whereas the positive control fMLF induced

approximately 40% migration (p < 0.001) (Fig. 2). Heat

inactivation of the H. pylori filtrate at 85 �C for 35 min

significantly abolished the capacity of the culture super-natant to induce transendothelial migration (Fig. 2),

indicating that the stimulatory agents are of proteina-

ceous nature. Similar results were obtained from culture

filtrates of strain Hel 333. As a negative control, culture

medium treated in the same way but without the addi-

tion of bacteria did not affect neutrophil transendothe-

lial migration (data not shown).

3.2. HP-NAP is a potent chemotactic factor for

neutrophils in culture filtrates of H. pylori

Since H. pylori culture filtrates induced neutrophil

transendothelial migration to the same extent as bacte-

ria, we examined if HP-NAP, a well known neutrophil

activating factor [5,7,8], is responsible for the induction

of neutrophil migration. rNAP was added to the lowerchambers at a concentration of 10 lg ml�1 and neutro-

phil transendothelial migration was measured. This con-

centration was selected based on in vitro titration

experiments in our model system to give the highest

induction of transendothelial migration (data not

shown). When rNAP was added to the assay, almost

25% (p < 0.01) of the added neutrophils transmigrated

(Fig. 3A). Again, the positive control fMLF inducedthe highest degree of transendothelial migration, attract-

ing about 35% (p < 0.001) of the added neutrophils in

this set of experiments (Fig. 3A). To avoid the possibil-

ity of LPS contamination, 200 lg ml�1 rNAP was incu-

bated with 20 lg ml�1 of polymyxin B for 1 h at room

temperature. After incubation, 100 ll of the HP-NAP–

polymyxin B mixture was added to each well, giving a

final concentration of 10 lg ml�1 HP-NAP and1 lg ml�1 polymyxin B. Limulus endotoxin test per-

formed before and after polymyxin treatment showed

that the HP-NAP preparation contained less than

1.4 EU ml�1 of LPS corresponding to 70 pg ml�1 E. coli

LPS, far less than the LPS concentration needed for

induction of transendothelial migration (10,000–

25,000 pg ml�1) [23]. Polymyxin itself had no effect on

transendothelial migration (data not shown).Thus, our findings show that rNAP is a potent indu-

cer of transendothelial migration. We next sought to

Fig. 3. Induction of neutrophil transendothelial migration by HP-

NAP. (A) rNAP (10 lg ml�1) was added to the lower wells of

Transwell cultures and 106 neutrophils to the upper wells. fMLF

(10�8 M) was used as a positive control. Bars indicate geometric

mean + SEM of percentage of transmigrated neutrophils. In (B)

culture filtrate, HP-NAP depleted (depl.) culture filtrate, or culture

filtrate treated with beads coated with a mAb against cholera toxin

(sham depleted) were added to the lower wells of Transwell cultures,

and 106 neutrophils to the upper wells. Bars indicate geometric

mean + SEM of percentage of transmigrated neutrophils relative to the

migration induced by the non-treated culture filtrate. Statistical

evaluation between negative control and stimulated samples was

performed by the Mann–Whitney test. ***p < 0.001, **p < 0.01,

*p < 0.05. Data are obtained in triplicates from at least three

independent experiments. (C) Depletion of HP-NAP from the culture

filtrate is shown as a radiographic photo were lane 1 is rNAP, lane 2 is

the HP-NAP depleted culture filtrate from strain Hel 312 and lane 3 is

the control of non-depleted culture filtrate from the same strain.

Fig. 4. Induction of neutrophil transendothelial migration by culture

filtrates from HP-NAP deficient mutants. Culture filtrates from H.

pylori wild-type (P1 or P12) or HP-NAP deficient mutants (P1Dnap or

P12Dnap) were added to the lower wells of Transwell cultures and 106

neutrophils to the upper wells. Bars indicate geometric mean + SEM of

percentage of transmigrated neutrophils expressed relative to the

migration induced by the respective wild type strains. Statistical

evaluation was performed by the Mann–Whitney test. *p < 0.05 and

**p < 0.01. Data are obtained in triplicates from at least three

independent experiments.

