G

I

Hm

MFMa

b

c

d

e

f

g

T

a

ARRAA

KHFSEL

I

ap(t

m

1h

ARTICLE IN PRESS Model

JMM-50732; No. of Pages 6

International Journal of Medical Microbiology xxx (2013) xxx– xxx

Contents lists available at ScienceDirect

International Journal of Medical Microbiology

jo ur nal homepage: www.elsev ier .co m/locate / i jmm

elicobacter pylori FliD protein is a highly sensitive and specificarker for serologic diagnosis of H. pylori infection

ohammad Khalifeh Gholi a, Behnam Kalalib, Luca Formichellab, Gereon Göttnerc,ereshteh Shamsipourd, Amir hassan Zarnanie,f, Mostafa Hosseinig, Dirk H. Buschb,ohammad Hasan Shirazia,∗, Markus Gerhardb,∗∗

Department of Pathobiology, School of Public Health and Institute of Public Health Research, Tehran University of Medical Sciences, Tehran, IranInstitute for Medical Microbiology, Immunology and Hygiene, Technische Universität München, Munich, GermanyMikrogen GmbH, Neuried, GermanyDepartment of Immunochemistry, Monoclonal Antibody Research Center, Avicenna Research Institute, ACECR, Tehran, IranNanobiotechnology Research Center, Avicenna Research Institute, ACECR, Tehran, IranImmunology Research Center, Tehran University of Medical Sciences, Tehran, IranDepartment of Epidemiology and Biostatistics, School of Public Health and Institute of Public Health Research, Tehran University of Medical Sciences,ehran, Iran

r t i c l e i n f o

rticle history:eceived 8 February 2013eceived in revised form 7 August 2013ccepted 18 August 2013vailable online xxx

eywords:elicobacter pyloriliDerological diagnosisLISAine assay

a b s t r a c t

Screening for H. pylori in large populations continues to be a challenging task, since available tests havelimited sensitivity and specificity, which, in population-based approaches, leads to significant numbersof false positive and false negative results. Various H. pylori proteins associated with virulence are highlyimmunogenic and therefore candidates to detect the infection. There are currently no defined markersthat are recognized in all H. pylori infected patients and that do not show cross-reactivity with otherbacterial proteins.

We identified the H. pylori “hook-associated protein 2 homologue”, FliD (UniProtKB/Swiss-Prot:P96786.4) as a novel marker of infection for serological analysis. The H. pylori FliD protein is an essentialelement in the assembly of the functional flagella. However, this virulence factor has not yet been testedas a diagnostic marker in serology. For this purpose FliD was recombinantly expressed in E. coli, purifiedby affinity chromatography and gel filtration and used to coat ELISA plates or immobilized on nitrocellu-lose stripes. To evaluate its antigenicity we screened a defined panel of patient sera. The recombinant H.pylori FliD protein reacted with a high percentage of human sera. Among 318 samples reported positive

by histology, 310 (97.4%) were tested positive by FliD Line assay, and 165 out of 170 samples were testedpositive by ELISA (97%). We could also reconfirm 297 out of 300 (99%) negative sera by Line assay and 73from 76 (96%) by ELISA. Taken together, application of FliD in serological diagnosis of H. pylori infectionpresents a high specificity of up to 99% and a sensitivity of up to 97%. This makes especially the FliD ELISAa simple, cost effective and highly efficient tool to detect H. pylori infection in developing countries whereprevalence is high and other screening methods are either not available or are unaffordable.

ntroduction

Helicobacter pylori (H. pylori), a microaerophilic, Gram-negativend spiral bacterium is colonizing approximately half of the world

Please cite this article in press as: Khalifeh Gholi, M., et al., Helicobacter pyldiagnosis of H. pylori infection. Int. J. Med. Microbiol. (2013), http://dx.doi.

opulation and considered to be a human-specific gastric pathogenMichetti et al., 1999). Most infected individuals show asymp-omatic chronic gastritis. However, in some subjects the infection

∗ Corresponding author. Tel.: +98 2188953021; fax: +98 2188954913.∗∗ Corresponding author. Tel.: +49 89 4140 2477; fax: +49 89 4140 4139.

E-mail addresses: mhshirazi@sina.tums.ac.ir (M.H. Shirazi),arkus.gerhard@mikrobio.med.tum.de (M. Gerhard).

438-4221/$ – see front matter © 2013 Elsevier GmbH. All rights reserved.ttp://dx.doi.org/10.1016/j.ijmm.2013.08.005

© 2013 Elsevier GmbH. All rights reserved.

causes chronic active gastritis, peptic ulceration and atrophy, andplays an important role in the development of mucosa-associatedlymphoid tissue (MALT) lymphoma, gastric adenocarcinoma andprimary gastric non-Hodgkin’s lymphoma (Suganuma et al., 2001).

