Interaction of cblA/adhesin-positive Burkholderiacepacia with squamous epithelium
Introduction
Adherence of bacteria to a specific host receptor is animportant initial step in pathogenesis. It determines tissuetropism and initiates molecular cross-talk between thehost and bacteria through mutual exchange of signals and responses (Svanborg et al., 1996). These processesdetermine the fate of the infecting organism: whether it iscleared by host natural defence mechanisms, or persistsand colonizes the host successfully. Initial adherence of bacteria to host mucosa is mediated by pili, fla-gella, lipopolysaccharides (LPS), or outer membrane proteins, and may lead to cytotoxicity, stimulation of a pro-inflammatory response, and/or tissue invasion. Forexample, pili-mediated adherence of Pseudomonasaeruginosa to airway epithelial cells has been reported tolead to bacterial invasion, cyto-toxicity (Comolli et al.,1999) and stimulation of a pro-inflammatory response viaNF-kB activation (DiMango et al., 1998). Some species ofpathogenic bacteria also target host signalling pathwaysto cause cytoskeletal rearrangements or to inhibit thenormal clearance mechanisms of host cells upon bindingand/or invasion (Mulvey et al., 1998; Nhieu et al., 2000;Steel-Mortimer et al., 2001).
Burkholderia cepacia, originally identified as a phy-topathogen, has been recognized as an important opportunistic respiratory pathogen in cystic fibrosis (CF)patients. About 4% of CF patients are culture-positive forthis organism world-wide (LiPuma, 1998), but in some CFcentres, for example, in the Toronto adult CF centre, theprevalence rate is as high as 40% (Johansen et al., 1998).Adult and adolescent CF patients are more susceptible toinfection by B. cepacia than paediatric patients, suggest-ing that chronic lung damage as a result of previous infec-tions is a major predisposing factor. Patients harbouring B.cepacia, show highly unpredictable and variable clinicalcourses, but about 30% develop fatal necrotizing pneu-monia and septicaemia (cepacia syndrome), often within afew months to a year (Tablan et al., 1985) but sometimesmuch later in the disease course (Dobbin et al., 2000).
Burkholderia cepacia, now known as the B. cepaciacomplex, has been divided into at least seven geno-movars depending on phenotypic and genetic character-istics (Coenye et al., 2001). Although strains from allgenomovars have been isolated from CF patients, themajority of strains from CF patients belong to genomovars
Cellular Microbiology (2002) 4(2), 73–86
Umadevi Sajjan,1* Cameron Ackerley2 and JanetForstner1*Departments of 1Structural Biology and Biochemistry,and 2Pediatric Lab Medicine, The Hospital for SickChildren, Toronto, Ontario, Canada.
Summary
A highly transmissible strain of Burkholderia cepaciafrom genomovar III carries the cable pilin gene,expresses the 22kDa adhesin (cblA +ve/Adh +ve),binds to cytokeratin 13 (CK13) and is invasive. CK13is expressed abundantly in the airway epithelia ofcystic fibrosis (CF) patients. We have now investi-gated whether binding of cblA +ve/Adh +ve B. cepaciato CK13 potentiates bacterial invasion and epithelialdamage using bronchial epithelial cell cultures dif-ferentiated into either squamous (CK13-enriched) ormucociliary (CK13-deficient) epithelia. Three differentB. cepacia isolates (cblA +ve/Adh +ve, cblA +ve/Adh–ve and cblA –ve/Adh –ve) showed minimal bindingto mucociliary cultures, and did not invade or causecell damage. In contrast, the cblA +ve/Adh +ve isolate,but not others, bound to CK13-expressing cells insquamous cultures, caused cytotoxicity and stimu-lated IL-8 release within 2h. By 24h, this isolateinvaded and migrated across the squamous culture,causing moderate to severe epithelial damage. A spe-cific antiadhesin antibody, which blocked the initialbinding of the cblA +ve/Adh +ve isolate to CK13, sig-nificantly inhibited all the pathologic effects. Trans-mission electron microscopy of squamous culturesincubated with the cblA +ve/Adh +ve isolate, revealedbacteria on the surface surrounded by filopodia by 2h, and within the cells in membrane-bound vesiclesby 24h. Bacteria were also observed free in the cyto-plasm, surrounded by intermediate filaments con-taining CK13. These findings suggest that binding ofB. cepacia to CK13 is an important initial event andthat it promotes bacterial invasion and epithelialdamage.
III or II (LiPuma et al., 2001). One clonal lineage fromgenomovar III, designated ET12, has been linked to thecepacia syndrome in the UK and Canada (Ledson et al.,1998; Clode et al., 2000). Isolates of this clonal lineagecarry the cblA gene (which encodes the major cable pilinsubunit) and the ‘epidemic’ DNA marker called BCESM(Sajjan et al., 1995; Sun et al., 1995; Mahenthiralingamet al., 1997). We have demonstrated previously thatselected isolates from this lineage express cable pili anda 22 kDa adhesin that mediates binding to mucin glyco-proteins, epithelial cells and lung sections of CF patients(Sajjan and Forstner, 1992; 1993; Sajjan et al., 2000a).The major receptor for the adhesin on epithelial cells wasidentified as cytokeratin 13 (CK13) (Sajjan et al., 2000b),the expression of which is increased in CF lungs (Sajjanet al., 2000a).
CF patients colonized with the ET12 strain of B. cepaciawere noted to have bacteria on pathologically thickenedalveolar septa, peribronchiolar and perivascular areas andin epithelia of terminal and respiratory bronchioles (Sajjanet al., 2001a). The ET12 strain was also shown to invadeand replicate in vitro in monolayers of non-polarized pul-monary epithelial cells and macrophages (Martin andMohr, 2000), and to persist in the lungs of CF mice andcause pneumonic consolidation (Sajjan et al., 2001b). Theunique distribution pattern of B. cepacia in the lungs of CFpatients combined with those in vitro and in vivo studiesclearly indicate that the ET12 strain of B. cepacia can beinvasive and virulent. Burkholderia cepacia other thanET12 strains were also demonstrated to invade and repli-cate in pulmonary epithelial cells, macrophages and
amoebae (Burns et al., 1996;Tipper et al., 1998; Sainiet al., 1999) suggesting that, under certain conditions,even non-ET12 strains can be invasive. However, theprocess of invasion by B. cepacia (both ET12 and non-ET12) is incompletely understood.
Pathogenesis usually depends on both the host and the infecting organism. In the present study, we sought to establish the relationship between binding of B.cepacia to CK13 and cellular damage and invasion. Ourfirst goal was to compare the initial cytotoxic and/or pro-inflammatory effects of an ET12 strain known to carry thecblA gene, and to express the 22kDa adhesin (cblA+ve/Adh +ve), with an ET12 strain that carries the cblAgene but does not express the adhesin (cblA +ve/Adh–ve) and a non-ET12 strain that does not carry either thecblA gene or express the adhesin (cblA –ve/Adh –ve). Oursecond goal was to determine if there was a correlationbetween the binding of B. cepacia to CK13 and invasive-ness of the bacteria. Third, we examined the interactionsbetween epithelial cells and B. cepacia at an ultrastruc-tural level. For these studies, we used bronchial epithe-lial cells that were differentiated into either squamouscultures expressing abundant CK13 on the surface, ormucociliary cultures deficient in surface CK13 (Gray et al.,1996; Koo et al., 1999).