100 M. Brisslert et al. / FEMS Microbiology Letters 249 (2005) 95–103

deplete Hel 312 culture filtrate of HP-NAP using mag-

netic beads coated with a mAb against HP-NAP. This

treatment dramatically reduced the capacity of the fil-

trate to induce neutrophil transendothelial migration

(Fig. 3B). In fact, the HP-NAP depleted filtrate induced

a 50% reduction in migration (p < 0.001) compared to

the untreated culture filtrate. Beads coated with amAb directed against the Vibrio cholerae toxin B sub-

unit were used as a negative control and had no effect

on the capacity to induce transendothelial migration

(Fig. 3B). The removal of HP-NAP from culture filtrates

is shown as a radiographic image of a western blot gel in

Fig. 3C.

3.3. Culture filtrates from HP-NAP deletion mutants but

not live mutants have decreased ability to recruit

neutrophils

To further assess the importance of HP-NAP, we first

tried to create an HP-NAP mutant in strain Hel312, butthis strain proved impossible to transform, possibly be-

cause it is a clinical isolate. We therefore constructed iso-

genic HP-NAP deletion mutants from strains P1 and

P12. There were no significant differences between the

wild-type and the mutant strains when analyzing the

induction of neutrophil transmigration by intact bacteria

(data not shown). However, when the effect of culture fil-

trates was examined, filtrates produced by the two HP-NAP mutants were reproducibly found to induce

approximately 30% less (p = 0.0025, and 0.0286, respec-

tively) neutrophil migration compared to culture filtrates

obtained from wild-type bacteria (Fig. 4).

3.4. HP-NAP does not activate endothelial cells

We then asked whether the effect of HP-NAP wasmediated by a direct interaction with the neutrophils

or if it was a secondary effect mediated by the endothe-

lial cells. For this purpose, we measured the production

of the neutrophil-recruiting chemokines IL-8, and

GRO-a, and surface expression of ICAM-1, VCAM-1

and E-selectin by endothelial cells after incubation with

10 lg ml�1 recombinant HP-NAP for 6.5 and 24 h,

respectively. This treatment had no effect on chemokinesecretion or adhesion molecule expression (data not

shown). In addition, the P1 wild-type and P1D nap

mutant were used for endothelial stimulation. Both the

wild-type and mutant strains induced a slightly

Fig. 5. Effect of HP-NAP on endothelial cells. 5 · 108 live H. pylori P1 wild-type or HP-NAP deficient mutant P1Dnap were added to cultured

endothelial cells and incubated for 6 h whereafter IL-8 production (A) was analyzed using ELISA and the expression of ICAM-1 (B), E-selectin (C)

and VCAM-1 (D) was evaluated by flow cytometry. Data are obtained in triplicates from at least three independent experiments. Bars indicate

geometric mean + SEM of percentage of transmigrated neutrophils. Statistical evaluation was performed by the student�s t test. ***p < 0.001,

**p < 0.01, and *p < 0.05. There was no significant difference between wild-type and mutant bacteria in any of the assays.

M. Brisslert et al. / FEMS Microbiology Letters 249 (2005) 95–103 101

increased production of IL-8 (Fig. 5) and increased

expression of the adhesion molecules ICAM-1,

VCAM-1, and E-selectin compared to unstimulated

controls (Fig. 5). Since the effect of the HP-NAP defi-

cient mutant was indistinguishable from that of the

wild-type bacteria, the results indicate that HP-NAP ex-erts transmigratory effects on the neutrophils themselves

rather than on the endothelial cells.