The World Health Organization has categorized H. pylori as aclass I carcinogen (Goto et al., 1999), and direct evidence of car-cinogenesis has been demonstrated in animal models (Honda et al.,1998; Watanabe et al., 1998). Eradication of H. pylori can preventgastric cancer in humans (Uemura et al., 2001). Test and treat strate-

ori FliD protein is a highly sensitive and specific marker for serologicorg/10.1016/j.ijmm.2013.08.005

gies have been considered in populations with high gastric cancerrisk (Yamaoka et al., 1998). However, such approach is hamperedby the lack of efficient and affordable screening systems especiallyfor countries of lower socioeconomic status. In these countries only

ING Model

I

2 nal of

sf

kdiictoTqbiimtafcp

ceBtmahT(a1s

tpeiipommboetc2

FHoni(pHFtptbsa

ARTICLEJMM-50732; No. of Pages 6

M. Khalifeh Gholi et al. / International Jour

erologic tests are applicable, most of which suffer from poor per-ormance or are not well validated.

For H. pylori serology there are several specific single markersnown and described. These factors have been applied in manyiagnostic approaches, but almost all of them have significant lim-

tations which make them unsuitable for H. pylori diagnosis. Fornstance, the cytotoxin-associated protein (CagA) is a very wellharacterized H. pylori protein. It is encoded on the cag-PAI (cyto-oxin associated gene Pathogenicity Island) and is described as anncogenic protein (Franco et al., 2005; Murata-Kamiya et al., 2007).his protein is also a highly immunogenic antigen, making it a fre-uently employed marker for serologic tests. CagA positivity cane used as an indicator of H. pylori virulence because individuals

nfected with CagA positive strains are at a higher risk for develop-ng gastroduodenal diseases. However, it is not suitable as a single

arker, since only a subgroup of H. pylori strains are CagA posi-ive. Moreover, CagA positivity is not a hallmark of active infections H. pylori eradicated patients maintain antibodies against CagAor many years (Fusconi et al., 1999). Therefore it should always beombined with other suitable antigens in serologic tests to confirmositivity.

Another well-characterized H. pylori protein is the vacuolatingytotoxin (VacA). It was reported to induce vacuolation in cellsxposed to H. pylori supernatants or purified protein (Cover andlaser, 1992). The vacA gene codes for a 140 kDa pro-toxin, wherehe amino-terminal signal sequence and the carboxy-terminal frag-

ent are proteolytically cleaved during secretion, leading to anctive protein with a molecular mass of 88 kDa that aggregates toexamers and forms a pore (Montecucco and de Bernard, 2003).his protein consists of two different regions. A signal sequences1a, s1b, s2) and a mid-region (m1, m2), both with high allelic vari-tions which appear to regulate cytotoxic activity (Atherton et al.,995). The high diversity of VacA makes this protein unsuitable forerologic testing.

Another well characterized H. pylori protein, GroEL, belongs tohe family of molecular chaperones, which are required for theroper folding of many proteins under stress conditions (Dunnt al., 1992). In different studies it was shown that this proteins highly conserved among different H. pylori strains and thatts seropositivity was even higher than for the CagA in infectedatients (Macchia et al., 1993; Suerbaum et al., 1994). Also, in ourwn studies we observed that a positive serostatus for GroEL wasore often found in German gastric cancer patients compared toatched controls (unpublished data). Also, it is suggested that anti-

odies against GroEL might persist longer after disease-related lossf H. pylori infection. Thus, GroEL may be a suitable marker ofither current or past infection, and may be helpful to overcomehe underestimation of H. pylori – related gastric cancer risk due tolearance of infection (Gao et al., 2009; Haas et al., 2002; Krah et al.,004).

In the present study, we employed the secreted H. pylori proteinliD as a novel marker of infection for serological analysis. The. pylori FliD protein is an essential element in the assemblyf the functional flagella and a FliD mutant strain is completelyon-motile. Flagellin plays a central role in bacterial motility and

s necessary for colonization and persistence of H. pylori infectionEaton et al., 1996). Motility of H. pylori is a virulent factor in theathogenesis of gastric mucosal injury (Watanabe et al., 1997). The. pylori FliD gene encodes a 76-kDa protein (Kim et al., 1999).TheliD operon of H. pylori consists of FlaG, FliD, and FliS genes, inhe order stated, under the control of a Sigma (28)-dependentromoter. In order to evaluate the applicability and suitability of

Please cite this article in press as: Khalifeh Gholi, M., et al., Helicobacter pyldiagnosis of H. pylori infection. Int. J. Med. Microbiol. (2013), http://dx.doi.

his protein in diagnostic assays, we first analyzed its sequence byioinformatics tools to assess conservation between Helicobacterpecies similarity to proteins of other bacteria. We then employed

recombinant FliD protein for the detection of specific antibodies

PRESSMedical Microbiology xxx (2013) xxx– xxx

in sera from H. pylori infected patients in a novel ELISA assay.Subsequently, the recombinant protein was applied for furtherdevelopment of a line immunoassay. The results of this study maycontribute to the development of a simple, accurate and sensitivemethod for diagnosis of the H. pylori infection.

