human and animal isolates of yersinia enterocolitica show ...human and animal isolates of yersinia...

Human and Animal Isolates of Yersinia enterocolitica Show SignificantSerotype-Specific Colonization and Host-Specific Immune DefenseProperties

Julia Schaake,a Malte Kronshage,a Frank Uliczka,a Manfred Rohde,b Tobias Knuuti,a Eckhard Strauch,c Angelika Fruth,d

Melissa Wos-Oxley,e Petra Derscha

Department of Molecular Infection Biology, Helmholtz Centre for Infection Research, Braunschweig, Germanya; Department of Medical Microbiology, Helmholtz Centrefor Infection Research, Braunschweig, Germanyb; Federal Institute for Risk Assessment, Berlin, Germanyc; Robert Koch Institute, Wernigerode, Germanyd; Department ofMicrobial Interactions and Processes, Helmholtz Centre for Infection Research, Braunschweig, Germanye

Yersinia enterocolitica is a human pathogen that is ubiquitous in livestock, especially pigs. The bacteria are able to colonize the intesti-nal tract of a variety of mammalian hosts, but the severity of induced gut-associated diseases (yersiniosis) differs significantly betweenhosts. To gain more information about the individual virulence determinants that contribute to colonization and induction of im-mune responses in different hosts, we analyzed and compared the interactions of different human- and animal-derived isolates of sero-types O:3, O:5,27, O:8, and O:9 with murine, porcine, and human intestinal cells and macrophages. The examined strains exhibitedsignificant serotype-specific cell binding and entry characteristics, but adhesion and uptake into different host cells were not host spe-cific and were independent of the source of the isolate. In contrast, survival and replication within macrophages and the induced proin-flammatory response differed between murine, porcine, and human macrophages, suggesting a host-specific immune response. In fact,similar levels of the proinflammatory cytokine macrophage inflammatory protein 2 (MIP-2) were secreted by murine bone marrow-derived macrophages with all tested isolates, but the equivalent interleukin-8 (IL-8) response of porcine bone marrow-derived macro-phages was strongly serotype specific and considerably lower in O:3 than in O:8 strains. In addition, all tested Y. enterocolitica strainscaused a considerably higher level of secretion of the anti-inflammatory cytokine IL-10 by porcine than by murine macrophages. Thiscould contribute to limiting the severity of the infection (in particular of serotype O:3 strains) in pigs, which are the primary reservoirof Y. enterocolitica strains pathogenic to humans.

The bacterium Yersinia enterocolitica is a food-borne zoonoticpathogen that causes various gut-associated diseases (yersini-osis) in humans, including acute enteritis, enterocolitis, diarrhea,and mesenteric lymphadenitis. The disease is typically self-limit-ing, although sequelae such as reactive arthritis, erythema nodo-sum, and thyroiditis are also common (1).

Based on their biochemical properties, pathogenic strains of Y.enterocolitica are classified into different biotypes and are furthergrouped into different serotypes (O:3, O:5,27, O:8, and O:9). Biotype1B strains (e.g., 1B/O:8) are highly virulent for mice, whereas biotype2 to 5 strains have lower pathogenicity in mouse models (1). Al-though less virulent in mice, biotype 2 to 5 strains account for themajority of human yersiniosis cases. In particular, bioserotype 4/O:3strains are by far the most frequent causes of outbreaks in Europe andJapan (80 to 90%). This bioserotype is less common in North Amer-ica, but lately, it has replaced Y. enterocolitica 1B/O:8 as the predom-inant serotype in these geographic regions (2, 3).

Y. enterocolitica was frequently isolated from wild animals (e.g.,boars) and livestock (e.g., sheep, cattle, goats, and poultry), butthe most important reservoir for human infections is pigs, fromwhich pathogenic Y. enterocolitica (in particular serotype O:3 andO:9 strains) can be routinely isolated (1, 4, 5). Pigs are generallysymptomless carriers, but they show clear seroconversion (e.g.,develop antibodies against the Yop proteins and lipopolysaccha-ride [LPS]), which has been used to detect swine carriers of thisenteroinvasive pathogen (6, 7). Pathogenic Y. enterocolitica car-riage rates in swine range from 35% to 70% of herds and 4.5% to100% of individual swine (8). The bacteria colonize their orophar-

ynx, nasopharynx, and intestinal tract for long periods withoutinduction of pathological changes (9, 10).

The primary route of human infection was shown to be foodborne. Transmission of Y. enterocolitica by the fecal-oral routethrough contaminated water or food has been frequently re-ported. The bacteria are most commonly found in raw or under-cooked pork or pork products (e.g., chitterlings), but also, out-breaks associated with other dairy and meat products andnosocomial transmissions have been documented (1, 4, 11).

The route of infection and virulence factors important forpathogenesis of Y. enterocolitica, in particular of the intensivelystudied bioserotype 1B/O:8 strains, are well known. After oralingestion, the bacteria reach the terminal ileum, where they crossthe intestinal epithelial layer through antigen-sampling M cells.Subsequently, they proliferate in the underlying lymphoid tissueand disseminate to the mesenteric lymph nodes, liver, and spleen(12). To invade and survive within different host tissues, Y. entero-

Received 9 May 2013 Returned for modification 30 May 2013Accepted 29 July 2013

Published ahead of print 19 August 2013

Editor: J. B. Bliska

Address correspondence to Petra Dersch, [email protected].

Supplemental material for this article may be found at http://dx.doi.org/10.1128/IAI.00572-13.

Copyright © 2013, American Society for Microbiology. All Rights Reserved.

doi:10.1128/IAI.00572-13

November 2013 Volume 81 Number 11 Infection and Immunity p. 4013–4025 iai.asm.org 4013

on March 31, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

http://dx.doi.org/10.1128/IAI.00572-13http://dx.doi.org/10.1128/IAI.00572-13http://dx.doi.org/10.1128/IAI.00572-13http://iai.asm.orghttp://iai.asm.org/

colitica utilizes a large set of virulence factors encoded on the chro-mosome and on the virulence plasmid (pYV) (13, 14). Host cellinteraction and invasion are promoted mainly by the adhesinsinvasin (InvA) and YadA. Invasin binds to �1-integrin receptorsand induces the most efficient uptake of the bacteria into mam-malian cells, which is important for rapid transcytosis of the in-testinal epithelial layer (15, 16). Expression of the invasin proteinis temperature and growth phase dependent, and this control pro-cess is mediated by the thermosensing transcriptional factor RovA(17, 18). YadA is a multifunctional virulence factor that promotesbinding to extracellular matrix proteins such as collagen, fi-bronectin, and laminin and protects the bacteria from comple-ment-mediated lysis and phagocytosis (19, 20). The yadA gene isencoded on the pYV plasmid and is coexpressed with the plasmid-encoded Ysc type III secretion system (T3SS) and the antiphago-cytic effector proteins (Yops) in a temperature-dependent man-ner (13). Enteropathogenic Yersinia strains have also been shownto modulate host immune responses. Invasin- or YadA-mediatedcontact with epithelial cells triggers activation of the nuclear factorNF-�B and induces production and secretion of the proinflam-matory chemokine interleukin-8 (IL-8) (21, 22). As IL-8 is a po-tent chemoattractant for neutrophils, which transiently separatecell-cell contacts to infiltrate infected tissues, it was assumed thatIL-8 secretion facilitates paracytosis and increases disseminationof the bacteria into deeper tissues. Furthermore, several Yop effec-tor proteins, which are injected into professional phagocytes, wereshown to manipulate the release of several proinflammatory cyto-kines such as tumor necrosis factor alpha (TNF-�), IL-6, IL-8, andIL-12 and the anti-inflammatory cytokine IL-10 (23). The releaseof these cytokines was shown to be an important part of the im-mune response (e.g., activation of natural killer cells and T cells)against Y. enterocolitica infection, and perturbation of their releasewas shown to be crucial to overcome innate and adaptive immuneresponses (24). A second and well-studied attribute of the Yopproteins is the ability to induce host cell apoptosis and preventuptake and lysis of Yersinia by professional phagocytes (23).

Y. enterocolitica generally persists and multiplies extracellu-larly in aggregates within lesions and abscesses when estab-lished in the lymphatic tissues. During this stage of infection,they are able to resist phagocytosis by macrophages and neu-trophils through T3SS and Yop protein functions. However, ithas been shown that pathogenic Yersinia species can also sur-vive and replicate within macrophages (25). It has been hy-pothesized that macrophages provide a replicative niche, inparticular during the very early stages of infection, before theT3SS and Yops are expressed. This would (i) allow the bacteriato replicate while they are protected from neutrophils recruitedto the site of infection, (ii) delay adaptive immune responses byhindrance of antigen representation, and (iii) permit the use ofmacrophages as transport vehicles for dissemination from gut-associated tissues to deeper organs (25, 26).