Results
Morphology and CK13 expression of bronchial epithelialcell cultures grown at an air–liquid interface
Bronchial epithelial cells grown on a membrane support
Fig 1. Morphology and expression of CK13 indifferentiated bronchial epithelial cell cultures.A and B. Light microscopy of H- and E-stained sections showing mucociliary andsquamous differentiation respectively.Asterisks in A represent mucus-producingcells.C and D. Fluorescence microscopy showingreactivity of mucociliary and squamouscultures, respectively, with mAb to CK13.
Epithelial invasion by cblA +ve/Adh +ve B. cepacia 75
at an air–liquid interface in the presence or absence of retinoic acid differentiated into mucociliary or squamouscultures, respectively, as described previously by others (Gray et al., 1996; Koo et al., 1999). Mucociliarycultures showed ciliated and mucus-producing cells on the surface, one or two layers of suprabasal cells and a layer of basal cells (Fig. 1A), whereas squamouscultures showed one or two layers of flattened desqua-mating cells on the surface, two to four layers ofsuprabasal epithelial cells and a layer of basal cells(Fig. 1B). In isolated areas of squamous cultures, some cells remained undifferentiated (not shown). Thiswas confirmed by scanning electron microscopy (seelater).
Expression of CK13, a marker of squamous differenti-ation (van Muijen et al., 1986), was determined by
both immunolocalization and Western blot analysis. Byimmunolocalization, no CK13 was detected on thesurface and only very little or no CK13 in the basal orsuprabasal layers of mucociliary differentiated cultures(Fig. 1C). In contrast, abundant CK13 was detected onthe surface, and low-to-moderate amounts in suprabasaland basal layers, respectively (Fig. 1D), of squamous cultures. Western blot analyses of cytokeratin-rich frac-tions prepared from three separate squamous culturesshowed a strong band of CK13 at a molecular mass of 55kDa, as observed previously for buccal epithelial cells (Sajjan et al., 2000b). Cytokeratin extracts frommucociliary cultures on the other hand, showed no bandor only a weakly positive CK13 band at 55kDa, suggest-ing very low expression of CK13 in these cultures (notpresented).
Fig. 2. Release of LDH (A) and IL-8 (B) fromsquamous-differentiated cultures in responseto B. cepacia infection. Cultures were treatedwith B. cepacia (106 cfu) for 2h, LDH and IL-8release were determined in the combinedmedia of apical and basal compartments asdescribed in Experimental procedures.Asterisks represent statistically significantdifferences between the groups, with P-values<0.005 in each case.
Burkholderia cepacia (1 ¥ 106 cfu) was incubated for 2h with the apical surface of squamous or mucociliary differentiated bronchial epithelial cellcultures, the cultures washed with PBS to remove unbound bacteria, bound bacteria desorbed by 0.2% Triton X-100 and plated on blood agarplates. Each value represents the mean ± SEM of triplicate experiments performed on three different batches of cultures.a. Difference between binding of BC7 and BC45 or ATCC 25416 to squamous cultures was statistically significant with a P-value <0.005 in eachcase.
Burkholderia cepacia stimulates a pro-inflammatoryresponse and cause cytotoxicity in squamous cultures
Certain bacterial pathogens have been shown to stimu-late expression of pro-inflammatory cytokines and/orcause cytotoxicity in epithelial cells (Svanborg et al.,1996; Kagnoff and Eckmann, 1997; Comolli et al., 1999).To determine whether different isolates of B. cepacia stim-ulate such responses in either squamous or mucociliarydifferentiated cultures, IL-1b, TNF-a, IL-8 and lactosedehydrogenase (LDH) levels were measured in the apicaland basolateral media from cultures that were incubatedwith B. cepacia or phosphate-buffered solutions (PBS) for2h. Only squamous cultures responded to B. cepaciatreatment, specifically by releasing LDH and IL-8. LDH,which is a measure of cytotoxicity (Comolli et al., 1999),was threefold higher in BC7-treated squamous culturesthan PBS-treated controls or cultures incubated withATCC 25416 or BC45 (Fig. 2A). Among the threecytokines measured, only IL-8 levels were significantly dif-ferent. All three B. cepacia isolates stimulated IL-8 releasefrom squamous cultures, but isolate BC7 was more effec-tive than the other two (Fig. 2B). Preincubation of BC7with an antibody to the 22 kDa adhesin inhibited LDHrelease by 87.8 ± 5.1% and IL-8 release by 68.8 ± 3.1%.By 2h, none of the B. cepacia isolates stimulated a mea-surable release of IL-1b or TNF-a from squamous cul-tures. This was not surprising because these twocytokines are normally upregulated at very early stages,usually within a few minutes of bacterial addition, and theyreturn to normal by 1h (Hedges et al., 1995; Svanborget al., 1996).
Adherence of B. cepacia to CK13 correlates withcytotoxic/pro-inflammatory effects of B. cepacia insquamous cultures
As only squamous cultures released higher levels of LDHand IL-8 within 2h of incubation with BC7, and these wereinhibited by a specific antibody to the adhesin, we exami-
ned whether the observed effects occurred as a result ofB. cepacia binding to its CK13 receptor. Burkholderiacepacia [106 colony-forming units (cfu)] were incubatedwith mucociliary or squamous-differentiated cultures for 2h, unbound bacteria were removed and the bacteriabound per culture were determined by the dilution platingmethod. All three B. cepacia isolates showed minimalbinding to mucociliary cultures (2.3–2.7 ¥ 103 cfu/cellculture (Table 1). Binding of B. cepacia to squamous cul-tures was 10 to 100-fold greater depending on the B.cepacia isolate (Table 1). Isolate BC7 showed the highestbinding, being fourfold and sixfold greater than BC45 andATCC 25416, respectively, and the differences betweenthem were statistically significant (P < 0.005) in bothcases. Preincubation of BC7 with a specific antibody tothe adhesin decreased the binding of BC7 to the squa-mous cultures by 91.9 ± 3.3%.
To localize bacteria in relation to CK13, intact culturesstill attached to the membrane support were double-labelled with a mixture of antibodies to B. cepacia and toCK13, and observed under a confocal microscope, scan-ning at successive 1mm intervals from the surface. Squa-mous cultures showed large areas of CK13-positive cells(squamous-differentiated cells) on the surface with inter-mittent areas of CK13-negative cells (undifferentiatedcells), whereas mucociliary cultures lacked surface CK13.Squamous cultures incubated with isolate ATCC 25416showed only a few bacteria on the surface, and thesewere not particularly confined to CK13-positive areas(Fig. 3A and C). Cultures incubated with BC45 gave asimilar result. In contrast, cultures incubated with BC7showed abundant bacteria in association with CK13 posi-tive areas (Fig. 3B and D). Mucociliary cultures showedvery few bacteria on the surface, but the bacteria werenot in close association with cells irrespective of the B.cepacia isolate used (not shown). Neither squamous normucociliary cultures showed bacteria in the deeper layersof the cells, indicating that B. cepacia had not penetratedthe cultures by 2h.