Materials and methods

Samples

A total of 618 patients (307 men, 309 women) were enrolledin the study. After receiving an explanation of the purpose of thestudy, informed consent was obtained from each patient and ablood sample was taken at the time of endoscopy, before anytherapy was initiated. Sera were separated and stored at −20 ◦C.Diagnosis of infection was based on the histopathology as goldstandard. Patients were considered H. pylori positive when theresults of histopathology were positive. All patients were screenedby FliD Line assay, and a subset of 246 sera was tested by FliD ELISA.

Cloning of the H. pylori FliD gene

All DNA manipulations were performed under standard condi-tions as described by Sambrook et al. (1989). Briefly, the FliD genewas amplified by PCR using genomic DNA from H. pylori strain J99 asthe template. Following oligonucleotides were used as primers: 5′-CAT ATG GCA ATA GGT TCA TTA A-3′ and 5′-CTC GAG ATT CTT TTTAGC CGC TGC-3′. Using this approach a NdeI site was introducedat the 5′-end of forward primers and a XhoI site at 5′-end of thereverse primers. After PCR amplification, the product (2058 bp) wasligated into the pTZ57R/T cloning vector (InsTAcloneTM PCR CloningKit, MBI Fermentas, Lithuania). Subsequently, the insert was con-firmed via PCR and sequencing, and was cloned into a PET-28a(+)expression vector (Qiagen, USA) using NdeI and XhoI restrictionenzymes.

Expression, purification and recognition of recombinant FliD

E. coli BL21 (Qiagen, USA) competent cells were transformedwith pET-28a(+)-fliD and inoculated in LB broth with antibiotic(kanamycin, 50 �g/ml). Expression was induced by addition of1 mM Isopropyl �-d-1-thiogalactopyranoside (IPTG) at an opti-cal density (OD600) of 0.6. After 4 h cells were harvested andprotein analysis of whole lysate was carried out by Sodiumdodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE).The soluble histidin-tagged proteins were purified using affinitychromatography (HisTrap crude, GE Healthcare). As a second pol-ishing step and for buffer exchange, size exclusion chromatography(Superdex 75, GE Helthcare) was performed. The relevant fractionswhere collected and concentrated with a centrifugal filter device(Millipore) with a cut off of 10 kDa and stored at −80 ◦C. Purifiedrecombinant protein was evaluated by western blot using an anti-His Tag-HRP antibody and also a mouse anti H. pylori-HRP antibody(Pierce, Rockford, USA) and detected by ECL system (GE Healthcare,Uppsala, Sweden).

Production and purification of rFliD specific antibody

A mature white New Zealand rabbit was immunized with puri-fied protein according to the protocol of Hay et al. with lightmodifications (Hay et al., 2002). Briefly, immunization was carried

ori FliD protein is a highly sensitive and specific marker for serologicorg/10.1016/j.ijmm.2013.08.005

out by i.m. injection of 250 �g purified recombinant protein (0.5 ml)with the same volume (0.5 ml) of Freund’s complete adjuvant.For the recall immunizations, the rabbit was boosted with 125 �gpurified protein prepared in the same volume (0.5 ml) of Freund’s

IN PRESSG Model

I

nal of Medical Microbiology xxx (2013) xxx– xxx 3

itwsaa

b5plIa2

E

ccc

tmH(tupve

D

ptaoolb

evhiiwstbsCraaistsuoaF

Fig. 1. Validation of recombinant FliD by western blot. 25 �g of purified rFliD or 1 �lof H. pylori rFliD was detectable in a western blot using commercial mouse anti H.

ARTICLEJMM-50732; No. of Pages 6

M. Khalifeh Gholi et al. / International Jour

ncomplete adjuvant, 4, 6, 8 and 10 weeks later. As a negative con-rol a serum sample was taken prior to immunization. Finally, twoeeks after the last immunization, blood was collected and sera

eparated. Polyclonal IgG antibody was purified by sepharose-4Bffinity chromatography using rFliD conjugated columns preparedccording to the manufacturer’s protocol (Pharmacia, 1988).

FliD expression of H. pylori (J99) was detected by westernlot using ultrasonic supernatant at the protein concentration of0 �g/ml. The rabbit polyclonal IgG antibody raised against rFliDrotein was used as the first antibody (1:5000 dilution), HRP-

abeling sheep antibody against rabbit IgG (Avicenna Researchnstitute, Tehran, Iran) as the second antibody (1:3000 dilution)nd ECL as a substrate were used for the detection (Chen et al.,001).

LISA

ELISA plates were coated with 100 �l rFliD protein at a con-entration of 1 �g/ml in PBS and incubated overnight at 4 ◦C. Theoated wells were blocked with phosphate buffered saline (PBS)ontaining 2.5% bovine serum albumin (BSA, Sigma) for 2 h at 37 ◦C.