Studies analyzing Y. enterocolitica virulence factors and theirrole in pathogenesis were performed mainly by using the highlymouse-virulent 1B/O:8 strains. However, most Y. enterocoliticapatient isolates belong to different bioserotypes (e.g., 4/O:3 and3/O:9), which are also frequently isolated from food and animals,but they are less virulent to mice. These Y. enterocolitica isolatescan infect a wide range of hosts with various severities of disease,but the connection between human and animal isolates and themolecular mechanisms that determine the different outcomes of

infections are not well understood. To gain more informationabout the individual virulence determinants that contribute tohost specificity, host adaptation, and disease development, we in-vestigated the abilities of different subtypes to infect and prolifer-ate in different murine, porcine, and human epithelial cells andmacrophages. We further analyzed host-specific immune re-sponses to different strains and found that human and animalisolates of Y. enterocolitica show significant serotype-specific col-onization and host-specific immune defense properties.

MATERIALS AND METHODSBacterial strains, cell culture, media, and growth conditions. The bac-terial strains used in this study are listed in Table 1. Y. enterocolitica strainsisolated from humans and animals, in particular pigs, were selected fromlibraries held at the Governmental Institute of Risk Assessment and theRobert Koch Institute in Germany. Escherichia coli strains were grown at37°C in Luria-Bertani broth (LB), and Y. enterocolitica was grown at 25°Cor 37°C in LB under aeration on a shaker (Multitron; Infors, Bottmingen,Switzerland) at 200 rpm.

All cell lines were grown at 37°C in the presence of 5% CO2. HumanHEp-2 cells were cultured in RPMI 1640 medium plus GlutaMax (Invit-

TABLE 1 Strains used in this studya

Strain DescriptionReference orsource

E. coli CC118 �pir F� �(ara-leu)7697 �(lacZ)74�(phoA)20 araD139 galEgalK thi rpsE rpoB arfEAm

recA1

65

Yersinia enterocoliticaY11 O:3/4, human, Germany, 1994 A. RakinYeO3 O:3/4, human, Finland, 1976 M. SkurnikY1/07 O:3/4, human, Germany, 2007 E. StrauchY2/07 O:3/4, human, Germany, 2007 E. StrauchY4/07 O:3/4, human, Germany, 2007 E. StrauchY5/07 O:3/4, human, Germany, 2007 E. StrauchY8/07 O:3/4, human, Germany, 2007 E. Strauch3446 O:3/4, human, Germany A. Fruth5025 O:3/4, human, Germany A. FruthY32/07 O:3/4, porcine, Germany, 2007 E. StrauchY33/07 O:3/4, porcine, Germany, 2007 E. StrauchY34/07 O:3/4, porcine, Germany, 2007 E. Strauch3056 O:5,27/3, human, Germany A. Fruth3192 O:5,27/3, human, Germany A. FruthY14 O:5,27/3, porcine, Germany E. StrauchY17 O:5,27/2, porcine, Germany E. StrauchY49 O:8/1B, human, Germany, 1991 E. Strauch8081v O:8, human, USA 35651 O:9/3, human, Germany A. Fruth8495 O:9/3, human, Germany A. FruthY23 O:9/2, cattle, Germany E. StrauchY24 O:9/3, human, Germany E. StrauchY26 O:9/3, porcine, Germany E. StrauchY48 O:9/3, porcine, Germany E. Strauch7836 O:9/3, human, Germany A. Fruth628 O:9/3, human, Germany A. FruthYE22 8081 �pYV This workYE23 Y1 �pYV This work

a For Y. enterocolitica strain descriptions, serovar/biotype, origin, and year ofisolation are indicated. Plasmid-cured derivatives of Y1 and 8081v were obtained byselection of fast-growing bacteria on Na-oxalate-containing solid medium aftergrowth at 37°C for 24 h.

Schaake et al.

4014 iai.asm.org Infection and Immunity


http://iai.asm.org/

Dow

nloaded from

http://iai.asm.orghttp://iai.asm.org/

rogen) supplemented with 7.5% newborn calf serum (Sigma-Aldrich).Other epithelial cell lines (LoVo, IPEC-J2, IPI-2I, and Mode-K) were cul-tured in Dulbecco’s modified Eagle’s medium (DMEM)–Ham’s F-12 me-dium (1:1; Biochrom) supplemented with 10% fetal calf serum (FCS)(Biochrom). Human U937 cells were grown in RPMI 1640 medium plusGlutaMax (Gibco) supplemented with 10% fetal calf serum (Biochrom).Human THP-1 macrophages were cultured in Iscove’s basal medium(Biochrom) supplemented with 10% fetal calf serum (Biochrom). BothU937 and THP-1 cells were induced by the addition of 40 nM PMA (phor-bol myristate acetate; Sigma-Aldrich) and incubated for 48 h (THP-1cells) or 72 h (U937 cells) prior to infection. Porcine macrophages(PLN-C2 and 3D4/31) were cultured in Iscove’s basal medium (Bio-chrom) supplemented with 10% porcine serum (Promocell). J774A.1murine macrophages were grown in RPMI 1640 medium plus GlutaMaxI(Gibco) supplemented with 10% fetal calf serum (Biochrom), andRaw264.7 murine macrophages were cultured in Iscove’s basal medium(Biochrom) supplemented with 10% fetal calf serum.

Gel electrophoresis, preparation of cell extracts, and Western blot-ting. Bacteria were grown overnight under aeration in LB at 25°C or 37°C.The optical density at 600 nm (OD600) of the cultures was adjusted, and analiquot was withdrawn from each culture. To visualize the YadA and InvAproteins, whole-cell extracts were prepared and separated on 10% SDS-polyacrylamide gels as described previously (27). For detection of theRovA protein, whole-cell extracts were separated on 15% polyacrylamidegels. Subsequently, proteins were transferred onto a polyvinylidene diflu-oride (PVDF) Immobilon transfer membrane (Millipore) and probedwith polyclonal antibodies directed against the YadA, InvA, or RovA pro-tein. The antigen-antibody complexes were visualized with a secondaryrabbit alkaline phosphatase antibody (Roth) with 5-bromo-4-chloro-3-indolylphosphate and nitroblue tetrazolium (Roth) as the substrates.

Cell adhesion and invasion assays. To test cell adhesion and uptake ofthe different Y. enterocolitica isolates into epithelial cells, 5 � 104 cells(HEp-2, Caco-2, LoVo, IPEC-J2, IPI-2I, or Mode-K) were seeded andgrown overnight in individual wells of 24-well cell culture plates. Forquantification of bacterial cell adhesion to and invasion of macrophages,2 � 105 cells (THP-1, U937, PLN-C2, 3D4/31, Raw264.7, or J774) wereseeded and grown overnight in individual wells of 24-well cell cultureplates. To prevent invasion by macrophages, the cells were treated withbinding buffer containing 5 �g/ml cytochalasin D (Sigma-Aldrich) for 60min at 37°C. Subsequently, cells were washed three times with phosphate-buffered saline (PBS) and incubated in binding buffer (culture mediumsupplemented with 20 mM HEPES [pH 7.0] and 0.4% bovine serumalbumin [BSA]) before infection. Bacteria were added to cells at a multi-plicity of infection (MOI) of 10 and centrifuged onto the monolayer. At 30min postinfection, the cells were washed extensively with PBS. The totalnumber of host cell-associated bacteria was determined by cell lysis using0.1% Triton X-100 and plating onto bacterial media (27, 28). Bacterialuptake was assessed 30 min after infection as the percentage of bacteriathat survived killing by gentamicin, as described previously (29). For eachstrain, the relative levels of bacterial adhesion and uptake were deter-mined by calculating the number of CFU relative to the total number ofbacteria introduced onto cells. The number of invaded bacteria is givenrelative to the number of cell-bound bacteria. The experiments were rou-tinely performed in triplicate.

Cell viability tests of macrophages. To examine the cell viability ofinfected macrophages, cells were infected with bacteria grown at 25°C or37°C at an MOI of 1, 10, or 50. At 2, 24, and 48 h postinfection, the cellviability of the macrophages was analyzed by using the Live/Dead Viabil-ity/Cytotoxicity kit for mammalian cells (Invitrogen). Intracellular es-terase activity generates a green fluorescent compound of a nonfluores-cent substrate of the kit. Vital green fluorescent cells (detectable at anexcitation wavelength of 494 nm and an emission wavelength of 517 nm)and ethidium bromide-stained dead red fluorescent cells (excitationwavelength of 528 nm and emission wavelength of 617 nm) were visual-

ized with a fluorescence microscope (Axiovert II with Axiocam HR; Zeiss,Germany) by using the AxioVision program (Zeiss, Germany).

Assays of bacterial survival in macrophages. Bacterial survival withinmacrophages was determined by a gentamicin protection assay, as de-scribed above. Cells were infected at an MOI of 10. One hour after infec-tion, cells were washed and incubated in fresh medium containing genta-micin (50 �g/ml) to kill extracellular bacteria. To determine the numberof invaded bacteria (T 0 h) a portion of the cells was lysed with 0.1%Triton X-100 1 h after gentamicin treatment. The remaining cells werewashed and incubated for 23 h in medium containing 12 �g/ml gentami-cin. Subsequently, the macrophages were lysed with 0.1% Triton X-100and plated onto bacterial medium (T 24 h) to determine the percentageof viable intracellular bacteria according to the following equation: per-cent surviving bacteria number of bacteria (T 24 h)/number of bac-teria (T 0 h) � 100.