4. Discussion

Endothelial cells have a key function in the recruit-

ment of neutrophils during inflammation due to theirability to produce cytokines and chemokines and

expressing adhesion molecules relevant for transmigra-

tion. Circulating neutrophils in turn are sensitive to che-

motactic factors such as bacterial peptides and

chemokines produced at the site of inflammation. When

encountering such stimuli, the neutrophils attach to the

endothelium, leave the blood stream, and migrate to-

wards increasing amounts of stimulus. During H. pylori

infection, a pronounced and continuous recruitment of

neutrophils to the inflamed gastric mucosa has been ob-

served [1,2]. Indeed, neutrophil infiltration has been sug-

gested to be one of the factors leading to mucosal

damage and DU formation [3,4]. In a recent study, we

showed that live H. pylori was able to activate endothe-

lial cells to express several adhesion molecules and neu-

trophil recruiting CXC-chemokines [17]. In this study,

using the transwell system where cultured endothelial

cells creates two separate compartments allowing pas-

sage from one side to another only through the endothe-

lial monolayer, we extend these findings by showing

that live H. pylori as well as bacterial culture filtrates

and purified HP-NAP can induce substantial transendo-thelial migration of human neutrophils in a manner

independent of a functional cag PAI or VacA. Further-

more, the chemotactic effect occurred without prior trig-

gering or other stimulatory signals.

Broom et al. [24] previously showed that H. pylori

can synthesize and secrete formylated peptides into cul-

ture medium. In addition, a study by Craig et al. [25]

showed that H. pylori secrete heat-stable and acid resis-tant neutrophil chemotactic factors into the culture

medium, which have a molecular weight of less than

3 kDa, indicating that they might be the same peptides

as those detected by Broom et al. [24]. In addition, Niel-

sen and Andersen [6] showed that sonicates from H. py-

lori contain chemotactic proteins with a molecular

weight between 25 and 35 kDa. The latter results are

in agreement with our finding that H. pylori culture fil-trates induce extensive neutrophil transendothelial

migration, and that heat treatment inhibits this activity,

indicating that the stimulatory factor(s) are of proteina-

ceous nature. Previous studies [6] have shown chemotac-

tic activity of H. pylori sonicates or water extracts, but

this is the first study to demonstrate the actual release

of chemotactic proteins from H. pylori to the

environment.

102 M. Brisslert et al. / FEMS Microbiology Letters 249 (2005) 95–103

HP-NAP has been shown to be chemotactic to hu-

man neutrophils [9], and we therefore depleted HP-

NAP from the Hel 312 culture filtrate. This treatment

substantially decreased their chemotactic effect. Further-

more, purified HP-NAP induced a large increase in neu-

trophil transendothelial migration. However, when weassessed the importance of HP-NAP using intact bacte-

ria, there was no significant difference in the induction of

migration between wild-type and HP-NAP mutant

strains, whereas culture filtrates from the knock-out

strains showed significantly reduced ability to attract

neutrophils as compared to culture filtrates from wild-

type H. pylori. These seemingly contradictory findings

might stem from the fact that intact bacteria are morecomplex than culture filtrate in terms of protein compo-

sition – the former might contain other chemotactic fac-

tors in addition to HP-NAP. Furthermore, strain

differences may explain why HP-NAP exhibits high lev-

els of chemotactic activity in Hel 312 derived culture fil-

trates but only a lower level of activity in the P1 and P12

culture filtrates. We therefore postulate that P1 and P12

may release other chemotactic factors, which are notproduced by Hel 312. It is interesting to note that a cor-

responding difference has recently been observed when

analyzing the effect of VacA blocking of NF-AT in T

cells [26]. In this system, while supernatants from wild-

type bacteria and not vacA mutants inhibited T cell pro-

liferation, no difference was observed with either live

wild-type H. pylori or vacA mutants.