All H. pylori positive and negative serologic samples used inhis study were screened for antibodies against FliD by using opti-

al dilution of patients’ sera (1:100 dilution) as the first antibody,RP-conjugated anti-human IgG (Promega, Mannheim, Germany)

1:100 dilution) as the secondary antibody and TMB (3,3′,5,5′-etra methyl benzidine) as a substrate. Moreover, wells were leftncoated as a control for each serum. The result of ELISA for aatient’s serum sample was considered to be positive if its OD450alue was over the mean plus 3 SD of negative serum samples (Chent al., 2001).

evelopment of the FliD line assay

The recomLine® is a line immunoassay based on recombinant H.ylori proteins immobilized on nitrocellulose. In contrast to ELISA,he test principle allows the identification of specific antibodiesgainst various antigens of H. pylori through separate applicationf different single antigens. In analogy to the production proceduref the commercial recomLine®, rFliD was immobilized on nitrocel-ulose membrane strips together with other highly purified recom-inant H. pylori Antigens (CagA, VacA, GroEL, UreA, HcpC, gGT).

The appropriate line conditions for rFliD were determinedmpirically with a selection of standard serum samples from a pre-iously described study population comprising 20 defined H. pyloriistologically positive samples and 20 defined H. pylori histolog-

cally negative samples. The optimal antigen concentration anddeal choice of additives like detergent, dithiothreitol, and NaCl

as adjusted for each antigen by repeated cycles of lining andcreening. The conditions with best presentation of antigen epi-opes and optimal binding to the membrane, observable by perfectand appearance and best discrimination of negative and positiveamples, were selected for ideal product specifications of first lots.ontrol bands were added on the upper end of the strip comprisingabbit anti-human IgG/IgM/IgA antibodies as incubation controlsnd human IgG, IgM or IgA antibodies as conjugate control as wells a cut off control that allows the assessment of the reactivity of thendividual antigen bands. After scanning and densitometric analy-is of the band intensities, the control was used as internal referenceo calculate ratios for each band. Usually, cut off control bands arecored between 20 and 30, while strong positive bands can score

Please cite this article in press as: Khalifeh Gholi, M., et al., Helicobacter pyldiagnosis of H. pylori infection. Int. J. Med. Microbiol. (2013), http://dx.doi.

p to 600 points. Every band scoring above the individual controlf the each stripe is considered positive (ratio > 1). The productionnd processing of this Line assay has been described in details byormichella et al. (submitted).

pylori polyclonal antibody (1:200) and ECL system (a). H. pylori lysate was similarlydetected by rabbit polyclonal anti H. pylori antibody (1:5000) (b).

Results

FliD is highly conserved among H. pylori species

Bioinformatic analysis revealed that FliD amino acid sequenceis present and highly conserved in all (>100) sequenced H. pyloristrains. FliD has a homology of 97% in around 200 H. pylori strainswe analyzed (Table S1). Interestingly, except for some other non-pylori Helicobacter species with partial homology, there is no otherknown organism (bacteria as well as eukaryotics) with a significantgenomic or proteomic homology to FliD of H. pylori. This analysistogether with high antigenicity prediction of this protein indicatesthat only low cross-reactivity might be expected. Taken together,these data suggest that FliD might be a suitable candidate for appli-cation in serological diagnostic assays for detection of H. pylori.

Supplementary material related to this article can be found,in the online version, at http://dx.doi.org/10.1016/j.ijmm.2013.08.005.

Production of the recombinant FliD protein

Amplification of the FliD gene from H. pylori strain J99 DNArevealed a single PCR product of 2.05 kb (data not shown) whichwas confirmed by sequencing and ligated into the expression vec-tor pET-28a(+). After transformation into E. coli expression strainBL21 DE3 and induction with IPTG, a clear single band could beobserved on Western blot using a commercial polyclonal anti- H.pylori antiserum. The protein was purified as described in Section2 to >90% purity and again confirmed by Western blot (Fig. 1a).

Furthermore, to test the antigenicity of the recombinant FliDand to compare it to the native protein, rabbit polyclonal anti-serum was produced. Antibody titers were already determinedafter the third immunization and reached high levels after thefourth boost, confirming the good immunogenicity of FliD. The rab-bit antiserum was able to recognize the purified rFliD and FliD in H.

ori FliD protein is a highly sensitive and specific marker for serologicorg/10.1016/j.ijmm.2013.08.005

pylori lysate (Fig. 1b). Importantly, in healthy animals no side effectswere observed, indicating that FliD could also be an interesting andsafe candidate for vaccination.

ARTICLE IN PRESSG Model

IJMM-50732; No. of Pages 6

4 M. Khalifeh Gholi et al. / International Journal of Medical Microbiology xxx (2013) xxx– xxx

Table 1Analysis of patients’ sera by ELISA. 246 patients’ sera already from histologicallyconfirmed individuals were analyzed by ELISA assay. From 170 sera positively eval-uated by histopathology, 165 (97%) samples were positive in ELISA, while from 76sera which were reported negative in pathology, 73 (96%) samples were scorednegative in the FliD ELISA.

Elisa Histology Total

Positive Negative

Positive 165 (97%) 3 (4%) 168

E

stts

wr(noaawf(n

O

ilcTttdsdrbiap

Ra

rbaptewsFhr

Table 2Analysis of patients’ sera by line assay. 618 patients’ already diagnosed byhistopathology, recomLine® and recomWell® were analyzed by FliD line assay. From318 sera evaluated positive in histopathology, 310 (97.4%) samples were positive inLine assay, and 297 (99%) out of 300 sera evaluated negative in histopathology werenegative by FliD line assay.