Electron microscopy. Infected macrophages were fixed with 2.5%glutaraldehyde and 2% tannin, washed with cacodylate buffer, and fixedwith 1% aqueous osmium for 1 h at room temperature. Dehydration wasachieved with a graded series of acetone (10, 30, and 50%) on ice. At 70%dehydration, samples were incubated overnight in 70% acetone and 2%uranyl acetate and dehydrated with 90% and 100% acetone on ice. Sam-ples were embedded in Spurr epoxy resin as described previously (30).Ultrathin sections were picked up with Butvar-coated grids, counter-stained with uranyl acetate and lead citrate, and examined with a TEM910transmission electron microscope (Carl Zeiss, Germany) at an accelera-tion voltage of 80 kV. Images were recorded digitally at calibrated magni-fications with a Slow-Scan charge-coupled-device (CCD) camera(ProScan; 1,024 by 1,024 pixels) with ITEM software (Olympus Soft Im-aging Solutions, Germany). Contrast and brightness were adjusted withAdobe Photoshop CS3.

Preparation and culture of murine and porcine bone marrow-de-rived macrophages. Six-week-old female C57BL/6 mice were obtainedfrom Harlan-Winkelmann (Borchen, Germany). The animals were main-tained, according to institutional guidelines, in individually ventilatedcages and were given food and water ad libitum. Harvesting of murine andporcine cells has been performed in compliance with the German animalprotection law (TierSchG BGBI S. 1206; 18 August 2006). Animals usedwere reported to the Lower Saxony State Office for Consumer Protectionand Food Safety according to the German laboratory animal reporting act(VersTierMeldV BGBI S. 2156; 4 November 1999). Bone marrow cellswere obtained by flushing the femurs and tibiae. Cells were strained by theuse of 100-�m cell strainers (BD Falcon) and pelleted by centrifugation.The pellets were resuspended and cultured in DMEM (Biochrom) supple-mented with 10% FCS (Biochrom) and 1% penicillin-streptomycin (Sigma-Aldrich) and containing 50 ng/ml murine macrophage colony-stimulat-ing factor (M-CSF) (PAN Biotech GmbH). On day 4, the amount ofM-CSF was reduced to 25 ng/ml. After an additional 4 days of cultivation,differentiated macrophages were obtained and used for infection experi-ments.

Five posterior ribs of German hybrid fattening pigs (German HybridPig Breeding Programme) were kindly provided by the Clinic for Swine,Small Ruminants and Forensic Medicine of the University of VeterinaryMedicine Hannover Foundation. The outer surface of the bones wascleaned with 70% ethanol, and the bone marrow was flushed. Prepara-tion, freezing, cultivation, and differentiation of the porcine bone mar-row-derived macrophages (BMDMs) were done according to methodsreported previously by Kapetanovic et al. (31). RPMI 1640 (GlutaMaxI;Gibco) supplemented with 10% porcine serum (Sigma-Aldrich) plus 1%penicillin-streptomycin (Sigma-Aldrich) was used as the culture medium.The medium contained 50 ng/ml human M-CSF (ProSpec) during thefirst 4 days of cultivation. At later time points, the M-CSF content wasreduced to 25 ng/ml. One, three, and seven days after preparation, differ-entiation of the porcine BMDMs was checked by flow cytometry analysisfor expression of the surface markers CD14, CD16, CD163 (AbD Serotec),

Host-Specific Colonization of Yersinia enterocolitica

November 2013 Volume 81 Number 11 iai.asm.org 4015


http://iai.asm.org/

Dow

nloaded from


and CD172a (Southern Biotec). Differentiated porcine BMDMs wereused for infection experiments on day 8 after preparation.

Determination of cytokine secretion by infected macrophages. Toexamine cytokine secretion induced by macrophages after infection, por-cine and murine BMDMs were infected with different Y. enterocoliticaisolates at an MOI of 10. At 0.5 h after infection, cells were washed andincubated in fresh medium containing gentamicin (50 �g/ml) to kill ex-tracellular bacteria. After 1 h, cells were washed and incubated for anadditional 23 h in medium containing 12 �g/ml gentamicin. Subse-quently, supernatants were collected and stored at �80°C. For experi-ments investigating the influence of IL-10, 10 ng/ml or 100 ng/ml recom-binant porcine IL-10 (Kingfisher Biosciences) was added to the medium30 min before and during infection. Levels of secreted murine and porcinecytokines and chemokines were determined with Procarta cytokine assaykits (Affymetrix) by using the Luminex detection system according to themanufacturer’s instructions.

Considering that each suite of measured cytokines comprises thehost’s response cytokine profile, the global cytokine pattern (using hier-archical clustering and ordination) can be used as a biomarker for dis-cerning pathogen exposure (32, 33). To do so, a data matrix comprisingthe concentrations of these cytokines, using the mean of n technical rep-licates for each sample, was compiled. Pairwise sample comparisons wereperformed by using the Bray-Curtis similarity algorithm, creating a re-semblance matrix that was then used to construct both group-averagehierarchical clustering (33) and principal coordinate analysis (PCoA)plots (32). The clusters are used to visualize and compare the influences ofdifferent Y. enterocolitica strains on cytokine secretion by BMDMs. Strainsthat provoke a similar pattern of cytokine secretion appear close togetherin the hierarchical cluster and ordination plot. Bubbles were superim-posed onto the PCoA ordination so as to represent the concentration ofeach cytokine within each particular sample. By using this method, differ-ences between the concentrations of the secreted cytokines are easily de-tectable, as reviewed by Genser et al. (32). All multivariate data analyseswere performed by using PRIMER v.6.1.6 according to the user manual/tutorial (PRIMER-E; Plymouth Marine Laboratory, United Kingdom).

Determination of cytokine secretion upon IL-10 induction by in-fected porcine macrophages. Blood of German hybrid fattening pigs(German Hybrid Pig Breeding Programme) was kindly provided by theClinic for Swine, Small Ruminants and Forensic Medicine of the Univer-sity of Veterinary Medicine Hannover Foundation. Fresh blood wasmixed with 0.85% NaCl and overlaid with Ficoll. After gradient centrifu-gation, all interphases were pooled and washed with RPMI medium. Cellswere incubated with magnetic anti-CD14 microbeads (Miltenyi Biotech),and monocytes were enriched by using the MACS magnetic column sys-tem (Miltenyi Biotech). Equilibration of the column was done by addingPBE buffer (1� PBS [pH 7.2], 0.5% BSA, 2 mM EDTA). Subsequently, thecell suspension was added onto the column, and all unlabeled cells werewashed off with PBE buffer. After the column was removed from themagnetic field, CD14-positive cells were eluted in PBE buffer. Differenti-ation of the cells was carried out as described above for porcine BMDMs.

Yop secretion and translocation. Y. enterocolitica strains of differentserotypes were grown overnight at 25°C in LB medium, diluted 1:50 infresh LB medium, and grown for 2 h at 25°C. To compare Yop secretionlevels, cultures were shifted to 37°C and grown for an additional 4 h in thepresence of 20 mM Mg2 and 20 mM Na-oxalate. Proteins in the super-natant were harvested, filtered (0.2 �m), and precipitated with trichloro-acetic acid (TCA). Precipitated proteins were washed with acetone, sus-pended in SDS sample buffer, separated on 12% SDS-polyacrylamide gels,and stained with Coomassie brilliant blue.

For the analysis of Yop translocation, bacteria were grown for an ad-ditional 2 h at 37°C and added to 2 � 105 porcine macrophages (PLN-C2)at an MOI of 70. At 2 h postinfection, cells were washed with PBS, resus-pended in SDS sample buffer, and separated on 12% SDS-polyacrylamidegels. Proteins were blotted onto a membrane, and intracellular Yop proteinswere visualized with an antiserum directed against all secreted Yop proteins.

RESULTSY. enterocolitica interaction with epithelial cells is serotype spe-cific but is not dependent on the source of the isolates and thehost. The ability of Y. enterocolitica to adhere to and invade epi-thelial cells is crucial for host tissue colonization and the progres-sion of the infection process. To compare cell binding and inva-sion properties of different Y. enterocolitica serotypes from varioussources, we tested their capacity to adhere to and invade threehuman (HEp-2, LoVo, and Caco-2), one murine (Mode-K), andtwo porcine (IPI-2I and IPEC-J2) intestinal-derived and nonin-testinal epithelial cell lines to ascertain potential host-specific col-onization properties. Since the expression patterns of predomi-nant Yersinia colonization factors such as invasin and YadAchange drastically upon a shift from 25°C to 37°C, cell adhesionand internalization were tested with bacteria grown at both tem-peratures. The adhesion rate of the well-studied Y. enterocolitica1B/O:8 strain 8081v grown at 25°C was used as the reference andset to 100%. We found that the overall cell adhesion and invasionpatterns of the different isolates on the tested human, murine, andporcine epithelial cells were very similar (Fig. 1 and 2; see also Fig.S1 and S2 in the supplemental material). All Y. enterocolitica sero-type O:5,27, O:8, and O:9 strains were able to adhere to and invadethe different epithelial cells when pregrown at 25°C and 37°C,although invasion of serotype O:8 strains was always reducedwhen the bacteria were pregrown at 37°C. Similar to what wasdescribed previously for human epithelial cells (34), serotype O:3strains were also unable to interact with porcine and murine epi-thelial cells when grown at 25°C (Fig. 1; see also Fig. S1 in thesupplemental material). In summary, this demonstrated that ad-hesiveness and invasiveness of the different isolates do not dependon the origin of the cell line and the source of the isolate. Observeddifferences of the colonization properties of Y. enterocolitica aremainly serotype specific.