In the original article, Evans et al. [5] describe in-creased adherence of neutrophils to HUVEC after incu-

bation with H. pylori water extracts, an effect that could

partly be blocked by antibodies directed against HP-

NAP. However, we could not demonstrate that purified

HP-NAP activates the HUVEC to express adhesion mol-

ecules or chemokines. Furthermore, there was no differ-

ence in the induction of endothelial activation

molecules between wild-type andHP-NAPmutantH. py-

lori in our system. These findings suggest that HP-NAP is

transported across the endothelial monolayer and acts di-

rectly on the neutrophils, which could in turn influence

the endothelial cells to promote transmigration. We have

thus extended previous studies on HP-NAP-induced che-

motaxis [9] showing that this protein is highly effective in

inducing transendothelial migration of neutrophils. In-

deed, our findings indicate thatHP-NAP is one of thema-jor inducers of transendothelial migration of neutrophils

present in H. pylori culture filtrates. However, we do not

discount the possibility that additional chemotactic H.

pylori products could act in parallel, or in synergy with

HP-NAP inmediating neutrophil chemotaxis in vivo. Be-

sides, chemokines such as IL-8 and ENA-78, which have

been detected in H. pylori positive individuals, are also

likely to contribute to neutrophil recruitment [27,28].Our results also showed that sufficient amounts of

HP-NAP to induce transendothelial migration were re-

leased into the culture medium by live H. pylori within

only 6 h of incubation. It is not known if HP-NAP

can gain access to the lamina propria during infection

in vivo. Given that H. pylori urease is released in a sim-

ilar fashion, and can be detected in the infected gastric

tissues, we propose that HP-NAP might be able to pen-etrate into the lamina propria. In addition, active trans-

port of HP-NAP across epithelial monolayers has

recently been demonstrated in vitro [29]. Our laboratory

has recently shown that expression of HP-NAP is much

higher in vivo than in vitro, suggesting that HP-NAP is

an important factor in H. pylori induced gastritis [30].

Furthermore, no clinical isolate of H. pylori lacking

the gene encoding HP-NAP has been reported, furthersupporting the postulation that this protein confers a

selective advantage for the bacteria in vivo.

In conclusion, H. pylori have the ability to induce sig-

nificant transendothelial migration of neutrophils

through a monolayer of human endothelial cells. Live

bacteria and bacterial filtrates may induce neutrophil

migration in different manners. HP-NAP thus may play

a significant role in neutrophil migration when presentin bacterial filtrates. Therefore, we suggest that in vivo

where live bacteria in the lamina propria are scarce,

HP-NAP may be a major factor contributing to the con-

tinuous recruitment of neutrophils to the gastric mucosa

in H. pylori-associated gastritis.

Acknowledgements

We thank Dr. Susanne Nystrom for kindly providing

the rNAP used in this study, Dr. Mattias Magnusson for

help with western blot analyses, and Dr. Terry Kwok for

critical reading of the manuscript. This study was sup-

ported by grants from the Swedish Science Council

(Grant No. 06X-13428) and the Faculty of Medicine

at Goteborg University.

References

[1] Go, M.F. (1997) What are the host factors that place an

individual at risk for Helicobacter pylori-associated disease?.

Gastroenterology 113, S15–20.

[2] Del Giudice, G., Covacci, A., Telford, J.L., Montecucco, C. and

Rappuoli, R. (2001) The design of vaccines against Helicobacter

pylori and their development. Annu. Rev. Immunol. 19, 523–

563.

[3] Hamlet, A., Thoreson, A.C., Nilsson, O., Svennerholm, A.M. and

Olbe, L. (1999) Duodenal Helicobacter pylori infection differs in

cagA genotype between asymptomatic subjects and patients with

duodenal ulcers. Gastroenterology 116, 259–268.

[4] Davies, G.R., Simmonds, N.J., Stevens, T.R., Sheaff, M.T.,

Banatvala, N., Laurenson, I.F., Blake, D.R. and Rampton, D.S.

(1994) Helicobacter pylori stimulates antral mucosal reactive

oxygen metabolite production in vivo. Gut 35, 179–185.