Quantification of the number of bacteria bound toCK13-positive and CK13-negative cells revealed that
Epithelial invasion by cblA +ve/Adh +ve B. cepacia 77
squamous cultures incubated with isolate BC7 showedthe highest number of bound bacteria per cell (63 ± 1.6bacteria/10 cells), the majority (95%) of which were boundto CK13-positive cells. Cultures incubated with ATCC25416 and BC45 however, showed only 5 ± 2.8 and14 ± 2.8 bacteria/10 cells, respectively, of which >90%were bound to CK13-negative cells. In mucociliary cul-tures, bacteria were rarely observed on cells but were inthe vicinity of the cells, entangled in mucus-like material(not presented).
Burkholderia cepacia persist, proliferate and invadesquamous cell cultures
For successful colonization of host tissues, bacteria mustpersist and proliferate efficiently. To test colonizationcapacity, cell cultures incubated with B. cepacia or PBSfor 2h were washed to remove unbound bacteria, andthen further incubated for 24h. Bacterial proliferation andpersistence were quantified by dilution plating. Comparedwith the number of bacteria observed at 2h of incubation(Table 1), the environmental isolate ATCC 25416 showedno increase in bacterial counts after 24h of incubationwith either type of culture, suggesting that it can persistbut not proliferate efficiently. In contrast, both of the clini-cal isolates BC7 and BC45 proliferated efficiently on both
types of culture, with the number of bacteria increasing bytwo to three logs (Table 2) over the bacterial countsobserved at 2 h of incubation (Table 1).
To determine if bound B. cepacia had invaded the cultures after 24h of incubation, medium from the basalchamber was tested for the presence of bacteria by dilution plating. Bacteria could be recovered only in thebasal medium of squamous cultures incubated with BC7(4600 ± 120cfuml-1 of culture medium) but not in identi-cal cultures incubated with BC45 or ATCC 25416, indi-cating that only BC7 was invasive. Preincubation of BC7
Fig. 3. Binding of B. cepacia to squamouscultures. Cultures incubated with B. cepaciafor 2h were washed and labelled with amixture of antibodies to CK13 (red) and B.cepacia (green) and observed under aconfocal microscope. C and D showlocalization of B. cepacia and CK13 on thesurface of the cultures incubated with isolateATCC 25416 and BC7 respectively. A and Bshow cell morphology as observed usingNomarski differential interference contrast,corresponding to areas presented in C and Drespectively.
Table 2. Proliferation of B. cepacia on squamous or mucociliary dif-ferentiated bronchial epithelial cell cultures.
Squamous or mucociliary differentiated bronchial epithelial cell cul-tures were incubated with B. cepacia (106 cfu per well) for 2 h. Thecultures were washed to remove unbound bacteria and further incu-bated for 24h. Cultures were treated with 0.2% Triton X-100 andplated on blood agar plates. The data represent the mean ± SEM of triplicate experiments performed on three different batches of cultures.
cblA +ve/Adh +ve B. cepacia cause localizedmorphologic damage to squamous cultures
To examine the effect of binding, proliferation and inva-sion of B. cepacia, paraffin sections, prepared from squamous cultures incubated with B. cepacia, werestained with haemotoxylin and eosin and observed undera light microscope. Squamous cultures incubated with B.cepacia for a short period (2h) showed no morphologicalchanges and resembled PBS-treated control cultures (notpresented). Cultures incubated with bound bacteria for24 h, however, showed variable morphological changesdepending upon the B. cepacia isolates used. BC7-treated cultures showed severe epithelial damage insome areas, with widespread disruption of cells (Fig. 5A),whereas isolates BC45 and ATCC 25416 caused mild orno damage respectively (not presented). Immunofluores-cence microscopy with anti-B. cepacia antibody, revealedabundant bacteria in both BC45- and BC7-treated, but notin ATCC-treated squamous cultures. BC7-treated culturesshowed bacteria on the surface as well as in the basaland suprabasal layers. Damage was severe in areas witha high bacterial density, and also where bacteria hadinvaded the full thickness of the culture (Fig. 5B). BC45-treated cultures showed bacteria mostly on the surfaceand only mild cellular damage was evident in isolated
Fig. 4. Invasion of squamous culture by B.cepacia isolate BC7. Cultures were incubatedwith BC7 (106 cfu) for 2 h, washed to removeunbound bacteria and further incubated for 24 h. Bacteria (green) and CK13 (red) weredetected using specific antibodies asdescribed in Experimental procedures, andthe cultures were scanned in two channelssimultaneously under a confocal microscope,taking optical sections every micron. Thedepth of each section is given on the leftupper corner of each panel. Full thickness ofthe culture was 29 m.
with an antibody to the 22kDa adhesin inhibited inva-sion of the squamous cultures by 98.4% ± 2.7%. Whensimilar experiments were carried out using mucociliary-differentiated cultures, no bacteria were recovered in thebasal medium irrespective of the B. cepacia isolate used,suggesting the resistance of these cultures to bacterialinvasion.
To investigate whether BC7 migration through the squa-mous cultures occurs mainly in the CK13-enriched areas,cell cultures with bound BC7 were further incubated for 24h as described above, then double-labelled with a mixture of antibodies to CK13 and B. cepacia andobserved under a confocal microscope, taking opticalsections at successive intervals of 1mm. Bacteria (green)were observed on the surface as well as in the deepercell layers of the culture and were usually associated withCK13-enriched (red) areas (Fig. 4). The depth of bacter-ial migration was measured by scanning three or morefields in three different cultures. Bacteria could be seenas deep as 28 and 35m, which represented full thicknessof various squamous cultures. Cultures incubated withUV-killed B. cepacia isolate BC7 for 24h, however,showed bacteria only on the surface in close associationwith CK13-positive areas, but not in the deeper layers(data not shown), suggesting that bacteria must be meta-bolically active to be invasive.
Epithelial invasion by cblA +ve/Adh +ve B. cepacia 79
areas. Morphology of mucociliary cultures was notaffected by any of the B. cepacia isolates, even after 24h of incubation, and immunolocalization revealed bac-teria only in the mucus-like secretions on the surface ofcultures. These results suggested that only cblA +ve/Adh+ve B. cepacia isolate BC7 caused localized severedamage, which correlated with bacterial density and invasion.
Interaction of B. cepacia with squamous cultures at anultrastructure level
Because isolate BC7 invaded and caused more cellulardamage than the other isolates, squamous culturestreated with isolate BC7 were used to characterize the
nature of specific pathogenic interactions. Scanning electron microscopy of cultures treated with BC7 for 2h(Fig. 6A), showed bacteria mostly on squamous differen-tiated areas that were characterized by flat cells withshallow ridges, and rarely on undifferentiated cells. Pili-like appendages were often noted, extending from bacte-rial surfaces connecting one bacterium to another and to the cell surface (Fig. 6B). In some areas, bacteriaappeared to be penetrating epithelial cells, promoting theformation of a fibrous network around it. In cultures incu-bated with bound BC7 for 24h, numerous bacterial micro-colonies were observed both on the surface (Fig. 6C),especially in areas of cellular damage as represented by detached and desquamated cells, and in crevicesbetween the desquamating cells. Often bacteria were
Fig. 5. Histology and immunolocalization of B.cepacia in squamous cultures. Cultures wereincubated with BC7 (106 cfu) for 2 h, washedto remove unbound bacteria and furtherincubated for 24 h, washed, fixed in 10%buffered formalin and embedded in paraffin.A. H- and E-stained section showing severedamage to epithelial cells.B. Localization of B. cepacia by anti-B.cepacia antibody (white).