Expression levels of YadA, InvA, and RovA differ stronglyamong the different Y. enterocolitica isolates. In order to inves-tigate the contribution of the main adhesion factors InvA andYadA to the different serotype-specific colonization patterns, wefirst analyzed expression levels of both pathogenicity factors. In-vasin of Y. enterocolitica O:8 strain 8081v was previously shown tobe maximally expressed at moderate temperatures, whereas YadAwas produced only at 37°C (35, 36). To compare the amounts ofYadA and InvA among all isolates, we grew the bacteria at 25°Cand 37°C. As shown in Fig. 3, no YadA, but invasin, was detectablein all isolates when the bacteria were grown at 25°C. At 37°C, YadAwas expressed in all strains at similar levels, whereby the apparentsize of the YadA molecules varied slightly among the isolates.However, YadA size alterations were not serotype specific andwere independent of the origin of the strains (Fig. 3). They seemedto be based on a various number of 15-mer repeats in the stalkregion, which determine the overall length of YadA but not thegeneral adhesion properties of the molecule (19). Large variationsof InvA and RovA levels were observed at 37°C (Fig. 3). Invasinlevels were generally very low in serotype O:8 strains and signifi-cantly reduced in all O:5,27 strains, whereas intermediate levelswere detectable in serotype O:9 strains, and larger amounts werefound in all serotype O:3 strains. To examine whether the differ-ential expression of invasin is based on different expression levelsof the InvA-regulatory protein RovA, we also tested the intracel-lular amounts of RovA in the Y. enterocolitica isolates. In general,

Schaake et al.



http://iai.asm.org/

Dow

nloaded from


RovA levels were high in serotype O:3 strains expressing high lev-els of invasin at 37°C, whereas smaller amounts were detected inO:5,27 and O:8 strains with significantly reduced invasin levels(Fig. 3). Larger variations of RovA levels were detected only

among the serotype O:9 strains. Taken together, the isolationsource of the Y. enterocolitica strains does not seem to have a majorimpact on the synthesis of invasin and YadA; only serotype-spe-cific variations of invasin levels were observed.

FIG 1 Interaction of Y. enterocolitica grown at 25°C with epithelial cells. Different Y. enterocolitica serotype O:3, O:5,27, O:8, and O:9 isolates from human patients oranimals were grown at 25°C overnight in LB medium. About 5 � 104 human (Caco-2), porcine (IPEC-J2), and murine (Mode-K) epithelial cells were infected with 5 �105 bacteria. After centrifugation of the bacteria onto the cells, the plate was incubated at 22°C to 25°C to monitor cell association or at 37°C to determine theinternalization efficiency of the bacteria by a gentamicin protection assay. Adhesion of Y. enterocolitica O:8 strain 8081v grown at 25°C was used as a reference and wasset to 100%. Data are presented as means � standard deviations of three independent experiments performed in duplicate. Data were analyzed by one-way analysis ofvariance with Dunnett’s multiple-comparison test. Asterisks indicate results that differed significantly from those for 8081v (��, P � 0.001).

FIG 2 Interaction of Y. enterocolitica grown at 37°C with epithelial cells. Different Y. enterocolitica serotype O:3, O:5,27, O:8, and O:9 isolates from human patients oranimals were grown at 37°C overnight in LB medium. About 5 � 104 intestine-derived human (Caco-2), porcine (IPEC-J2), and murine (Mode-K) epithelial cells wereinfected with 5�105 bacteria. After centrifugation of the bacteria onto the cells, the plate was incubated at 22°C to 25°C to monitor cell association or at 37°C to determinethe internalization efficiency of the bacteria by a gentamicin protection assay. Adhesion of Y. enterocolitica O:8 strain 8081v grown at 25°C was used as a reference and wasset to 100%. Data are presented as means � standard deviations of three independent experiments performed in duplicate.




http://iai.asm.org/

Dow

nloaded from


Y. enterocolitica interactions with and entry into macro-phages are serotype but not host specific. It is well documentedthat Y. enterocolitica can trigger or prevent uptake by phagocyticimmune cells such as macrophages. The invasin protein of Y. en-terocolitica interacts with �1-integrins on the macrophage surfaceand has been shown to induce actin cytoskeleton rearrangementsthat lead to phagocytic uptake of the pathogen (37). In contrast,bacteria primed to express the Ysc/Yop type III secretion pathwayinhibit phagocytosis by translocation of the Yop effector proteins,which inhibit cytoskeletal rearrangements (23). To assess the abil-ity of the different Y. enterocolitica isolates to trigger uptake byphagocytes, we tested their cell adhesion and invasion propertieswith two human (THP-1 and U937), two porcine (PLN-C2 and3D4/31), and two murine (J774A.1 and Raw264.7) macrophagecell lines (see Fig. S3 to S6 in the supplemental material). Strainswere pregrown at 25°C and 37°C to ensure the expression of inva-sin or YadA, the type III secretion system, and the Yop effectors.The Y. enterocolitica strains bound to and internalized in macro-phage cell lines in a serotype-specific manner, similar to what hasbeen observed with epithelial cell lines; e.g., the lowest adhesion

and invasion rates were observed with the Y. enterocolitica O:3strain at 25°C (see Fig. S3 to S6 in the supplemental material). It ispossible that the O:3-specific LPS structure might inhibit invasinfunction, which is known to be important for uptake by macro-phages (37). However, significant variations based on the differentorigins of the bacteria or the source of the cell lines were notdetectable.

Persistence of Y. enterocolitica in macrophages. Recent stud-ies also demonstrated that pathogenic yersiniae are able to surviveand replicate within macrophages (25, 38–40). However, it hasalso been demonstrated that the presence of a functional Ysc/Yoptype III secretion system during phagocytosis decreases survival ofinternalized bacteria and at the same time induces apoptosis of themacrophages by a Toll-like receptor 4 (TLR4)- and YopP/J-de-pendent mechanism (41). In order to address the abilities of dif-ferent Y. enterocolitica isolates to persist in macrophages, we firstdetermined the viability of infected macrophages by cytotoxicityassays. As shown in Fig. 4 and in Fig. S7 in the supplementalmaterial, the viability of macrophages infected with a Y. enteroco-litica O:3 or O:8 strain was strongly dependent on the MOI and the

FIG 3 Comparison of invasin, RovA, and YadA expression levels in strains of different Y. enterocolitica bioserotypes grown at 25°C and 37°C. Strains of differentY. enterocolitica serotypes (O:3, O:5,27, O:8, and O:9) were grown overnight in LB at 25°C and 37°C. Whole-cell extracts of identical amounts of bacteria wereprepared, separated on SDS-polyacrylamide gels, and analyzed by Western blotting using polyclonal antibodies directed against InvA, YadA, and RovA.Unspecific bands detected with the polyclonal antiserum served as the loading control (c, loading control for RovA blots [data not shown]). As a molecularmarker, a PageRuler prestained protein ladder was loaded on the left.

Schaake et al.



http://iai.asm.org/

Dow

nloaded from


incubation time after infection but was independent of the sero-type and the growth temperature of the bacteria. The latter sug-gested that induction of the Ysc/Yop type III secretion system at37°C had no influence on the killing of macrophages. To test thisassumption, we compared survival rates of human (THP-1) andporcine (PLN-C2) macrophages challenged with O:8 strain 8081v,O:3 strain Y1, and their pYV-cured derivatives YE22 and YE23,but no considerable difference was detectable (data not shown).To a small extent, cell survival was also host dependent. More than40% of the porcine macrophages underwent cell death after 48 h atan MOI of 1, whereas more than 90% of the human and murinephagocytes remained alive (Fig. 4).

Based on these results, Y. enterocolitica survival assays were per-formed with the human, porcine, and murine macrophage cell linesat an MOI of 10 for 24 h. To determine the percentage of phagocy-tized intracellular bacteria, a cell sample was lysed at 1 h postinfection,and numbers of cell-associated and internalized bacteria were deter-mined by plating. The numbers of bacterial CFU were determined 24h after infection to measure intracellular survival. Survival and repli-cation of the bacteria were somewhat less efficient in murine macro-phages and were mostly serotype dependent (Fig. 5; see also Fig. S8 inthe supplemental material). For instance, serotype O:5,27 isolateswere not able to survive and replicate in murine and porcine macro-phages; persistence was observed only in macrophages from humans.Interestingly, replication did not correlate with the number of in-vaded bacteria; e.g., serotype O:3 strains grown at 25°C were less in-

vasive than strains of other serotypes under the same conditions.However, significantly more bacteria were recovered 24 h after infec-tion, indicating that replication of serotype O:3 strains inside themacrophages is more efficient (Fig. 5; see also Fig. S3 in the supple-mental material).