[5] Evans Jr., D.J., Evans, D.G., Takemura, T., Nakano, H.,

Lampert, H.C., Graham, D.Y., Granger, D.N. and Kvietys,

M. Brisslert et al. / FEMS Microbiology Letters 249 (2005) 95–103 103

P.R. (2001) Characterization of a Helicobacter pylori neutrophil-

activating protein. Infect. Immun. 63, 2213–2220.

[6] Nielsen, H. and Andersen, L.P. (1992) Chemotactic activity of

Helicobacter pylori sonicate for human polymorphonuclear leu-

cocytes and monocytes. Gut 33, 738–742.

[7] Nielsen, H. and Andersen, L.P. (1995) Activation of phagocytes

by Helicobacter pylori correlates with the clinical presentation of

the gastric infection. Scand. J. Infect. Dis. 27, 347–350.

[8] Rautelin, H., Blomberg, B., Fredlund, H., Jarnerot, G. and

Danielsson, D. (1993) Incidence of Helicobacter pylori strains

activating neutrophils in patients with peptic ulcer disease. Gut

34, 599–603.

[9] Satin, B., Del Giudice, G., Della Bianca, V., Dusi, S., Laudanna,

C., Tonello, F., Kelleher, D., Rappuoli, R., Montecucco, C. and

Rossi, F. (2000) The neutrophil-activating protein (HP-NAP) of

Helicobacter pylori is a protective antigen and a major virulence

factor. J. Exp. Med. 191, 1467–1476.

[10] Tonello, F., Dundon, W.G., Satin, B., Molinari, M., Tognon, G.,

Grandi, G., Del Giudice, G., Rappuoli, R. and Montecucco, C.

(1999) The Helicobacter pylori neutrophil-activating protein is an

iron-binding protein with dodecameric structure. Mol. Microbiol.

34, 238–246.

[11] Zanotti, G., Papinutto, E., Dundon, W., Battistutta, R., Seveso,

M., Giudice, G., Rappuoli, R. and Montecucco, C. (2002)

Structure of the neutrophil-activating protein from Helicobacter

pylori. J. Mol. Biol. 323, 125–130.

[12] Thoreson, A.C., Hamlet, A., Celik, J., Bystrom, M., Nystrom, S.,

Olbe, L. and Svennerholm, A.M. (2000) Differences in surface-

exposed antigen expression between Helicobacter pylori strains

isolated from duodenal ulcer patients and from asymptomatic

subjects. J. Clin. Microbiol. 38, 3436–3441.

[13] Namavar, F., Sparrius, M., Veerman, E.C., Appelmelk, B.J. and

Vandenbroucke-Grauls, C.M. (1998) Neutrophil-activating pro-

tein mediates adhesion of Helicobacter pylori to sulfated carbo-

hydrates on high-molecular-weight salivary mucin. Infect.

Immun. 66, 444–447.

[14] Teneberg, S., Miller-Podraza, H., Lampert, H.C., Evans Jr., D.J.,

Evans, D.G., Danielsson, D. and Karlsson, K.A. (1997) Carbo-

hydrate binding specificity of the neutrophil-activating protein of

Helicobacter pylori. J. Biol. Chem. 272, 19067–19071.

[15] Muller, W.A. (1999) Leukocyte-endothelial cell adhesion mole-

cules in transendothelial migration In: Inflammation: Basic

Principles and Clinical Correlates (Gallin, J. and Snyderman,

R., Eds.), 3rd edn. Lippincott Williams & Wilkins, Philadelphia.

[16] Silverstein, R. (2000) The vascular endothelium In: Inflammation:

Basic Principles and Clinical Correlates (Gallin, J. and Snyder-

man, R., Eds.), 3rd edn, pp. 207–227. Lippincott Williams &

Wilkins, Philadelphia.

[17] Innocenti, M., Thoreson, A.C., Ferrero, R.L., Stromberg, E.,

Bolin, I., Eriksson, L., Svennerholm, A.M. and Quiding-Jarbrink,

M. (2002) Helicobacter pylori-induced activation of human

endothelial cells. Infect. Immun. 70, 4581–4590.