Fig. 6. Scanning and transmission electronmicroscopy of squamous cultures treated withB. cepacia isolate BC7. Squamous cultureswere incubated with BC7 for 2 or 24 h asdescribed in Experimental procedures.A. Squamous cultures incubated for 2 h withB. cepacia, showing bacterial binding (arrows)to regions of squamous cells (within theadded white boundary).B. Higher magnification of a section of A (pili-like extensions connecting one bacterium toanother are represented by single arrows andbacteria to host cell by double arrows;arrowheads represent bacteria penetrating theepithelial cell).C. Squamous cultures incubated with BC7 for24 h, showing desquamating cells withnumerous bacterial colonies (arrows).Asterisks represent crevices between cells.D. Bacteria in one of the crevices entrappedin a fine fibrous network and appearing topenetrate a cell.
entangled in an extended fibrous network and positionedat right angles to the surface of a cell, as though pene-trating the membrane (Fig. 6D). It is concluded thatbinding of cblA +ve/Adh +ve B. cepacia to CK13-enrichedsquamous cells facilitates cellular invasion.
Transmission electron microscopy of squamous cul-tures was carried out to observe the phenomenon of inva-sion in greater detail. After incubation for 2h, bacteriawere observed on apical cell surfaces causing membraneinvaginations (Fig. 7A). Some bacteria that were enteringthe cells were almost completely enveloped by filopodia(Fig. 7B). Using immunoelectron microscopy with anti-actin antibody, enrichment of actin was observed in thefilopodia (not presented), suggesting that penetrating bac-teria cause rearrangement of the epithelial cell cytoskele-ton to accommodate their entry. After 24h of incubation(Fig. 8), bacteria were observed at different stages of
invasion. Areas of heavy colonization showed completedestruction of cells with a loss of recognizable cellorganelles (data not shown). Mild-to-moderately-affectedareas showed colonized bacteria in the apical andsuprabasal cell layers that were structurally still intact(Fig. 8A). Bacteria inside the cells were in membrane-bound vacuoles (Fig. 8B) and often more than one bac-terium was observed in the same vacuole, suggestingintracellular replication. Sometimes bacteria were alsoobserved free in the cytoplasm, and some of theseappeared to be pushing through the basolateral mem-brane (Fig. 8C) and exiting from the cell. An adjacent cellbelow shows incipient membrane invagination, as thoughready to accommodate the exiting bacteria. In Fig. 8D,both entry at the apical surface and exit of a bacteriumfrom the basal surface of the same epithelial cell is shown,indicating that bacteria migrate by transcytosis. The bac-
Fig. 7. Transmission electron microscopy ofsquamous cultures incubated with B. cepaciaBC7 for 2 h.A. Close association of bacteria with thesurface of a squamous cell.B. Invading bacteria surrounded by filopodiaof epithelial cells (arrow).
Fig. 8. Transmission electron microscopy ofsquamous cultures incubated with B. cepaciaBC7 for 24 h. Cultures were incubated withBC7 for 2 h, washed and further incubated for24 h.A. Bacteria in the apical and suprabasal cellsof a culture.B. Membrane-bound intracellular bacteria.C. Cytoplasmic bacteria exiting from a cell(arrows) and membrane invagination on thecell below (arrowheads).D. A bacterium entering at the apical surface,intracellular bacteria in vacuoles and abacterium exiting from the basal side of cell,which is partially covered by basolateralmembrane (arrow) and membraneinvagination in the cell below (arrowhead).
Epithelial invasion by cblA +ve/Adh +ve B. cepacia 81
terium in the intercellular space is partially surrounded bybasolateral membrane of the cell from which the bac-terium had probably exited. The cell below shows mem-brane invagination to accommodate the exiting bacteria.There seems little doubt, therefore, that isolate BC7 iscapable of invading epithelial cells, replicating within themand moving from one cell to another. This probablyexplains how cblA +ve/Adh +ve B. cepacia crossed thewhole thickness of squamous cultures.
One of the most interesting phenomena that wereobserved in this study was that intracellular bacteria, freein the cytoplasm, were surrounded by bundles of fila-ments (Fig. 9A), suggesting that these interactions mayplay a role in the movement of B. cepacia within the cells.These filaments were not enriched in actin, but reactedwith a multicytokeratin antibody (not presented) and withthe specific monoclonal antibody to CK13 (Fig. 9B).Damaged epithelial cells containing a high density ofintracellular bacteria showed many of these filamentsaround the bacteria (Fig. 9C), implying that intracellularbacteria adhere to and disrupt the intermediate filamentnetwork.
Discussion
In the present study, we compared the adherence-mediated pathologic effects and invasive properties ofcblA +ve/Adh +ve, cblA +ve/Adh –ve and cblA –ve/Adh–ve B. cepacia isolates utilizing well differentiatedbronchial epithelial cell cultures as a model system. ThecblA +ve/Adh +ve B. cepacia isolate BC7, which bound to
its CK13 receptor, caused cytotoxicity, stimulated exces-sive release of IL-8 and invaded the CK13-enrichedepithelium efficiently within 24h. These effects were inhibited by an antibody to the 22kDa adhesin. On theother hand cblA +ve/Adh –ve(BC45) and cblA –ve /Adh–ve(ATCC 25416) isolates, which did not bind to CK13,neither caused cytotoxicity nor invaded the epithelium,although both isolates moderately stimulated IL-8 release.Squamous differentiated cultures, which express abun-dant CK13 on the surface, were thus vulnerable to bindingand invasion by the cblA +ve/Adh +ve B. cepacia isolateBC7. Mucociliary differentiated cultures, which lack sur-face CK13, were resistant to these adherence-mediatedeffects and invasion by all three B. cepacia isolates.
Both B. cepacia isolates BC7 and BC45 are clinical isolates. They are identical using pulsed-field gel elec-trophoretic analysis, both belong to Genomovar III, carrythe epidemiological marker designated BCESM and thecblA gene. Both isolates proliferated on mucociliary aswell as squamous differentiated cultures, indicating thatthey have adapted to a eukaryotic environment and haveacquired resistance to antimicrobial factors produced bybronchial epithelial cells (Baird et al., 1999). However,when the adherence and pathogenic potential of theseisolates were compared, BC7 was found to be more effec-tive than BC45 in all respects. One of the major pheno-typic differences between these two isolates is their abilityto bind to the CK13 receptor, a process that is mediatedby the 22 kDa adhesin protein (Sajjan and Forstner, 1992;Sajjan et al., 2000a). The adhesin is expressed by BC7,and immunoblocking of the adhesin significantly reduced
Fig. 9. Immunolocalization of CK13 insquamous cultures incubated with BC7 for 24h. Cultures were treated as described inFig. 8.A. Intracellular bacteria showing aggregationof intermediate filaments (arrowheads) aroundbacteria.B. CK13 in the aggregated intermediatefilaments (arrows).C. Enrichment of CK13 in the filamentssurrounding the bacteria in the damaged cells(arrows).
the pathogenic potential of isolate BC7. The environmen-tal isolate ATCC 25416 belongs to genomovar I, is cblA–ve/Adh –ve and did not bind to CK13 or cause cytotox-icity. Thus, the initial binding of B. cepacia to CK13 seemsto play an important role in pathogenesis. ATCC 25416also failed to proliferate on bronchial cell cultures, sug-gesting that in addition to absence of the adhesin, it lacksthe ability to utilize nutrients from eukaryotic cells and/ormay be susceptible to antibacterial activity of bronchialepithelial cells.