Line assay Histology Total

Positive Negative

Positive 310 (97.4%) 3 (1%) 313

Negative 5 (3%) 73 (96%) 78Total 170 (100%) 76 (100%) 246

LISA

A sandwich ELISA was generated to screen patient’s sera. 246era from patients which were previously assessed H. pylori posi-ive or negative in histopathological analysis of biopsies were usedo evaluate the FliD ELISA. During histopathological analysis 170amples reported positive whereas 76 samples were negative.

Using the FliD ELISA, among 170 positive reported samples,e detected 165 positive samples, whereas among 76 samples

eported negative we could reconfirm 73 as negative by ELISATable 1). Taken together, application of FliD in ELISA based diag-osis of H. pylori infection has a specificity of 96% and a sensitivityf 97%. Interestingly, the five cases which were ELISA negative hadlso low but barely positive scores in the line blot which were justbove the cut off (ratios ranging from 1.2 to 2.2). One of theseas also regarded H. pylori negative by line blot, while the other

our were line blot positive, reacting with several other antigensdata not shown). It is important to note that only one sample wasegative by both tests.

ptimization of line blot assay

The promising data obtained with ELISA prompted us to furthernvestigate the usefulness of rFliD in another assays system, theine blot. Here, single recombinant proteins are applied to a nitro-ellulose strip, which then enables detection of individual bands.he performance of this assay highly depends on the concentra-ion of antigen applied as well as correct buffer conditions. Initially,o optimize the assay conditions for FliD, antigen was applied atifferent concentrations and defined sera were screened (data nothown). Then, to exclude cross-reactivity in the line blot assay, vali-ation testing was performed using different buffers in repetitiveounds on a panel of blood donors with H. pylori status verifiedy histology (npos = 20, nneg = 15). The rFliD antigen showed a clear

mmune response with high sensitivity and low cross reactivity. Itlso showed a good discrimination between negative and positiveatients (Fig. 2).

ecognition of antibodies against FliD in patients sera by linessay

Using the optimized Line assay for FliD, we analyzed antibodyesponses against FliD in the sera of 618 patients (part of which hadeen screened by ELISA). Here, we observed that the line assay has

high sensitivity of 97.4% with 310 out of 318 patients evaluatedositive in histopathology being positive by line assay, whereashe line assay reaches a specificity of 99% (Table 2). We carefullyxamined the results from the patients in which discrepant resultsere obtained. 8 sera were negative for FliD on the line assay but

Please cite this article in press as: Khalifeh Gholi, M., et al., Helicobacter pyldiagnosis of H. pylori infection. Int. J. Med. Microbiol. (2013), http://dx.doi.

howed reactivity with other antigens, indicating that here indeedliD was not recognized as antigen. Within these 8 samples, onead no reactivity against the FliD band at all. Seven had a weakeactivity which was barely below the cut off (ratios between 0.6

Negative 8 (2.6%) 297 (99%) 305Total 318 (100%) 300 (100%) 618

and 0.95), and four of these had weak reactivities against all otherrecognized antigens in general (not shown). All three samples inwhich FliD gave a “false positive” result showed reactivities withother bands as well. All these bands including FliD were relativelyweak, but clearly above cut off.

Analysis of the cross reactivity

To exclude the possible cross reaction of FliD from H. pyloriwith sera from patients infected with other flagellated pathogens,we firstly performed bioinformatics analysis. The highest overallhomology of the H. pylori’s FliD to E. coli, Salmonella and Campy-lobacter species’ proteins is: 9% to Flagellar protein of E. coli(accession number YP 006137341), 24% to flagellar hook associatedprotein 2 of Salmonella (accession number YP 004730643) and 25%to FliD of Campylobacter (accession number ZP 18210110.1). Thereis no homologous sequence of more than 8 amino acids among thecompared sequences. We further tested the cross-reactivity of H.pylori’s FliD with sera from patients infected with Campylobacterjejuni (C. jejuni) as the closest related flagellated bacterium to H.pylori. Briefly, thirty C. jejuni positive but H. pylori negative pre-screened sera were tested by line-assay based analysis to assesspossible cross reactive antibodies detecting FliD of H. pylori. Asshown in Fig. 3, we did not observed any cross reactivity betweenFliD of H. pylori and sera from C. jejuni infected individuals.

