Published online before print August 21, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Schools of
* Biomedical Sciences and
Clinical Laboratory Sciences, University of Nottingham Medical School, Nottingham NG7 2UH, United Kingdom;
Institut für Immunologie, Philipps-Universität, 35032 Marburg, Germany;
¶ INSERM U431, Université de Montpellier II, 34095 Montpellier, France
1 Correspondence: University of Nottingham Medical School, School of Biomedical Sciences, Queen’s Medical Centre, Nottingham NG7 2UH, UK; E-mail:michael.rittig{at}nottingham.ac.uk
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, IL-8) and CC (MIP-1, MIP-1ß, MCP-1, RANTES) chemokines, as well as pro- (IL-6, TNF-) and anti-inflammatory (IL-10) cytokines released by challenged monocytes. Upon uptake, phagosomes containing rough brucellae develop selective fusion competence to form spacious communal compartments, whereas phagosomes containing smooth brucellae are nonfusiogenic. Collectively, our data suggest that rough brucellae attract and infect monocytes more effectively than smooth brucellae, but only smooth LPS phenotypes establish a specific host cell compartment permitting successful parasitism. These novel findings link the LPS phenotype of Brucella and its virulence for humans at the level of the infected host cells. Whether this is due to a direct effect of the LPS molecules or to upstream bacterial mechanisms remains to be established.
Key Words: phagocytosis • intracellular trafficking • intracellular parasitism • host–pathogen interaction
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The coccobacterium Brucella—a facultative intracellular pathogen of various mammals that is the source of a major zoonotic disease for humans worldwide—displays two distinct LPS phenotypes, a smooth form characterized by the fully expressed O-chain and a rough form with a substantially reduced or absent O-chain and maybe outer core section [3 ]. Rough forms arise spontaneously from smooth forms and accumulate in culture, suggesting that O-chain expression is nonessential and metabolically demanding under culture conditions [4 ]. Thus, it can be assumed that the that O-chain has to be of selective advantage in other conditions, presumably in the host. The "major" species (i.e., strongly pathogenic for humans) B. melitensis, B. suis, and B. abortus indeed are fully virulent only in their natural smooth form and non- or much less virulent in their mutant rough forms, both in humans and their primary hosts [5 ]. With LPS being the most prominent component of the outer cell membrane of brucellae and dominating the antibody response of the host, it seems reasonable to assume that LPS is a key factor in the pathogenicity of human brucellosis [5 ]. This assumption is supported by the fact that the identification of other classical virulence factors of Gram-negative bacteria—such as flagella, fimbriae, capsules, and exotoxins—in Brucella has been elusive, and the genome analysis of two of the major Brucella species revealed that most functional gene sequences encoding for known virulence factors are missing [6 , 7 ]. Accordingly, it has been suggested that LPS may be important for the extracellular survival of brucellae in the host [8 , 9 ], for the intracellular survival in the host cells [10 ], or both [11 ].
However, several exceptions from this rule have confused interpretation [4 , 5 ]. First, among the "minor" species (i.e., nonpathogenic or rarely pathogenic for humans), B. canis is naturally rough and B. neotomae naturally smooth, but the former causes anecdotal infections in humans and the latter not. Second, although generally being nonvirulent in humans, the minor species are fully virulent in their primary hosts. Third, some rough Brucella mutants are able to survive for long periods in cultured host cells. Fourth, the LPS of Brucella is by some orders of magnitude less biologically potent than that of other Gram-negative bacteria such as Escherichia coli [12 , 13 ]. At present, it is not clear whether the correlation between virulence and LPS phenotype in human brucellosis is causal or not, in what stage of infection virulence is being established, and what factors exactly are affected by the LPS phenotype—the adaptability of the bacterial parasite to changing environments, the modulation of the host cell, or the effectiveness of the host immune response.