Several cytotoxic and pro-inflammatory factors that arecommon to all species of the B. cepacia complex, irre-spective of their source, have been recognized by otherinvestigators (Palfreyman et al., 1997; Hutchison et al.,2000; Melnikow et al., 2000). A cytotoxic effect onmacrophages and mast cells is linked to bacterialenzymes that generate phosphorylated and non-phosphorylated adenine nucleosides. These, in turn, acti-vate purinergic receptors and promote cell death (Mel-nikow et al., 2000). As bronchial epithelial cells alsoexpress purinergic receptors on their surface (Laubingeret al., 2001), it is possible that the same enzymes wouldcause similar cytotoxic effects on these cells. LPS, andan uncharacterized exoproduct other than LPS, havebeen identified as important pro-inflammatory factors, butonly the latter was demonstrated to stimulate pro-inflam-matory cytokines from undifferentiated lung epithelial cells(Palfreyman et al., 1997; Hutchison et al., 2000). Howeverthe activity of these cytotoxic or pro-inflammatory factorson well differentiated bronchial epithelial cell cultures hasnot yet been investigated. As only the cblA +ve/Adh +veisolate BC7 caused cytotoxicity and excessive stimulationof IL-8, and only in CK13-enriched squamous differenti-ated cultures, it is presumed that initial binding of B.cepacia to CK13 may be necessary to trigger signal trans-duction events leading to LDH and IL-8 release. Alterna-tively, binding may be necessary to promote close and/orlong enough contact between the bacteria and the cellsto stimulate the secretion and/or facilitate the delivery of cytotoxic or pro-inflammatory factors to the plasmamembrane.
The initial binding of B. cepacia to its CK13 receptoralso appears to be necessary for subsequent cellularinternalization of the bacteria. Isolate BC7 invaded theCK13-enriched squamous cells and even while migratingacross the depth of the squamous cultures, BC7 wasalmost always found in association with CK13-positivecells, as observed by confocal microscopy. The cblA+ve/Adh –ve and cblA –ve/Adh –ve isolates did not invadeeither type of culture. By contrast, other investigatorshave demonstrated that all species of B. cepacia complexcan invade pulmonary epithelial cells (Burns et al., 1996;Tipper et al., 1998; Martin and Mohr, 2000). The appar-ent discrepancy between our results is probably as a
result of their use of non-polarized epithelial cells in whichtight junctions do not form and, thus, basolaterally locatedglycolipid receptors are readily accessible for binding.Previously, it has been shown that all three strains of B.cepacia can bind to glycolipids, including asialo GM1(Sylvester et al., 1996). We assume that these receptorsare not accessible to B. cepacia in well differentiatedbronchial epithelial cell cultures with tight junctions.Support for this interpretation comes from studies in whichpolarized airway epithelial cell cultures, which have noasialo GM1 on the surface, were shown to be resistant toP. aeruginosa binding and invasion. Compromising thetight junctions by EGTA treatment, however, made the cul-tures susceptible (Fleiszig et al., 1997). In the presentstudy, mucociliary cultures, in which neither galactolipidsnor CK13 are expressed on the surface, were resistant toinvasion by all three isolates tested.
For invasion of well differentiated squamous cultures,viability of bacteria was required, as UV-killed BC7 did notinvade. This suggests that the invasion process requiresboth adherent and metabolically active bacteria, presum-ably to synthesize, secrete and/or translocate effectormolecules into host cells, as described for other invasiveGram-negative bacterial pathogens such as Salmonellasubsp. and Shigella subsp. (Galan, 1999; Nhieu et al.,2000). In these bacterial species, type III secretionsystems serve to deliver bacterial effector proteins into theplasma membrane and cytoplasm of host cells, whichthen triggers a series of events leading to remodelling ofthe plasma membrane to accommodate the entry of theinvasive pathogens. The entire process generally followsstrong bacterial adherence to host epithelial cells. Asimilar delivery system may mediate the invasion of squa-mous epithelial cells by B. cepacia because a putativetype III secretion gene cluster has been shown to existwidely in the B. cepacia complex, except for genomovarI strains (Parsons et al., 2001).
The initial entry of squamous epithelial cells by the cblA+ve/Adh +ve BC7 resembles the entry process describedfor Salmonella subsp., Streptococcus pyogens andShigella subsp. (Steele-Mortimer et al., 1999; Molinariet al., 2000; Nhieu et al., 2000). At the site of entry, B.cepacia caused membrane ruffling and formation of actin-enriched filopodia, suggesting that actin rearrangementoccurs during the initial stages of invasion. At later stages,unlike other invasive bacteria that promote actin tail for-mation to propel themselves and to migrate from one cellto another (Steel-Mortimer et al., 2001), cblA +ve/Adh +veB. cepacia appeared to migrate by causing rearrange-ment and/or disruption of the intermediate filamentnetwork. Intermediate filaments are a family of structurallyheterologous and functionally similar proteins, whichinclude cytokeratins, vimentin and desmin. They organizeinto a complex supramolecular network extending from
Epithelial invasion by cblA +ve/Adh +ve B. cepacia 83
the nucleus to the peripheral-most portion of the cell(Lazarides, 1982). Disruption of such a complex networkby pathogens might be expected, therefore, to alter thefunctional organization of the plasma membrane and/orcytoplasm, thus affecting various signal transductionevents and cell growth. Some viral proteins and bacterialtoxins have been shown to cause disruption by interferingwith phosphorylation of intermediate filaments or deple-tion of regulatory proteins that are involved in the organi-zation of intermediate filaments (Sharpe et al., 1980;Coulombe, 1993). This is a promising lead to explore infuture research of B. cepacia pathogenesis.
The present study demonstrates that cblA +ve/Adh +veB. cepacia cause more cell damage than cblA +ve/Adh-ve or cblA-ve/Adh-ve isolates, especially in CK13-enriched squamous epithelia. In normal individuals, thedegree of squamous metaplasia in the lung is very low,but in cystic fibrosis patients, squamous metaplasia isobserved frequently in both large and small airways,probably as a result of repeated infection-related injuryand repair (Simel et al., 1984). These areas expressabundant CK13 and are probably preferred targets for thebinding of cblA +ve/Adh +ve B. cepacia (Sajjan et al.,2000a). Thus CF airways provide a suitable environmen-tal niche for adherence and colonization by the highlytransmissible cblA +ve/Adh +ve strain of B. cepacia.Based on the present study, we propose that after initialbinding, B. cepacia proliferate on squamous epithelia,invade and migrate across the epithelial barrier, causingfurther epithelial damage and, finally, colonize the deeperparenchyma, especially in peribronchiolar and perivascu-lar regions, in which B. cepacia was noted in earlierstudies (Sajjan et al., 2001a; 2001b). During this process,B. cepacia stimulate excessive release of IL-8 by hostcells (analogous to KC release observed in CF knockoutmice (Sajjan et al., 2001b), which, in turn, recruits neu-trophils to the site of infection. This may lead to severeinflammation, abscess formation and septicaemia. Thus,the ability of cblA +ve/Adh +ve B. cepacia to bind to CK13seems to correlate highly with its potential to invade andcause lung damage in CF patients. These findings con-tribute new insights into the mechanisms by which cblA+ve/Adh +ve B. cepacia can provoke pathologic changesin airway epithelia.