Discussion

In this study we describe FliD as a highly sensitive and specificantigen for application in serological diagnosis of H. pylori. Serologyis the noninvasive technique of choice to detect H. pylori infectionbecause it is simple, widely available and inexpensive. The reportedsensitivity and specificity of IgG serology is highly variable, ran-ging from 30% to 90%. (Urita et al., 2004). Although there are manyserological tests available for diagnosis of H. pylori, most of themhave significant limitations. Often, antigens used in these tests areeither not expressed in all H. pylori isolates (e.g. CagA) or they havea high diversity that makes them unsuitable for application in aserological diagnostic test (e.g. VacA). While combination of severalantigens together improves the reliability, sensitivity and speci-ficity of a given test system, many of these antigens could not beproduced recombinantly in a native and active form. Also, combin-ing several recombinant antigens leads to higher production costsand thus test kit prices, which makes such tests unaffordable inmany developing countries. Obviously, there is a need for a simplereliable test which is not only inexpensive but also technically fea-sible and applicable in most areas. Here we demonstrate that FliDis not only easily producible in recombinant form, also it has allproperties required to consider an antigen suitable for application

ori FliD protein is a highly sensitive and specific marker for serologicorg/10.1016/j.ijmm.2013.08.005

in diagnostic approaches.Initially, we were interested to establish an expression system

which would allow the reproducible production of recombinantFliD protein under standard conditions in E. coli. Such approach is

ARTICLE IN PRESSG Model

IJMM-50732; No. of Pages 6

M. Khalifeh Gholi et al. / International Journal of Medical Microbiology xxx (2013) xxx– xxx 5

F and aP ty of gs

oHopdaecccop

aprsr

a

Fr

ig. 2. (a) FliD optimization on Line assay concerning antigen concentration, bufferatient ID B02-B41: H. pylori negative patients verified by histology, Y-axis: intensiera were selected randomly to illustrate the read out of the test.

ften hampered by the fact that outer membrane proteins from. pylori can not efficiently be expressed in E. coli based systems,r, after successful expression, are not soluble and thus difficult tourify. In this case, proteins must be purified using denaturing con-itions and subsequently refolded, which may affect native foldingnd antigenicity. Here, we were able to achieve stable and robustxpression of FliD using the pET-28a(+) expression system, and weould produce the recombinant FliD protein efficiently in BL21DE3ells. rFliD was mainly present as a soluble protein at a high con-entration upon induction by IPTG. The high yield and solubilityf rFliD will be beneficial in large scale production of recombinantrotein for the assembly of a diagnostic test.

Recognition of the FliD by a commercially available antibodygainst whole cell lysates of H. pylori indicated that the purifiedrotein presents antigenic epitopes. Moreover, we could show inabbit immunization experiments that rFliD efficiently induces apecific antibody response with a high titer, indicating that the

Please cite this article in press as: Khalifeh Gholi, M., et al., Helicobacter pyldiagnosis of H. pylori infection. Int. J. Med. Microbiol. (2013), http://dx.doi.

ecombinant protein exhibits favorable antigenicity.In a diagnostic ELISA and line assay, we observed that serum

ntibodies against FliD were present in approximately 97% of H.

ig. 3. Analysis of cross-reactivity to sera from C. jejuni infected individuals. Sera from theaction was observed.

dditives. X-axis: Patient ID 104-273: H. pylori positive patients verified by histologyray. (b) Line assay for diagnosis of the anti-FliD antibody in patients’ sera. Here 20

pylori infected patients. In only four patients regarded positive byhistology and screened with both tests, neither the line blot northe ELISA detected antibodies against FliD. Thus, FliD detectionrates were significantly higher than those of vacuolating cytotoxinA (VacA) (42.4%, 68%), heat shock protein (68%) and FlaB (92.8%)and similar to FlaA (98.4%) (Opazo et al., 1999; Yan et al., 2003; Yanand Mao, 2004). After the Line assay had been optimized, we couldreach a sensitivity of 97.4% which is close to the sensitivity of 97% inthe simplified ELISA while specificity of the line assay was superior.In this setup, the specificity of the ELISA was up to 96%, which issufficient for a diagnostic test to be used in screening approaches.

Interestingly, most sera which were false negative or false pos-itive by either test gave results close to the respective cut off.Sensitivity and specificity of assays was calculated by dividing thenumber of detected positive samples by all positive reported sam-ples considering various different cut offs in both tests. For example,the specificity of the line blot could be improved by slightly raising

ori FliD protein is a highly sensitive and specific marker for serologicorg/10.1016/j.ijmm.2013.08.005

the cut off from a ratio of 1.0 to a ratio of 1.2. While this would elim-inate two false positive samples, five true positives would then fallbelow the cut off. Similarly, increasing the sensitivity by lowering

irty C. jejuni infected patients were analyzed by FliD line-assay. No detectable cross

ING Model

I

6 nal of

tiTbrpFesomitmvt2asiis

piatrH

A

m

R

A

A

C

C

D

E

F

F

ARTICLEJMM-50732; No. of Pages 6

M. Khalifeh Gholi et al. / International Jour

he cut off to 0,9 would render 6 more samples positive, but alsoncrease the false positives by five, leading to a loss of specificity.herefore, we think that the current system is optimally titrated forest sensitivity and specificity. The fact that some patients showather weak reactivates against FliD (although the protein is sup-osed to be present in all strains) can have several explanations.irst, upon initial infection, presentation of epitopes is a stochasticvent, and in some cases the FliD protein might be underrepre-ented, leading to weak B cell activation. However, since in mostf these cases also all other antigens have given weak signals, it isore likely that the H. pylori infection in these patients in general

nduces a weak immune response. In this context, we recognizedhat all these patients with low humoral responses only showed

ild gastritis. Several recent publications have substantiated theiew that H. pylori has the potential to induce tolerance ratherhan only immune activation (Arnold et al., 2012; Oertli et al.,012). It is tempting to speculate that in the patients presentings false negative by serology, such tolerogenic effects prevent atronger immune response against the infection. Alternatively, its conceivable that some patients are receiving immunosuppress-ve medication which was not reported during enrollment in thistudy.