To this end, we have compared the interaction of human mononuclear phagocytes with smooth or rough LPS phenotypes of Brucella. We were focusing on the two main pathogenic smooth species B. melitensis and B. suis and a fully characterized constructed rough mutant each. Our results confirm and extend previous findings in other systems that mononuclear phagocytes differentiate considerably between smooth and rough strains in terms of kinetics of attachment, uptake, and replication. We show for the first time that the two LPS phenotypes establish different host cell compartments and that the challenged host cells release a different chemokine/cytokine pattern. Thus, the LPS phenotype of brucellae obviously is linked with two early effects in humans, one on the local inflammatory response and one on the intracellular compartments occupied in the preferential host cells. Whether this is due to a direct effect of the smooth or rough LPS molecules themselves or to bacterial mechanisms upstream of LPS biosynthesis remains to be established.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(CXCL1) from Sigma-Aldrich; mAb to MCP-1 (CCL1), IL-6, IL-10 and TNF- from BD PharMingen; mAb to IL-8 (CXCL8), MIP-1 (CCL3), MIP-1ß (CCL4) and RANTES (CCL5), as well as a biotinylated secondary Ab from R&D Systems Biotech; and a streptavidin-peroxidase conjugate from Boehringer Mannheim (now Roche Diagnostics). Recombinant human GRO-, MCP-1, IL-10, IL-8, MIP-1, MIP-1ß, and RANTES were all obtained from R&D Systems Biotech; TNF- from BD PharMingen; and IL-6 from PBH Pharma Biotechnologie, Hannover, Germany. Items from the multinational companies were ordered at the respective German, English, French, or European offices as listed on the companies’ web pages.
Bacteria
The Brucella strains used in this study are listed in Table 1 . The two mutant strains have the mini-Tn5 Km transposon inserted in the manB or perA gene encoding for phosphomannomutase or perosamine synthetase, respectively, resulting in a truncated LPS with rough phenotype [14
, 15
]. Liquid cultures of brucellae in trypticase soy broth were raised overnight from stocks growing on trypticase soy agar plates. Immediately before the experiments, the density of the late stationary-phase cultures were adjusted in a spectrophotometer at OD600. Aliquoted bacteria were washed with PBS and resuspended in host cell culture medium. If necessary, bacteria were killed by exposing aliquots to heat (60°C for 60 min) or p-formaldehyde (6% in PBS for 6 h).
|
View this table: [in a new window] |
Table 1. Brucella Strains Used in This Study
|
Flow cytometry
Killed bacteria were directly labeled with 50 µg/ml FITC (isomer I) for 15 min at room temperature following the protocol of Miller and Quarles [18
] and stored protected from light at 4°C. Aliquots of PBMC (1x106 each) in polypropylene test tubes were challenged with FITC-labeled brucellae at a bacterium:host cell ratio (multiplicity of infection, MOI) of 500:1 in a total volume of 0.5 ml RPMI+ for 30 min, washed thoroughly with PBS containing 0.1% BSA, incubated with a saturating concentration of a directly fluorochrome-labeled mAb to CD14 for 20 min, which results in the labeling of 
85% of the heterogenous PBM [19
], washed again twice with 0.1% PBS-BSA, and resuspended in 0.5% formaldehyde in PBS before flow cytometric analysis. All these steps were performed at 4°C or on ice, and all samples were thoroughly vortexed to reverse any aggregation. 104 events for each sample were analyzed using an Beckman Coulter EPICS ALTRA flow cytometer with Expo32 software. For comparison, the fluorescence intensities of the FITC-labeled bacteria were adjusted to the same level to compensate for minor differences in the labeling intensities between different strains.
Electron microscopy
Phagocytosis was synchronized by mixing precooled (to 4°C) aliquots of PBM (1x106 each) and brucellae in polypropylene tubes in a total volume of 0.5 ml RPMI+ and a MOI of 500:1. In some experiments monocytes were coincubated with thoroughly vortexed precooled aliquots of brucellae and latex beads at a MOI of 250:1 each. After 15 min at 4°C, the suspensions were placed in a water bath at 37°C for another 15 min before being washed twice with large volumes of PBS. The pelleted monocytes were resuspended in RPMI+, seeded into Falcon Primaria T25 tissue culture flasks, and chased for up to 24 h in presence of 50 µg/ml gentamicin in order to kill any remaining extracellular brucellae. The chase was stopped by adding an excess amount of warm cacodylate-buffered aldehyde fixative [20
] to the flasks. Following overnight fixation at 4°C, the adherent cells were gently scraped off the flasks and processed for transmission electron microscopy following established protocols [20
]. Ultrathin sections were placed on 200-square mesh grids and examined with a Phillips TEM410. Events were quantified by the stereological approach of Mayhew et al. [21
], which allows for the simple and nonbiased scoring of a large number of electron microscopical samples. Briefly, having selected the optimal magnification, photos were taken from randomly chosen cells at predetermined corners of several grid windows. The host cell profiles on these micrographs were superimposed with test-line lattices to count the number of sites at which test-lines intersected events of interest or not. The 
-squared test was used to statistically compare "expected" and "observed" distributions within the experimental groups.