Experimental procedures
Bacteria and growth conditions
Burkholderia cepacia isolates BC7 and BC45 were isolated fromthe sputa of cystic fibrosis patients attending the Toronto CFclinic. Both isolates belong to genomovar III, are identical bypulsed-field gel electrophoresis (PFGE) and are cblA +ve. BC7is a Adh +ve isolate and binds to CK13, whereas BC45 is Adh–ve and does not bind to CK13 (Sajjan and Forstner, 1993;
Sajjan et al. 2000a; Sun et al., 1995). An environmental typestrain ATCC 25416 was purchased from the American TissueCulture Collection, Bethesda, MD. This strain belongs togenomovar I, is cblA –ve/Adh –ve and does not bind to CK13.All bacterial stock cultures were stored at -80∞C. For binding andinvasion assays, B. cepacia was subcultured on brain–heart infu-sion (BHI) agar plates, a single colony was inoculated into 10mlof tryptic soy broth (Difco) and grown for 18h on an orbital shakerat 37∞C. Bacteria were harvested by centrifugation, washed threetimes with phosphate-buffered solution (PBS) and suspended inPBS to a concentration of 1 ¥ 108 cfuml-1.
Antibodies
Monoclonal antibodies to CK13 and multicytokeratins were pur-chased from Vector diagnostic Laboratories. Anti-actin antibodywas from Becton Dickinson Canada. A polyclonal rabbit antibodyto B. cepacia (R418) and an antibody to the 22kDa adhesinprotein have been described previously (Sylvester et al., 1996;Sajjan et al., 2000a).
Culture conditions for human bronchial epithelial cells atan air–liquid interface
Primary cultures of normal bronchial epithelial cells at passageone and bronchial epithelial cell growth medium (BEGM) werepurchased from Clonetics. Dulbecco’s minimal essential mediumwith high glucose (DMEM-H) was purchased from Life Tech-nologies. Cells were thawed and seeded into T75 tissue cultureflasks at a seeding density of 2500 cells per cm2. Cells weregrown in BEGM at 37∞C in an atmosphere of 95% air/5% CO2
until they reached 60–80% confluence. The cells were dissoci-ated using trypsin/EDTA, and cryopreserved as cells at passagetwo according to the supplier’s instructions.
Air–liquid interface culture of normal human bronchial epithelial (NHBE) cells to obtain squamous or mucociliary dif-ferentiated cultures was carried out as described by Gray andcolleagues (Gray et al., 1996), with some modifications. Briefly, 25000 cells from passage two were seeded into 12mmtranswell-clear inserts (Costar) coated with rat tail collagen (10mg/cm2) (Cohesion Technologies). Cells were grown undersubmerged conditions in a 1:1 mixture of BEGM:DMEM-H con-taining all the components of BEGM plus 10-8 M retinoic acid and13ml of bovine pituitary extract (equivalent to 65 mg of totalprotein). The medium was changed after 24h and then everysecond day until the cells were 90% confluent (5–7 days). Anair–liquid interface was created by removing the medium fromthe apical chamber and feeding the cultures basolaterally toobtain mucociliary differentiation. Cultures were fed with retinoicacid-depleted BEGM:DMEM-H medium under identical condi-tions to obtain squamous differentiation.
Cytotoxicity assay and determination of pro-inflammatory cytokines
On the day of the assay, media in the basal chamber waschanged and the apical surface of air–liquid interface cultureswas rinsed with PBS pH 7.2. B. cepacia [106 colony-forming units(cfu)] suspended in 50ml PBS, or PBS alone, was added to squa-mous or mucociliary differentiated cultures on the apical surface
and incubated for 2h at 37∞C, in 95% air/5% CO2. Media fromthe apical and basal chambers were combined, immediatelymixed with 1% BSA to prevent proteolytic degradation of proteinsreleased from the cultures and centrifuged to remove bacteriaand cell debri. The supernatant was either assayed immediatelyfor LDH activity or stored at -20∞C for determination of pro-inflammatory cytokines. LDH present in the supernatant (50ml)was analysed by the non-radioactive cytotoxicity kit® fromPromega, following the supplier’s instructions.
Pro-inflammatory cytokines IL-1b, TNF-a and IL-8 in the super-natant were determined by enzyme-linked immunoabsorbanceassay (ELISA) (R and D Systems) according to the manufac-turer’s instructions.
Bacterial binding and proliferation
Burkholderia cepacia (106 cfu) were incubated with squamous ormucociliary differentiated cultures for 2h as described above.Control cultures were incubated with PBS alone under identicalconditions. The apical surface of the culture was washed gentlysix times with PBS to remove unbound bacteria, and the cultureswere treated with 0.2% Triton X-100 in water, serially diluted inPBS and plated on either blood agar or B. cepacia isolation agar(Henry et al., 1997) to determine the number of bacteria perculture. To determine the bacteria bound to individual cells withinthe culture, cultures were dissociated with 0.05% Trypsin/1mMEDTA to give single cell suspensions, and the total number ofcells was determined by haemocytometer. An aliquot containing500 cells was cytocentrifuged, fixed in 2% paraformaldehyde anddouble-labelled with a mixture of antibodies to cytokeratin 13 andB. cepacia. The number of bound bacteria per CK13-positive and CK13-negative cells was counted in 10 microscopic fieldsand averaged.
In some experiments, after removing unbound bacteria, cul-tures were further incubated for 24 h at 37∞C, in 95% air/5% CO2,and the number of bacteria associated with the cultures was thendetermined as described above.
Inhibition assay
When antibody to the adhesin protein was used as an inhibitor,B. cepacia isolate BC7 was preincubated with antibody (1:50diluted) or with normal mouse serum for 1h, the mixture thenincubated with squamous cultures and the experiment continuedas described above.
SDS–PAGE and Western blot analysis
A cytokeratin-rich fraction was isolated from cell cultures, and theamount of CK13 was determined by Western blot analysis asdescribed previously (Sajjan et al., 2000b).
Morphological evaluation of cultures
Cultures were fixed in 10% buffered formalin overnight at 4∞C,and embedded in agar-paraffin (Bourgeois et al., 1982). Sections(5m thick) were stained with haematoxylin/eosin and examinedby light microscopy (Leica Dialux 22).
Immunological detection of CK13 and B. cepacia
CK13 and B. cepacia present in paraffin sections of cultures weredetected by double-labelling as described previously (Sajjanet al., 2000a; 2000b).
When intact cultures were to be analysed, cultures that werestill attached to the membrane support were fixed in 4% para-formaldehyde, permeabilized by incubating in 0.2% Triton X-100,washed and equilibrated in Tris-buffered saline (TBS) pH 7.8.Cultures were blocked with 5% normal donkey serum for 2h atRT and then incubated in a mixture of mouse monoclonal anti-body to CK13 (diluted 1:100) and rabbit polyclonal antibody to B. cepacia (diluted 1:800) for 1h at room temperature. The cul-tures were washed with TBS to remove unbound antibody and the bound antibodies were detected with a mixture of CY5-conjugated anti-rabbit IgG (for B. cepacia) and CY3-conjugatedanti-mouse IgG (for CK13). Cultures were observed under aZeiss Laser scanning microscope 510 with an inverted axivert100 and equipped with HeNe 1 and HeNe 2 lasers, and scannedsimultaneously in three channels (blue for cell morphology, redfor cytokeratin 13 and green for bacteria). Optical images werethen superimposed to determine the location of bacteria in rela-tion to CK13 on cells. Optical sections (at 1m intervals) of threeor more fields were taken to evaluate the depth of bacterial invasion.