Taken together, the data presented here indicate that the H.ylori FliD efficiently induces specific antibodies in almost allnfected individuals, which can easily be detected by serologicssays based on recombinant FliD, implying a great potential forhe establishment of a simple H. pylori diagnostic assay. Thus, theecombinant FliD protein presents an ideal candidate to screen for. pylori infection in large populations with high prevalence.

cknowledgment

The authors thank Sundus Javed for critical reading theanuscript.

eferences

rnold, I.C., Hitzler, I., Muller, A., 2012. The immunomodulatory properties of Heli-cobacter pylori confer protection against allergic and chronic inflammatorydisorders. Front. Cell. Infect. Microbiol. 2, 10.

therton, J.C., Cao, P., Peek Jr., R.M., Tummuru, M.K., Blaser, M.J., Cover, T.L., 1995.Mosaicism in vacuolating cytotoxin alleles of Helicobacter pylori, associationof specific vacA types with cytotoxin production and peptic ulceration. J. Biol.Chem. 270 (30), 17771–17777.

hen, Y., Wang, J., Shi, L., 2001. In vitro study of the biological activities and immuno-genicity of recombinant adhesin of Heliobacter pylori rHpaA. Zhonghua Yi XueZa Zhi 81 (5), 276–279.

over, T.L., Blaser, M.J., 1992. Purification and characterization of thevacuolating toxin from Helicobacter pylori. J. Biol. Chem. 267 (15),10570–10575.

unn, B.E., Roop 2nd, R.M., Sung, C.C., Sharma, S.A., Perez-Perez, G.I., Blaser, M.J.,1992. Identification and purification of a cpn60 heat shock protein homologfrom Helicobacter pylori. Infect. Immun. 60 (5), 1946–1951.

aton, K.A., Suerbaum, S., Josenhans, C., Krakowka, S., 1996. Colonization of gnotobi-otic piglets by Helicobacter pylori deficient in two flagellin genes. Infect. Immun.64 (7), 2445–2448.

ranco, A.T., Israel, D.A., Washington, M.K., Krishna, U., Fox, J.G., Rogers, A.B., et al.,

Please cite this article in press as: Khalifeh Gholi, M., et al., Helicobacter pyldiagnosis of H. pylori infection. Int. J. Med. Microbiol. (2013), http://dx.doi.

2005. Activation of beta-catenin by carcinogenic Helicobacter pylori. Proc. Natl.Acad. Sci. U.S.A. 102 (30), 10646–10651.

usconi, M., Vaira, D., Menegatti, M., Farinelli, S., Figura, N., Holton, J., et al., 1999.Anti-CagA reactivity in Helicobacter pylori-negative subjects: a comparison ofthree different methods. Dig. Dis. Sci. 44 (8), 1691–1695.

PRESSMedical Microbiology xxx (2013) xxx– xxx

Gao, L., Michel, A., Weck, M.N., Arndt, V., Pawlita, M., Brenner, H., 2009. Helicobac-ter pylori infection and gastric cancer risk: evaluation of 15 H. pylori proteinsdetermined by novel multiplex serology. Cancer Res. 69 (15), 6164–6170.

Goto, T., Nishizono, A., Fujioka, T., Ikewaki, J., Mifune, K., Nasu, M., 1999. Localsecretory immunoglobulin A and postimmunization gastritis correlate with pro-tection against Helicobacter pylori infection after oral vaccination of mice. Infect.Immun. 67 (5), 2531–2539.

Haas, G., Karaali, G., Ebermayer, K., Metzger, W.G., Lamer, S., Zimny-Arndt, U., et al.,2002. Immunoproteomics of Helicobacter pylori infection and relation to gastricdisease. Proteomics 2 (3), 313–324.

Hay, F.C., Westwood, O.M.R., Nelson, P.N., Hudson, L., 2002. Practical Immunology.Wiley-Blackwell, Chichester, West Sussex, UK.

Honda, S., Fujioka, T., Tokieda, M., Satoh, R., Nishizono, A., Nasu, M., 1998. Devel-opment of Helicobacter pylori-induced gastric carcinoma in Mongolian gerbils.Cancer Res. 58 (19), 4255–4259.

Kim, J.S., Chang, J.H., Chung, S.I., Yum, J.S., 1999. Molecular cloning and characteriza-tion of the Helicobacter pylori fliD gene, an essential factor in flagellar structureand motility. J. Bacteriol. 181 (22), 6969–6976.