Persistence of serotype O:3, O:8, and O:9 strains within themacrophages was generally increased when the bacteria were pre-grown at 37°C (Fig. 5). Using pYV-cured derivatives of serotypeO:8 and O:3 strains 8081 and Y1 (YE22 and YE23), we demon-strated that this effect is not caused by the induction of pYV-encoded virulence factors, since the survival rate of the pYV-curedstrains was still somewhat higher when the strains were pregrownat 37°C (see Fig. S9 in the supplemental material). On the con-trary, loss of the virulence plasmid enhanced intracellular replica-tion in both types of macrophages (see Fig. S9 in the supplementalmaterial), indicating that expression of components of the Ysc/Yop secretion machinery prevents intracellular replication of thebacteria, similar to what has been observed with Yersinia pseudo-tuberculosis (41). We further tested expression of the Ysc/Yop se-cretion machinery and found that comparable amounts of theeffectors are secreted and/or translocated into porcine macro-phages by the different Y. enterocolitica serotypes (see Fig. S10 inthe supplemental material). We also analyzed the intracellular lo-cation of these strains by transmission electron microscopy anddemonstrated that the bacteria reside and replicate within a mem-

FIG 4 MOI-dependent cell viability of infected macrophages. About 5 � 105 human (THP-1), porcine (PLN-C2), and murine (J774.A) macrophages wereinfected with Y. enterocolitica 8081v and Y1 grown at 25°C and 37°C. To examine cell viability, cells were infected at an MOI of 1, 10, or 50. At 2, 24, and 48 hpostinfection, the cell viability of the macrophages was analyzed by using the Live/Dead Viability/Cytotoxicity kit for mammalian cells (Invitrogen).




http://iai.asm.org/

Dow

nloaded from


brane-bound vacuole (phagosomal compartment) in both humanand porcine macrophages (Fig. 6).

Y. enterocolitica isolates induce a host- and serotype-specificcytokine response. To examine whether the amount and profile ofreleased cytokines vary between macrophages of different host spe-cies, we used Luminex cytokine analysis technology to determine thecytokine levels in supernatants of infected porcine and murine bonemarrow-derived macrophages (BMDMs). BMDMs were used be-cause a significantly reduced intrinsic activation state and low cyto-kine expression levels for lipopolysaccharides were observed withmacrophage cell lines (IC-21, J774A.1, and Raw264.7) comparedwith primary macrophages (42). Secretion of several murine and por-cine cytokines (TNF-�, macrophage inflammatory protein 2 [MIP-2]/IL-8, IL-1�, IL-4, IL-6, IL-12, alpha interferon [IFN-�], IFN-,and IL-10) were tested. TNF-�, IL-1�, and IL-8/MIP-2 are mainlyinvolved in the coordination of primary host immune responses toinfection and are strongly induced during a Y. enterocolitica infection.They lead to the recruitment of neutrophils and macrophages andinitiate reactions of the adaptive immune system (43–45). In contrast,IL-10 is a potent anti-inflammatory cytokine that normally serves tocounterregulate proinflammatory processes that can be detrimentalfor the host (e.g., bacterial sepsis), and it is involved in mediatingimmune tolerance in the gastrointestinal tract (46, 47). Infection wasperformed with Y. enterocolitica strains of serotypes O:3, O:5,27, O:9,and O:8 grown at 25°C at an MOI of 10. Supernatants were harvestedand analyzed 24 h after infection.

Mainly TNF-�, IL-8/MIP-2, and also IL-10 were induced with

all tested Y. enterocolitica serotypes by both porcine and murineBMDMs compared to uninfected cells. Global cytokine secretionprofiles comparing the secretion patterns of these cytokines byinfected murine BMDMs revealed two main clusters (Fig. 7A andC). The cytokine secretion pattern induced by Y. enterocoliticastrains Y1 (serotype O:3), 3192 (serotype O:5,27), and 651 (sero-type O:9) differed considerably from the pattern induced by strain8081v (serotype O:8), which is much more virulent for mice. Nodistinct cluster can be defined for cytokine secretion by porcineBMDMs. The most distant strain is Y1, which belongs to serotypeO:3, the most prevalent serotype in pigs (Fig. 7B and D).

Secretion of TNF-� and IL-1� (Fig. 8 and data not shown; seealso Fig. S11 in the supplemental material) was strongly upregu-lated in porcine and murine BMDMs, but no host-specific differ-ences in secretion levels were observed. MIP-2 levels secreted byinfected murine BMDMs were also comparable among the differ-ent Y. enterocolitica strains. However, Y. enterocolitica-inducedIL-8 secretion by porcine BMDMs varied strongly between thedifferent serotypes. IL-8 levels were strongly increased after infec-tion with Y. enterocolitica O:8 strains, and intermediate levels weredetected with Y. enterocolitica O:5,27 and O:9 strains, whereasonly a very small amount of secreted IL-8 was detectable afterinfection with Y. enterocolitica O:3 strains. Furthermore, host-spe-cific differences in the secretion of IL-10 were detected. Most in-terestingly, IL-10 secretion was only slightly increased after infec-tion of murine BMDMs with the different Y. enterocoliticaserotypes, whereas much higher levels of IL-10 were detected for

FIG 5 Y. enterocolitica survival within macrophages. Y. enterocolitica isolates were grown at 25°C or 37°C overnight in LB medium. About 5 � 105 human(THP-1), porcine (PLN-C2), and murine (J774.A) macrophages were infected with 5 � 106 bacteria and incubated at 37°C for 1 h (T 0 h) to determine theinternalization efficiency of the bacteria by a gentamicin protection assay. The determination of survival was performed after an incubation of 24 h ingentamicin-containing buffer (T 24 h). The percentage of surviving bacteria was calculated. Data are presented as means � standard deviations of threeindependent experiments performed in duplicate. The gray lines indicate 100% survival.

Schaake et al.



http://iai.asm.org/

Dow

nloaded from


all Y. enterocolitica-infected porcine BMDMs (Fig. 8). These dif-ferences might contribute to the different inflammatory responsesobserved during infection of swine and mice. In agreement withthis assumption, we found that the addition of IL-10 reduced Y.enterocolitica Y1- and 8081-mediated induction of proinflamma-tory cytokines in porcine macrophages (see Fig. S11 in the supple-mental material).

DISCUSSION

Y. enterocolitica is transmitted via the fecal-oral route among hu-mans and animals. A wide range of animal species can be infectedby Y. enterocolitica, but the severity of infection varies significantlybetween different hosts (1). Y. enterocolitica infections often prog-ress asymptomatically in pigs. As a consequence, apparentlyhealthy pigs are a prevalent reservoir of Y. enterocolitica strains inthe United States and Europe and are known to be the majorsource of Y. enterocolitica infections (9, 48).

The aim of this study was to investigate virulence strategies ofthe pathogens that might contribute to the potential of Y. entero-colitica to cause disease symptoms in humans and rodents but notin swine. A set of different Y. enterocolitica isolates of serotypesO:3, O:5,27, O:8, and O:9 collected between 1976 and 2008 fromanimals and patients in Europe were used to infect intestinal epi-

thelial cells and macrophages from human, mouse, and swine (foran overview, see Table S1 in the supplemental material). All testedanimal and human isolates were able to adhere to and enter thedifferent host epithelial and macrophage cell lines. This observa-tion is supported by a previous study demonstrating that British Y.enterocolitica isolates from biotypes 2 to 4 are more invasive thanbiotype 1A strains, but similarly to our study, invasiveness ap-peared to be independent of the source of the isolate (49).

In addition, the averaged association with and invasion by anindividual isolate remained similar, suggesting that adhesivenessand invasiveness are not host specific. This is different from otherenteric pathogens. Host specificity is well known for some Salmo-nella enterica serovars. For instance, S. enterica serovar Typhi andS. Paratyphi infect only humans, S. Gallinarum colonizes chick-ens, S. Dublin infects cattle, and S. Cholerasuis infects pigs (50–52). S. Typhimurium causes infections without any specificity, butthe severity varies between host species; e.g., the type 3 secretionsystems encoded on Salmonella pathogenicity islands (SPIs) 1 and2 are important for efficient colonization of cattle, but their dis-ruption impairs virulence in chickens only mildly (53).

Although the host cell binding and entry properties of patho-genic Y. enterocolitica isolates did not seem to be host and origin

FIG 6 Localization of Y. enterocolitica inside macrophages. Human (THP-1), porcine (PLN-C2), or murine (J774A.1) macrophages were infected with Y.enterocolitica strains 8081v, YE22 (8081 �pYV), Y1, and YE23 (Y1 �pYV) at an MOI of 70 for 2 h. Localization of the bacteria after infection was determined bytransmission electron microscopy. Arrowheads mark the bacterial as well as the host cell membranes, indicating the localization of the bacteria inside phagosomalcompartments.




http://iai.asm.org/

Dow

nloaded from


specific, certain serotype- and strain-specific differences were ob-served. In contrast to serotype O:5,27, O:8, and O:9 strains, alltested O:3 isolates exhibited a very low ability to interact and in-vade human, porcine, and murine intestinal cells when grown at25°C. This phenotype is caused by the presence of a specificbranched LPS structure consisting of a special O antigen and anouter core which inhibits invasin-mediated cell contact (34).