Scanning electron microscopy
Cultures were fixed in universal fixative (mixture of 1% glu-taraldehye/4% formaldehyde in 0.1M phosphate buffer) for 1h,washed with 0.1M phosphate buffer (pH 7.4), and incubated in1% osmium tetroxide in PBS for 1h. Cultures were washed withphosphate buffer followed by water, dehydrated in a gradedalcohol series, critical point-dried and mounted on carbon-coatedstubs. The samples were then sputter-coated with gold andobserved under a JEOL JSM 820 scanning electron microscope(JEOL USA).
Transmission electron microscopy
Cultures were fixed for 1h in universal fixative, washed, fixed in1% osmium tetroxide for 1h and dehydrated in a graded ethanolseries. Cultures were then infiltrated and embedded in eponresin. Thin (70–80nm) sections were cut, mounted on coppergrids, stained with 5% uranyl acetate/2.6% lead citrate andobserved under a JEOL 1200 EXII transmission electron micro-scope (JOEL USA).
Immunoelectron microscopy
Cultures were fixed in 4% paraformaldehyde containing 0.1%glutaraldehyde in 0.1M phosphate buffer, washed in buffer,infused with 2.3M sucrose, frozen in liquid nitrogen and substi-tuted in methanol containing uranyl acetate at -85∞C. They werethen infiltrated and embedded in Lowicryl IHM20 and cold-polymerized at -20∞C under UV light. Thin (70–80 nm) sectionswere cut and mounted on formavar-coated nickel grids, blockedwith PBS containing 0.15% glycerol and 2% BSA, and then incu-bated with 1:100 diluted anti-CK13 antibody, undiluted multicy-tokeratin antibody or 1:10 diluted antiactin antibody for 1h at RT.
Epithelial invasion by cblA +ve/Adh +ve B. cepacia 85
Sections were rinsed with PBS containing BSA to removeunbound antibody and then incubated for 1h at RT with anti-mouse IgG conjugated with 10nm colloidal gold (1:20 diluted).Sections were rinsed with distilled water, counterstained withuranyl acetate and lead acetate and observed under a JEOL1200 EXII transmission electron microscope (JOEL USA).
Acknowledgements
We thank Drs B. Fisher and J.A. Voynow, Duke UniversityMedical Centre, Durham, NC, for helpful suggestions to obtainwell differentiated bronchial cell cultures, and Aina Tilups andYewMeng Heng, Department of Pediatric Laboratory Medicine,the Hospital for Sick Children, for help in processing samples forelectron microscopy. Financial support was received from theCanadian Cystic Fibrosis Foundation.
References
Baird, R.M., Brown, H., Smith, A.W., and Watson, M.L. (1999)Burkholderia cepacia is resistant to the antimicrobial activity ofairway epithelial cells. Immunopharmacology 44: 267–272.
Bourgeois, N., Braspenninckx, E., Wijnen, M., and Buyssens, N.(1982) The agar-paraffin embedding technique applied to smalldiagnostic biopsies. Stain Technol 57: 251–254.
Burns, J.L., Jonas, M., Chi, E.Y., Clark, D.K., Berger, A., and Griffith, A. (1996) Invasion of respiratory epithelial cells byBurkholderia (Pseudomonas) Cepacia. Infect Immun 64:4054–4059.
Clode, F.E., Kaufmann, M.E., Malnick, H., and Pitt, T.L. (2000)Distribution of genes encoding putative transmissibility factorsamong epidemic and non-epidemic strains of Burkholderiacepacia from cystic fibrosis patients in the United Kingdom. J Clin Microbiol 38: 1763–1766.
Coenye, T., Vandamme, P., Govan, J.R.W., and LiPuma, J.J.(2001) Taxonomy and identification of the Burkholderiacepacia complex. J Clin Microbiol 39: 3427–3436.
Comolli, J.C., Waite, L.L., Mostov, K.E., and Engle, J.N. (1999)Pili binding to asialo-GM1 on epithelial cells can mediate cyto-toxicity or bacterial internalization by Pseudomonas aerugi-nosa. Infect Immun 67: 3207–3214.
Coulombe, P.A. (1993) The cellular and molecular biology of keratins: beginning a new era. Curr Opin Cell Biol 5: 17–29.
DiMango, E., Ratner, A.J.R.B., Tabibi, S., and Prince, A. (1998)Activation of NF-kB by adherent Pseudomonas aeruginosain normal and cystic fibrosis respiratory epithelial cells. J ClinInvest 101: 2598–2605.
Dobbin, C.J., Soni, R., Jelihovsky, T., and Bye, P.T.P. (2000)Cepacia syndrome occurring following prolonged colonizationwith Burkholderia cepacia. Aust NZ J Med 30: 288–289.
Fleiszig, S.M.J., Evans, D.J., Do, N., Vallas, V., Shin, S., andMostov, K.E. (1997) Epithelial cell polarity affects susceptibil-ity to Pseudomonas aeruginosa invasion and cytotoxicity.Infect Immun 65: 2861–2867.
Galan, J.E. (1999) Interaction of Salmonella with host cellsthrough the centisome 63 type III secretion system. Curr OpinMicrobiol 2: 46–50.
Gray, T.E., Guzman, K.C., Davis, W., Abdullah, L.H., andNettesheim, P. (1996) Mucociliary differentiation of seriallypassaged normal human tracheobronchial epithelial cells. AmJ Respir Cell Mol Biol 14: 102–112.
Hedges, S., Agace, W., and Svanborg, C. (1995) Epithelial
cytokine responses and mucosal cytokine networks. TrendsMicrobiol 3: 266–270.
Henry, D.A., Campbell, M.E., LiPuma, J.J., and Speert, D.P.(1997) Identification of Burkholderia cepacia isolates frompatients with cystic fibrosis and use of a simple new selectivemedium. J Clin Microbiol 35: 614–619.
Johansen, H.K., Kovesi, T.A., Koch, C., Corey, M., Hoiby, N., andLevison, H. (1998) Pseudomonas aeruginosa and Burkholde-ria cepacia infection in cystic fibrosis patients treated in Torontoand Copenhagen. Pediatr Pulmonol 26: 89–96.
Kagnoff, M.F., and Eckmann, L. (1997) Epithelial cells as sensorsfor microbial infection. J Clin Invest 100: S51–S55.
Koo, J.S., Yoon, J.H., Gray, T., Norford, D., Jetten, A.M., andNettesheim, P. (1999) Restoration of the mucous phenotypeby retinoic acid in retinoid-deficient human bronchial cell cul-tures: changes in mucin gene expression. Am J Respir CellMol Biol 20: 43–52.
Laubinger, W., Streubel, G., and Reiser, G. (2001) Physiologicalevidence for a P2Y receptor responsive to diadenosinepolyphosphates in human lung via Ca(2+) release studies in bronchial epithelial cells. Biochem Pharmacol 61: 623–629.
Lazarides, E. (1982) Intermediate filaments: a chemically het-erogeneous, developmentally regulated class of proteins. AnnRev Biochem 51: 219–250.