Krah, A., Miehlke, S., Pleissner, K.P., Zimny-Arndt, U., Kirsch, C., Lehn, N., et al., 2004.Identification of candidate antigens for serologic detection of Helicobacter pylori-infected patients with gastric carcinoma. Int. J. Cancer 108 (3), 456–463.

Macchia, G., Massone, A., Burroni, D., Covacci, A., Censini, S., Rappuoli, R., 1993. TheHsp60 protein of Helicobacter pylori: structure and immune response in patientswith gastroduodenal diseases. Mol. Microbiol. 9 (3), 645–652.

Michetti, P., Kreiss, C., Kotloff, K.L., Porta, N., Blanco, J.L., Bachmann, D., et al., 1999.Oral immunization with urease and Escherichia coli heat-labile enterotoxin issafe and immunogenic in Helicobacter pylori-infected adults. Gastroenterology116 (4), 804–812.

Montecucco, C., de Bernard, M., 2003. Molecular and cellular mechanisms of actionof the vacuolating cytotoxin (VacA) and neutrophil-activating protein (HP-NAP)virulence factors of Helicobacter pylori. Microbes Infect. 5 (8), 715–721.

Murata-Kamiya, N., Kurashima, Y., Teishikata, Y., Yamahashi, Y., Saito, Y., Higashi, H.,et al., 2007. Helicobacter pylori CagA interacts with E-cadherin and deregulatesthe beta-catenin signal that promotes intestinal transdifferentiation in gastricepithelial cells. Oncogene 26 (32), 4617–4626.

Oertli, M., Sundquist, M., Hitzler, I., Engler, D.B., Arnold, I.C., Reuter, S., et al., 2012.DC-derived IL-18 drives Treg differentiation, murine Helicobacter pylori-specificimmune tolerance, and asthma protection. J. Clin. Invest. 122 (3), 1082–1096.

Opazo, P., Muller, I., Rollan, A., Valenzuela, P., Yudelevich, A., Garcia-de la Guarda, R.,et al., 1999. Serological response to Helicobacter pylori recombinant antigens inChilean infected patients with duodenal ulcer, non-ulcer dyspepsia and gastriccancer. APMIS 107 (12), 1069–1078.

Pharmacia., 1988. Affinity Chromatography LKB Biotechnology Uppsala, Sweden.Sambrook, J., Fritsch, E., Maniatis, T., 1989. Moleculer Cloning: A Laboratory Manual,

Book 1. Cold Spring Harbor Laboratory Press, New York.Suerbaum, S., Thiberge, J.M., Kansau, I., Ferrero, R.L., Labigne, A., 1994. Helicobac-

ter pylori hspA-hspB heat-shock gene cluster: nucleotide sequence, expression,putative function and immunogenicity. Mol. Microbiol. 14 (5), 959–974.

Suganuma, M., Kurusu, M., Okabe, S., Sueoka, N., Yoshida, M., Wakatsuki, Y., et al.,2001. Helicobacter pylori membrane protein 1: a new carcinogenic factor ofHelicobacter pylori. Cancer Res. 61 (17), 6356–6359.

Uemura, N., Okamoto, S., Yamamoto, S., Matsumura, N., Yamaguchi, S., Yamakido, M.,et al., 2001. Helicobacter pylori infection and the development of gastric cancer.N. Engl. J. Med. 345 (11), 784–789.

Urita, Y., Hike, K., Torii, N., Kikuchi, Y., Kurakata, H., Kanda, E., et al., 2004. Comparisonof serum IgA and IgG antibodies for detecting Helicobacter pylori infection. Intern.Med. 43 (7), 548–552.

Watanabe, S., Takagi, A., Tada, U., Kabir, A.M., Koga, Y., Kamiya, S., et al., 1997. Cyto-toxicity and motility of Helicobacter pylori. J. Clin. Gastroenterol. 25 (Suppl. 1),S169–S171.

Watanabe, T., Tada, M., Nagai, H., Sasaki, S., Nakao, M., 1998. Helicobacter pyloriinfection induces gastric cancer in mongolian gerbils. Gastroenterology 115 (3),642–648.

Yamaoka, Y., Kodama, T., Kita, M., Imanishi, J., Kashima, K., Graham, D.Y., 1998.Relationship of vacA genotypes of Helicobacter pylori to cagA status, cytotoxinproduction, and clinical outcome. Helicobacter 3 (4), 241–253.

Yan, J., Liang, S.H., Mao, Y.F., Li, L.W., Li, S.P., 2003. Construction of expression systemsfor flaA and flaB genes of Helicobacter pylori and determination of immunoreac-

ori FliD protein is a highly sensitive and specific marker for serologicorg/10.1016/j.ijmm.2013.08.005

tivity and antigenicity of recombinant proteins. World J. Gastroenterol. 9 (10),2240–2250.

Yan, J., Mao, Y.F., 2004. Construction of a prokaryotic expression system of vacAgene and detection of vacA gene, VacA protein in Helicobacter pylori isolates andant-VacA antibody in patients’ sera. World J. Gastroenterol. 10 (7), 985–990.