[18] Backert, S., Ziska, E., Brinkmann, V., Zimny-Arndt, U., Fau-

connier, A., Jungblut, P.R., Naumann, M. and Meyer, T.F.

(2000) Translocation of the Helicobacter pylori CagA protein in

gastric epithelial cells by a type IV secretion apparatus. Cell

Microbiol. 2, 155–164.

[19] Lee, A., O�Rourke, J., De Ungria, M.C., Robertson, B., Dask-

alopoulos, G. and Dixon, M.F. (1997) A standardized mouse

model of Helicobacter pylori infection: introducing the Sydney

strain. Gastroenterology 112, 1386–1397.

[20] Lindholm, C., Osek, J. and Svennerholm, A.M. (1997) Quanti-

fication of conserved antigens in Helicobacter pylori during

different culture conditions. Infect. Immun. 65, 5376–5380.

[21] Boyum, A. (1968) Isolation of mononuclear cells and granulo-

cytes from human blood. Isolation of monuclear cells by one

centrifugation, and of granulocytes by combining centrifugation

and sedimentation at 1g. Scand. J. Clin. Lab. Invest. Suppl. 97,

77–89.

[22] Klein, C.L., Bittinger, F., Kohler, H., Wagner, M., Otto, M.,

Hermanns, I. and Kirkpatrick, C.J. (1995) Comparative studies

on vascular endothelium in vitro. 3. Effects of cytokines on the

expression of E-selectin, ICAM-1 and VCAM-1 by cultured

human endothelial cells obtained from different passages. Patho-

biology 63, 83–92.

[23] Shen, Y., Sultana, C., Arditi, M., Kim, K.S. and Kalra, V.K.

(1998) Endotoxin-induced migration of monocytes and PECAM-

1 phosphorylation are abrogated by PAF receptor antagonists.

Am. J. Physiol. 275, E479–486.

[24] Broom, M.F., Sherriff, R.M., Munster, D. and Chadwick, V.S.

(1992) Identification of formyl Met-Leu-Phe in culture filtrates of

Helicobacter pylori. Microbios 72, 239–245.

[25] Craig, P.M., Territo, M.C., Karnes, W.E. and Walsh, J.H. (1992)

Helicobacter pylori secretes a chemotactic factor for monocytes

and neutrophils. Gut 33, 1020–1023.

[26] Gebert, B., Fischer, W., Weiss, E., Hoffmann, R. and Haas, R.

(2003) Helicobacter pylori vacuolating cytotoxin inhibits T lym-

phocyte activation. Science 301, 1099–1102.

[27] Shimoyama, T., Everett, S.M., Dixon, M.F., Axon, A.T. and

Crabtree, J.E. (1998) Chemokine mRNA expression in gastric

mucosa is associated with Helicobacter pylori cagA positivity and

severity of gastritis. J. Clin. Pathol. 51, 765–770.

[28] Rieder, G., Einsiedl, W., Hatz, R.A., Stolte, M., Enders, G.A. and

Walz, A. (2001) Comparison of CXC chemokines ENA-78 and

interleukin-8 expression inHelicobacter pylori-associated gastritis.

Infect. Immun. 69, 81–88.

[29] Montemurro, P., Nishioka, H., Dundon, W.G., de Bernard, M.,

Del Giudice, G., Rappuoli, R. and Montecucco, C. (2002) The

neutrophil-activating protein (HP-NAP) of Helicobacter pylori is

a potent stimulant of mast cells. Eur. J. Immunol. 32, 671–676.

[30] Blom, K., Lundin, B.S., Bolin, I. and Svennerholm, A. (2001)

Flow cytometric analysis of the localization of Helicobacter pylori

antigens during different growth phases. FEMS Immunol. Med.

Microbiol. 30, 173–179.