LiPuma, J.J. (1998) Burkholderia cepacia. Management issuesand new insights. Clin Chest Med 19: 473–486.
LiPuma, J.J., Spilker, T., Gill, L.H., Campbell, P.W.R., Liu, L., andMahenthiralingam, E. (2001) Disproportionate distribution ofBurkholderia cepacia complex species and transmissibilitymarkers in cystic fibrosis. Am J Respir Crit Care Med 163:92–96.
Mahenthiralingam, E., Simpson, D.A., and Speert, D.P. (1997)Identification and characterization of a novel DNA markerassociated with epidemic Burkholderia cepacia strains recov-ered from patients with cystic fibrosis. J Clin Microbiol 35:808–816.
Martin, D.W., and Mohr, C.D. (2000) Invasion and intracellularsurvival of Burkholderia cepacia. Infect Immun 68: 24–29.
Melnikow, A., Zaborina, O., Dhiman, N., Prabhakara, B.S.,Chakrabarty, A.M., and Hendrickson, W. (2000) Clinical andenvironmental isolates of Burkholderia cepacia exhibit differ-ential cytotoxicity towards macrophages and mast cells. 36:1481–1493.
Molinari, G., Rohde, M., Ruzman, G.A., and Chhatwal, G.S.(2000) Two distinct pathways for the invasion of Streptococcuspyogenes in non-phagocytic cells. Cell Microbiol 2: 145–154.
van Muijen, G.N.P., Ruiter, D.J., Franke, W.W., Achtstatter, T.,Haasnoot, W.H.B., Ponec, M., and Warnaar, S.O. (1986) Celltype heterogeneity of cytokeratin expression in complexepithelia and carcinomas as demonstrated by monoclonal antibodies specific for cytokeratins nos. 4 and 13. Cell162: 97–113.
Mulvey, M.A., Lopez-Boado, Y.S., Wilson, C.L., Roth, R., Parks,W.C., Heuser, J., and Hultegren, S.J. (1998) Induction and
evasion of host defenses by type I-piliated uropathogenicEscherichia coli. Science 282: 1494–1497.
Nhieu, G., Bourdet-Sicard, R., Dumenil, G., Blocker, A., and San-sonetti, P.J. (2000) Bacterial signals and cell responses duringShigella entry into epithelial cells. Cell Microbiol 2: 19–33.
Palfreyman, R.W., Watson, M.L., Eden, C., and Smith, A.W.(1997) Induction of biologically active interleukin 8 from lungepithelial cells by Burkholderia (Pseudomonas) cepaciaproducts. Infect Immun 65: 617–622.
Parsons, Y.N., Glendinning, K.J., Thornton, V., Hales, B.A., Hart, C.A., and Winstanley, C. (2001) A putative type III se-cretion gene cluster is widely distributed in the Burkholderiacepacia complex but absent from genomovar I. FEMS Microbiol Letts 203: 103–108.
Saini, L.S., Galsworthy, S.B., John, M.A., and Valvano, M.A.(1999) Intracellular survival of Burkholderia cepacia complexisolates in the presence of macrophage cell activation. Micro-biol 145: 3465–3475.
Sajjan, U.S., and Forstner, J.F. (1992) Identification of the mucin-binding adhesin of isolated Pseudomonas cepacia frompatients with cystic fibrosis. Infect Immun 60: 1434–1440.
Sajjan, U.S., and Forstner, J.F. (1993) Role of a 22-kilodalton pilin protein in binding of Pseudomonas cepacia to buccalepithelial cells. Infect Immun 61: 3157–3163.
Sajjan, U.S., Sun, L., Goldstein, R., and Forstner, J.F. (1995)Cable (Cbl) type II pili of cystic fibrosis-associated Burkholde-ria (Pseudomonas) cepacia: nucleotide sequence of the cblAmajor subunit pilin gene and novel morphology of the assem-bled appendage fibers. J Bacteriol 177: 1030–1038.
Sajjan, U.S., Wu, Y., Kent, G., and Forstner, J. (2000a) Prefer-ential adherence of cable-piliated Burkholderia cepacia to respiratory epithelia of CF knockout mice and human CF lungexplants. J Med Micro 49: 875–885.
Sajjan, U.S., Sylvester, F.A., and Forstner, J. (2000b) Cable-piliated Burkholderia cepacia bind to cytokeratin 13 of epithe-lial cells. Infect Immun 68: 1787–1795.
Sajjan, U., Corey, M., Humar, A., Tullis, E., Cutz, E., Ackerley, C.,and Forstner, J. (2001a) Immunolocalization of Burkholderiacepacia in the lungs of cystic fibrosis patients. J Med Micro-biol 50: 535–546.
Sajjan, U.S., Thanassoulis, G., Cherapanov, V., Lu, A., Sjolan,C., Steer, B., et al. (2001b) Enhanced susceptibility to pul-monary infection with Burkholderia cepacia in Cftr -/- mice.Infect Immun 69: 5138–5150.
Sharpe, A.H., Chen, L.B., Murphy, J.R., and Fields, B.N. (1980)Specific disruption of vimentin filament organization in monkeykidney CV-1 cells by diphtheria toxin, exotoxin A and cyclo-heximide. Proc Natl Acad Sci USA 77: 7267–7271.
Simel, D.L., Mastin, J.P., Pratt, P.C., Wisseman, C.L., Shelburne,J.D., and Spock, A. (1984) Scanning electron microscopicstudy of the airways in normal children and in patients withcystic fibrosis and other lung diseases. Pediatr Pathol 2:47–64.
Steele-Mortimer, O., Meresse, S., Gorvel, J.-P., Toh, B.-H., andFinlay, B.B. (1999) Biogenesis of Salmonella typhimurium-containing vacuoles in epithelial cells involves interactions withthe early endocytic pathway. Cell Microbiol 1: 33–49.
Steel-Mortimer, O., Knodler, L.A., and Finlay, B.B. (2001)Poisons, ruffles, rockets: bacterial pathogens and the hostcytoskeleton. Traffic 1: 107–118.
Sun, L., Jiang, R.Z., Steinbach, S., Holmes, A., Campanelli, C.,Forstner, J., et al. (1995) The emergence of a highly trans-missible lineage of cbl+ Pseudomonas (Burkholderia) cepaciacausing CF centre epidemics in North America and Britain.Nature Med 1: 661–666.
Svanborg, C., Hedlund, M., Connell, H., Agace, W., Duan, R.D.,Nilsson, A., and Wult, B. (1996) Bacterial adherence andmucosal cytokine responses. Receptors and transmembranesignalling. Ann NY Acad Sci 797: 177–190.
Sylvester, F.A., Sajjan, U.S., and Forstner, J.F. (1996) Burk-holderia (basonym Pseudomonas) cepacia binding to lipidreceptors. Infect Immun 64: 1420–1425.
Tablan, O.C., Chobra, T.L., Schidlow, D.V., White, J.W., Hardy,K.A., Gilligan, P.H., et al. (1985) Pseudomonas cepacia colo-nization in patients with cystic fibrosis: risk factors and clinicaloutcome. J Pediatr 107: 382–387.
Tipper, J.L., Ingham, E., Cove, J.H., Todd, N.J., and Kerr, K.G.(1998) Survival and multiplication of Burkholderia cepaciawithin respiratory epithelial cells. Clin Microbiol Infect 4:450–459.