Serotype-specific differences were also observed for the inva-sion of epithelial cells. Downregulation of inv expression at 37°C,as characterized for YeO:8 strain 8081v (35), results in a reductionof host cell entry of serotype O:8 strains at the host temperature.However, only a slight reduction of the invasin expression levelwas seen in O:5,27 strains at 37°C, but no significant changes weredetected in the highly prevalent serotype O:3 or O:9 strains. Inter-estingly, nucleotide differences in the invO:9 promoter, recentlyidentified in YeO:9 strain Y127, led to a 4-fold-higher inv expres-sion level at 37°C than for YeO:8 strain 8081v (54). Large amountsof invasin in serotype O:3 strains were shown to result from anadditional temperature-independent promoter encoded by an in-

serted IS1667 element within the inv promoter region and a morestable RovA protein activating inv transcription (34). Taken to-gether, this indicates that invasin of Y. enterocolitica might bemore important for the establishment of an infection in humansand domestic animals, as previous mouse infections with O:8strains have indicated.

After invasion and transmigration through the M cells of theintestinal layer, yersiniae are encountered by invading phagocyticimmune cells, in particular neutrophils and macrophages, as partof the first immune response. Multiple reports demonstrated theability of Y. enterocolitica to prevent phagocytosis by macrophagesthrough the function of the type III secreted Yop effector proteins(23). However, increasing evidence exists that yersiniae activelyinvade, persist, and replicate within intact macrophages, which islikely to be important during early stages of infection (23, 25, 38).Analysis of the adhesion, uptake, and intracellular location of dif-ferent Yersinia isolates within human, murine, and porcine mac-rophages revealed no differences, indicating that interactions withthese immune cells are also not host specific. In contrast, striking

FIG 7 Global cytokine secretion profiles of murine and porcine BMDMs upon infection by different Y. enterocolitica isolates. Y. enterocolitica isolates were grown at 25°Covernight in LB medium. About 5 � 105 porcine or murine BMDMs were infected with 5 � 106 bacteria and incubated at 37°C for 30 min using gentamicin protectionassay protocols. After 24 h of incubation, levels of porcine and murine TNF-�, IL-8/MIP-2, and IL-10 were determined by using Procarta cytokine assay kits (Affymetrix)according to the manufacturer’s instructions. (A and B) Group-average hierarchical clustering illustrating the influence of different Y. enterocolitica isolates on cytokinesecretion by murine (A) and porcine (B) BMDMs. (C and D) Principal coordinate analysis (PCoA) plot of different Y. enterocolitica isolates used to infect murine (C) andporcine (D) BMDMs. The analysis demonstrates the impacts of different Y. enterocolitica strains on cytokine secretion by BMDMs. In both PCoA plots, PC1 and PC2 axescollectively account for �96% of the explained variation, thus leading to little chance of misinterpretation.

Schaake et al.



http://iai.asm.org/

Dow

nloaded from


differences were observed when the persistence of bacteria and levelsof secreted cytokines were compared. Both porcine macrophage celllines seem more sensitive to Y. enterocolitica, and the overall survivalrate of the bacteria was lower in cultured murine than in human andporcine macrophages. Also, serotype-specific differences became ev-ident; i.e., all tested serotype O:5,27 strains were able to persist inhuman but not in porcine and murine macrophages. Several geneshave been identified as being important for survival and adaptation tothe intracellular stress situation. The PhoP/PhoQ two-componentsystem was shown to be essential for the survival of Yersinia (55, 56).The osmotic regulator OmpR and the global stress requirement pro-tein GsrA are necessary for the stress adaptation of Y. enterocolitica

inside macrophages (40, 57). In addition, genes encoding glucose-1-phosphate-uridyltransferase (galU), a UDP-N-acetylglucosamine-2-epimerase (wecB), and a UDP-N-acetyl-D-mannosamine dehydroge-nase (wecC) were shown to be important for survival of Yersinia inmacrophages (58). However, whether variations in macrophage per-sistence and replication result from differences in the expression orcoding sequence of these components is unknown.

Several studies analyzed the inflammatory responses of celllines or the intestinal mucosa of mice, but the characteristics anddistinctions of the immune response during an asymptomaticinfection in pigs are still unclear. To evaluate the response of mac-rophages in inflammation in pigs and mice upon Yersinia infec-

FIG 8 Y. enterocolitica triggers TNF-�, IL-8/MIP-2, and IL-10 secretion by porcine and murine BMDMs. Y. enterocolitica isolates were grown at 25°C overnightin LB medium. About 5 � 105 porcine or murine BMDMs were infected with 5 � 106 bacteria and incubated at 37°C for 30 min using gentamicin protection assayprotocols. After 24 h of incubation, levels of porcine and murine TNF-�, IL-8/MIP-2, and IL-10 were determined by using Procarta cytokine assay kits(Affimetrix) according to the manufacturer’s instructions. Shown are PCoA plots of TNF-� (A and B), MIP-2 (C), IL-8 (D), and IL-10 (E and F) for different Y.enterocolitica isolates used to infect murine (A, C, and E) and porcine (B, D, and F) BMDMs. The diameters of the symbols illustrate the concentrations of secretedcytokines, which can be compared between the different Y. enterocolitica strains by PCoA.




http://iai.asm.org/

Dow

nloaded from


tion, we used murine and porcine BMDMs and analyzed Yersinia-triggered cytokine production by these cells. Induction of themurine and porcine BMDMs with different Y. enterocolitica sero-types resulted in an increase of the secretion of several proinflam-matory cytokines, including TNF-�, IL-1�, and IL-8/MIP-2.These cytokines have a wide spectrum of activity, coordinatinghost responses to infections, e.g., activation of complement op-sonization, and initiation of various innate and adaptive immuneresponses. They were also found to be secreted by human mac-rophage-type cell lines upon infection with biotype 3, 4, and 1Astrains, which are classically defined as nonpathogenic due to theirlack of lethality in the mouse infection model (49, 59). Further-more, expression levels of these cytokines were increased uponinvasin-mediated binding of YeO:8 strain 8081v to HeLa cells, andthey were shown to play an important role in clearing YeO:8 in-fection in the murine model (22, 26, 43–45).

We found that Yersinia-triggered IL-8 induction by porcinemacrophages varied considerably between the Y. enterocolitica se-rotypes. IL-8 secretion was maximally induced by YeO:8 strains,whereas YeO:3 and YeO:9 strains showed only a very weak or lowlevel of induction of IL-8 secretion. IL-8 secretion triggers primar-ily migration of neutrophils and other granulocytes toward thesite of infection, which is associated with inflammation but leadsto rapid clearance of the infection (60).

Secretion of the anti-inflammatory cytokine IL-10 was studied,which counterregulates proinflammatory responses, amelioratesimmunopathology, and plays a role in mediating immune toler-ance in the gastrointestinal tract (46, 47). It was previously shownthat the Y. enterocolitica V antigen can drive IL-10 production, andIL-10�/� mice were highly resistant to Y. enterocolitica O:8 infec-tion, as shown by the lower bacterial loads in spleen and liver,absent abscess formation in these organs, and survival (61–63).We found that IL-10 secretion is induced by both murine andporcine macrophages with all tested Y. enterocolitica serotypes.However, IL-10 production was much less induced in murineBMDMs than in porcine BMDMs. IL-10 is known to be producedby many different myeloid and lymphoid cells, e.g., macrophages,dendritic cells, B cells, and regulatory T cells (64). During infec-tion, it suppresses functions of dendritic cells, NK cells, and mac-rophages, all of which are required for optimal pathogen clear-ance, but they also contribute to tissue damage. As a consequence,IL-10 can both impede pathogen clearance and ameliorate immu-nopathology. It is therefore tempting to speculate that the highprevalence of serotype O:3 in pigs could be a consequence of bothelevated IL-10- and decreased IL-8-mediated immune responses.This would directly inhibit pathogen clearance and may be in-duced by the pathogen to promote its own survival. Furthermore,it would reduce inflammation, prevent severe immunopathology,and give rise to persistent (asymptomatic) infections. A more in-depth analysis of the immune responses to Y. enterocolitica infec-tion in pigs is required to address this assumption and is part ofour future work.

ACKNOWLEDGMENTS

We thank Martin Fenner and Fabio Pisano for critical reading of themanuscript; Lothar Wieler and Karsten Tedin for the supply of cell lines;the Clinic for Swine, Small Ruminants and Forensic Medicine of the Uni-versity of Veterinary Medicine Hannover Foundation for the ribs and theblood of German hybrid fattening pigs (German Hybrid Pig BreedingProgramme); Maria Fällman for the anti-Yop antibodies; Ina Schleicher

for technical assistance in the electron microscopy studies; Jörn Pezoldtfor advice on using Luminex; and Tanja Thiermann for technical supportin the preparation of BMDMs.

This research project has been supported by the German Federal Min-istry for Research and Education (BMBF) (Consortium FBI-Zoo), theFonds der Chemischen Industrie, and the President’s Initiative and Net-working Fund of the Helmholtz Association of German Research Centres(HGF) under contract number VH-GS-202.

REFERENCES1. Bottone EJ. 1999. Yersinia enterocolitica: overview and epidemiologic cor-

relates. Microbes Infect. 1:323–333.2. Tauxe RV. 2002. Emerging foodborne pathogens. Int. J. Food Microbiol.

78:31– 41.3. Jones TF. 2003. From pig to pacifier: chitterling-associated yersiniosis

outbreak among black infants. Emerg. Infect. Dis. 9:1007–1009.4. Fredriksson-Ahomaa M. 2012. Isolation of enteropathogenic Yersinia

from non-human sources. Adv. Exp. Med. Biol. 954:97–105.5. McNally A, Cheasty T, Fearnley C, Dalziel RW, Paiba GA, Manning G,

Newell DG. 2004. Comparison of the biotypes of Yersinia enterocoliticaisolated from pigs, cattle and sheep at slaughter and from humans withyersiniosis in Great Britain during 1999-2000. Lett. Appl. Microbiol. 39:103–108.

6. Thibodeau V, Frost EH, Quessy S. 2001. Development of an ELISAprocedure to detect swine carriers of pathogenic Yersinia enterocolitica.Vet. Microbiol. 82:249 –259.

7. Nielsen B, Heisel C, Wingstrand A. 1996. Time course of the serologicalresponse to Yersinia enterocolitica O:3 in experimentally infected pigs. Vet.Microbiol. 48:293–303.

8. Bhaduri S, Wesley IV, Bush EJ. 2005. Prevalence of pathogenic Yersiniaenterocolitica strains in pigs in the United States. Appl. Environ. Micro-biol. 71:7117–7121.

9. Tauxe RV, Vandepitte J, Wauters G, Martin SM, Goossens V, De MolP, Van Noyen R, Thiers G. 1987. Yersinia enterocolitica infections andpork: the missing link. Lancet i:1129 –1132.

10. Najdenski H, Golkocheva E, Kussovski V, Ivanova E, Manov V, Iliev M,Vesselinova A, Bengoechea JA, Skurnik M. 2006. Experimental pigyersiniosis to assess attenuation of Yersinia enterocolitica O:8 mutantstrains. FEMS Immunol. Med. Microbiol. 47:425– 435.

11. de Boer E, Nouws JF. 1991. Slaughter pigs and pork as a source of humanpathogenic Yersinia enterocolitica. Int. J. Food Microbiol. 12:375–378.

12. Dube P. 2009. Interaction of Yersinia with the gut: mechanisms of patho-genesis and immune evasion. Curr. Top. Microbiol. Immunol. 337:61–91.

13. Cornelis GR, Boland A, Boyd AP, Geuijen C, Iriarte M, Neyt C, SoryMP, Stainier I. 1998. The virulence plasmid of Yersinia, an antihost ge-nome. Microbiol. Mol. Biol. Rev. 62:1315–1352.

14. Revell PA, Miller VL. 2001. Yersinia virulence: more than a plasmid.FEMS Microbiol. Lett. 205:159 –164.

15. Pepe JC, Miller VL. 1993. Yersinia enterocolitica invasin: a primary role inthe initiation of infection. Proc. Natl. Acad. Sci. U. S. A. 90:6473– 6477.

16. Isberg RR, Leong JM. 1990. Multiple beta 1 chain integrins are receptorsfor invasin, a protein that promotes bacterial penetration into mammaliancells. Cell 60:861– 871.

17. Herbst K, Bujara M, Heroven AK, Opitz W, Weichert M, ZimmermannA, Dersch P. 2009. Intrinsic thermal sensing controls proteolysis of Yer-sinia virulence regulator RovA. PLoS Pathog. 5:e1000435. doi:10.1371/journal.ppat.1000435.

18. Lawrenz MB, Miller VL. 2007. Comparative analysis of the regulation ofrovA from the pathogenic yersiniae. J. Bacteriol. 189:5963–5975.

19. El Tahir Y, Skurnik M. 2001. YadA, the multifaceted Yersinia adhesin.Int. J. Med. Microbiol. 291:209 –218.

20. Kirjavainen V, Jarva H, Biedzka-Sarek M, Blom AM, Skurnik M, MeriS. 2008. Yersinia enterocolitica serum resistance proteins YadA and Ailbind the complement regulator C4b-binding protein. PLoS Pathog.4:e1000140. doi:10.1371/journal.ppat.1000140.

21. Eitel J, Heise T, Thiesen U, Dersch P. 2005. Cell invasion and IL-8production pathways initiated by YadA of Yersinia pseudotuberculosis re-quire common signalling molecules (FAK, c-SRC, Ras) and distinct cellfactors. Cell. Microbiol. 7:63–77.

22. Schmid Y, Grassl GA, Buhler OT, Skurnik M, Autenrieth IB, Bohn E.2004. Yersinia enterocolitica adhesin A induces production of interleu-kin-8 in epithelial cells. Infect. Immun. 72:6780 – 6789.

Schaake et al.



http://iai.asm.org/

Dow

nloaded from

http://dx.doi.org/10.1371/journal.ppat.1000435http://dx.doi.org/10.1371/journal.ppat.1000435http://dx.doi.org/10.1371/journal.ppat.1000140http://iai.asm.orghttp://iai.asm.org/

23. Viboud GI, Bliska JB. 2005. Yersinia outer proteins: role in modulation ofhost cell signaling responses and pathogenesis. Annu. Rev. Microbiol. 59:69 – 89.

24. Autenrieth SE, Autenrieth IB. 2008. Yersinia enterocolitica: subversion ofadaptive immunity and implications for vaccine development. Int. J. Med.Microbiol. 298:69 –77.

25. Pujol C, Bliska JB. 2005. Turning Yersinia pathogenesis outside in: sub-version of macrophage function by intracellular yersiniae. Clin. Immunol.114:216 –226.

26. Autenrieth IB, Kempf V, Sprinz T, Preger S, Schnell A. 1996. Defensemechanisms in Peyer’s patches and mesenteric lymph nodes against Yer-sinia enterocolitica involve integrins and cytokines. Infect. Immun. 64:1357–1368.

27. Eitel J, Dersch P. 2002. The YadA protein of Yersinia pseudotuberculosismediates high-efficiency uptake into human cells under environmentalconditions in which invasin is repressed. Infect. Immun. 70:4880 – 4891.

28. Uliczka F, Kornprobst T, Eitel J, Schneider D, Dersch P. 2009. Cellinvasion of Yersinia pseudotuberculosis by invasin and YadA requires pro-tein kinase C, phospholipase C-gamma1 and Akt kinase. Cell. Microbiol.11:1782–1801.

29. Dersch P, Isberg RR. 1999. A region of the Yersinia pseudotuberculosisinvasin protein enhances integrin-mediated uptake into mammalian cellsand promotes self-association. EMBO J. 18:1199 –1213.

30. Spurr AR. 1969. A low-viscosity epoxy resin embedding medium forelectron microscopy. J. Ultrastruct. Res. 26:31– 43.

31. Kapetanovic R, Fairbairn L, Beraldi D, Sester DP, Archibald AL, TuggleCK, Hume DA. 2012. Pig bone marrow-derived macrophages resemblehuman macrophages in their response to bacterial lipopolysaccharide. J.Immunol. 188:3382–3394.

32. Genser B, Cooper PJ, Yazdanbakhsh M, Barreto ML, Rodrigues LC.2007. A guide to modern statistical analysis of immunological data. BMCImmunol. 8:27. doi:10.1186/1471-2172-8-27.

33. Chromy BA, Fodor IK, Montgomery NK, Luciw PA, McCutchen-Maloney SL. 2012. Cluster analysis of host cytokine responses to biode-fense pathogens in a whole blood ex vivo exposure model (WEEM). BMCMicrobiol. 12:79. doi:10.1186/1471-2180-12-79.

34. Uliczka F, Pisano F, Schaake J, Stolz T, Rohde M, Fruth A, Strauch E,Skurnik M, Batzilla J, Rakin A, Heesemann J, Dersch P. 2011. Uniquecell adhesion and invasion properties of Yersinia enterocolitica O:3, themost frequent cause of human yersiniosis. PLoS Pathog. 7:e1002117. doi:10.1371/journal.ppat.1002117.

35. Pepe JC, Badger JL, Miller VL. 1994. Growth phase and low pH affect thethermal regulation of the Yersinia enterocolitica inv gene. Mol. Microbiol.11:123–135.

36. Lambert de Rouvroit C, Sluiters C, Cornelis GR. 1992. Role of thetranscriptional activator, VirF, and temperature in the expression of thepYV plasmid genes of Yersinia enterocolitica. Mol. Microbiol. 6:395– 409.

37. Wiedemann A, Linder S, Grassl G, Albert M, Autenrieth I, AepfelbacherM. 2001. Yersinia enterocolitica invasin triggers phagocytosis via beta1integrins, CDC42Hs and WASp in macrophages. Cell. Microbiol. 3:693–702.

38. Pujol C, Bliska JB. 2003. The ability to replicate in macrophages is con-served between Yersinia pestis and Yersinia pseudotuberculosis. Infect. Im-mun. 71:5892–5899.

39. Yamamoto T, Hanawa T, Ogata S, Kamiya S. 1997. The Yersinia entero-colitica GsrA stress protein, involved in intracellular survival, is induced bymacrophage phagocytosis. Infect. Immun. 65:2190 –2196.

40. Brzostek K, Raczkowska A, Zasada A. 2003. The osmotic regulatorOmpR is involved in the response of Yersinia enterocolitica O:9 to envi-ronmental stresses and survival within macrophages. FEMS Microbiol.Lett. 228:265–271.

41. Zhang Y, Murtha J, Roberts MA, Siegel RM, Bliska JB. 2008. Type IIIsecretion decreases bacterial and host survival following phagocytosis of Yer-sinia pseudotuberculosis by macrophages. Infect. Immun. 76:4299–4310.

42. Chamberlain L, Godek M, Gonzalez-Juarrero M, Grainger DW. 2009.Phenotypic non-equivalence of murine (monocyte-) macrophage cells inbiomaterial and inflammatory models. J. Biomed. Mater. Res. 88:858–871.

43. Kampik D, Schulte R, Autenrieth IB. 2000. Yersinia enterocolitica invasinprotein triggers differential production of interleukin-1, interleukin-8,monocyte chemoattractant protein 1, granulocyte-macrophage colony-stimulating factor, and tumor necrosis factor alpha in epithelial cells: im-plications for understanding the early cytokine network in Yersinia infec-tions. Infect. Immun. 68:2484 –2492.

44. Dube PH, Handley SA, Lewis J, Miller VL. 2004. Protective role of inter-leukin-6 during Yersinia enterocolitica infection is mediated through the mod-ulation of inflammatory cytokines. Infect. Immun. 72:3561–3570.

45. Dube PH, Revell PA, Chaplin DD, Lorenz RG, Miller VL. 2001. A rolefor IL-1 alpha in inducing pathologic inflammation during bacterial in-fection. Proc. Natl. Acad. Sci. U. S. A. 98:10880 –10885.

46. Couper KN, Blount DG, Riley EM. 2008. IL-10: the master regulator ofimmunity to infection. J. Immunol. 180:5771–5777.

47. Saraiva M, O’Garra A. 2010. The regulation of IL-10 production byimmune cells. Nat. Rev. Immunol. 10:170 –181.

48. Rosner BM, Stark K, Werber D. 2010. Epidemiology of reported Yersiniaenterocolitica infections in Germany, 2001-2008. BMC Public Health 10:337. doi:10.1186/1471-2458-10-337.

49. McNally A, Dalton T, La Ragione RM, Stapleton K, Manning G, NewellDG. 2006. Yersinia enterocolitica isolates of differing biotypes from hu-mans and animals are adherent, invasive and persist in macrophages, butdiffer in cytokine secretion profiles in vitro. J. Med. Microbiol. 55:1725–1734.

50. Alpuche-Aranda CM, Berthiaume EP, Mock B, Swanson JA, Miller SI.1995. Spacious phagosome formation within mouse macrophages corre-lates with Salmonella serotype pathogenicity and host susceptibility. In-fect. Immun. 63:4456 – 4462.

51. Barrow PA, Huggins MB, Lovell MA. 1994. Host specificity of Salmonellainfection in chickens and mice is expressed in vivo primarily at the level ofthe reticuloendothelial system. Infect. Immun. 62:4602– 4610.

52. Pascopella L, Raupach B, Ghori N, Monack D, Falkow S, Small PL.1995. Host restriction phenotypes of Salmonella typhi and Salmonella gal-linarum. Infect. Immun. 63:4329 – 4335.

53. Morgan E, Campbell JD, Rowe SC, Bispham J, Stevens MP, Bowen AJ,Barrow PA, Maskell DJ, Wallis TS. 2004. Identification of host-specificcolonization factors of Salmonella enterica serovar Typhimurium. Mol.Microbiol. 54:994 –1010.

54. Trcek J, Fuchs TM, Trulzsch K. 2010. Analysis of Yersinia enterocoliticainvasin expression in vitro and in vivo using a novel luxCDABE reportersystem. Microbiology 156:2734 –2745.

55. Grabenstein JP, Marceau M, Pujol C, Simonet M, Bliska JB. 2004. Theresponse regulator PhoP of Yersinia pseudotuberculosis is important forreplication in macrophages and for virulence. Infect. Immun. 72:4973–4984.

56. Bozue J, Mou S, Moody KL, Cote CK, Trevino S, Fritz D, Worsham P.2011. The role of the phoPQ operon in the pathogenesis of the fully viru-lent CO92 strain of Yersinia pestis and the IP32953 strain of Yersinia pseu-dotuberculosis. Microb. Pathog. 50:314 –321.

57. Yamamoto T, Hanawa T, Ogata S, Kamiya S. 1996. Identification andcharacterization of the Yersinia enterocolitica gsrA gene, which protectivelyresponds to intracellular stress induced by macrophage phagocytosis andto extracellular environmental stress. Infect. Immun. 64:2980 –2987.

58. Klein KA, Fukuto HS, Pelletier M, Romanov G, Grabenstein JP, PalmerLE, Ernst R, Bliska JB. 2012. A transposon site hybridization screenidentifies galU and wecBC as important for survival of Yersinia pestis inmurine macrophages. J. Bacteriol. 194:653– 662.

59. Tennant SM, Grant TH, Robins-Browne RM. 2003. Pathogenicity ofYersinia enterocolitica biotype 1A. FEMS Immunol. Med. Microbiol. 38:127–137.

60. Baggiolini M, Clark-Lewis I. 1992. Interleukin-8, a chemotactic andinflammatory cytokine. FEBS Lett. 307:97–101.

61. Sing A, Rost D, Tvardovskaia N, Roggenkamp A, Wiedemann A,Kirschning CJ, Aepfelbacher M, Heesemann J. 2002. Yersinia V-antigenexploits Toll-like receptor 2 and CD14 for interleukin 10-mediated im-munosuppression. J. Exp. Med. 196:1017–1024.

62. Sing A, Roggenkamp A, Geiger AM, Heesemann J. 2002. Yersinia en-terocolitica evasion of the host innate immune response by V antigen-induced IL-10 production of macrophages is abrogated in IL-10-deficientmice. J. Immunol. 168:1315–1321.

63. Sing A, Reithmeier-Rost D, Granfors K, Hill J, Roggenkamp A, Heese-mann J. 2005. A hypervariable N-terminal region of Yersinia LcrV deter-mines Toll-like receptor 2-mediated IL-10 induction and mouse viru-lence. Proc. Natl. Acad. Sci. U. S. A. 102:16049 –16054.

64. Moore KW, de Waal Malefyt R, Coffman RL, O’Garra A. 2001. Inter-leukin-10 and the interleukin-10 receptor. Annu. Rev. Immunol. 19:683–765.

65. Manoil C, Beckwith J. 1985. TnphoA: a transposon probe for proteinexport signals. Proc. Natl. Acad. Sci. U. S. A. 82:8129 – 8133.




http://iai.asm.org/

Dow

nloaded from

http://dx.doi.org/10.1186/1471-2172-8-27http://dx.doi.org/10.1186/1471-2180-12-79http://dx.doi.org/10.1371/journal.ppat.1002117http://dx.doi.org/10.1186/1471-2458-10-337http://iai.asm.orghttp://iai.asm.org/

Human and Animal Isolates of Yersinia enterocolitica Show Significant Serotype-Specific Colonization and Host-Specific Immune Defense PropertiesMATERIALS AND METHODSBacterial strains, cell culture, media, and growth conditions.Gel electrophoresis, preparation of cell extracts, and Western blotting.Cell adhesion and invasion assays.Cell viability tests of macrophages.Assays of bacterial survival in macrophages.Electron microscopy.Preparation and culture of murine and porcine bone marrow-derived macrophages.Determination of cytokine secretion by infected macrophages.Determination of cytokine secretion upon IL-10 induction by infected porcine macrophages.Yop secretion and translocation.

RESULTSY. enterocolitica interaction with epithelial cells is serotype specific but is not dependent on the source of the isolates and the host.Expression levels of YadA, InvA, and RovA differ strongly among the different Y. enterocolitica isolates.Y. enterocolitica interactions with and entry into macrophages are serotype but not host specific.Persistence of Y. enterocolitica in macrophages.Y. enterocolitica isolates induce a host- and serotype-specific cytokine response.

DISCUSSIONACKNOWLEDGMENTSREFERENCES

human and animal isolates of yersinia enterocolitica show ...human and animal isolates of yersinia...

Documents