CFU counts
Aliquots of PBM (5x105 each/well) in Falcon PrimariaTM 24-well tissue culture plates were challenged with Brucella in a total of 0.5 ml RPMI+ at a MOI of 5:1 to 500:1 for 15 min. Some PBM were left untreated or were exposed to 10 ng/ml LPS from E. coli as negative and positive controls, respectively, for the ELISA measurements described below. Having rinsed the wells thoroughly with PBS and refilled with 1 ml each of fresh RPMI+, the PBM were given another 45 min to complete phagocytosis before adding 50 µg/ml gentamicin to the culture medium. At the end of the respective chase periods of up to 48 h, the culture supernatants were harvested for the ELISAs, whereas the host cells were osmotically lysed with 0.5 ml/well of 0.2% Triton X-100 in cold distilled water. Trypticase soy agar plates were inoculated in triplicate with 100 µl each of supernatant in serial dilutions and evaluated for presence of CFU. As the naturally rough strains grow somewhat slower than the naturally smooth strains and their rough mutant strains used in this study, care was taken to count CFUs of about the same size (typically 3-4 days old for smooth species and 4-5 days old for rough species).
ELISAs
The culture supernatants mentioned above were filtered through 0.2 µm Gelman Acrodisc filters, centrifuged at 16,000 x g for 15 min, and stored deep frozen until the actual measurements. Chemokine and cytokine release was determined by specific triple-sandwich ELISAs developed in our laboratory, as described in detail before [22
]. Briefly, Nunc Maxisorp 96-well microtiter plates were coated with Ab specific for IL-6, IL-8, IL-10, TNF-, GRO-, MCP-1, MIP-1, MIP-1ß, and RANTES, and blocked with 2% PBS-BSA. The wells were incubated with aliquots of culture supernatants (100 µl/well) followed by a biotinylated secondary Ab and a streptavidin-peroxidase complex. Conversion of the chromogenic substrate o-phenylenediamine dihydrochloride was measured and evaluated by means of a Dynatech MR 7000 photometric plate reader with intrinsic software (Dynex Technologies, Denkendorf, Germany). Readouts were plotted against a standard curve obtained with the respective recombinant chemokines and cytokines. The sensitivities of the established ELISAs were <20 pg/ml for IL-8, MCP-1, MIP-1ß, RANTES, IL-6, and IL-10, <50 pg/ml for GRO- and <100 pg/ml for TNF- and MIP-1, with less than 5% intra- and interassay variances.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
 View larger version (54K): [in a new window] |
Figure 1. Binding of Brucella organisms with rough or smooth LPS to human monocytes. Human PBMC were incubated with FITC-labeled brucellae at a MOI of 500:1 for 30 min at 4°C and subsequently labeled with an APC-conjugated antibody to CD14. 104 events for each sample were analyzed with a flow cytometer. The data are representative of results obtained from five blood donors. The dot plots show that CD14+ cells bind higher numbers of the constructed rough LPS mutant (B. suis  manB; 1C) than the parental smooth LPS organisms (B. suis wt; 1B). This pattern was consistent for all Brucella strains tested. In addition to CD14+ monocytes, a large number of CD14- cells also showed binding to Brucella strains; these cells have the scatter characteristics of lymphocytes when gated onto a forward/side scatter plot (FS/SS; Fig. 1D
1E
1F
).
|
 View larger version (23K): [in a new window] |
Figure 2. Survival and replication of Brucella organisms with rough or smooth LPS in human monocytes. Human monocytes were challenged with brucellae at a MOI of 100:1 for 30 min and chased for up to 36 h in presence of 50 µg/ml gentamicin. The graph shows the number of CFU of brucellae reisolated at various intervals from osmotically lysed host cells. Data are representative of results obtained from eight blood donors. SD are in the range of 0.1-0.2 log units. Higher numbers of constructed rough LPS mutants (B. suis manB and B. melitensis perA) than parental smooth LPS organisms (B. suis wt and B. melitensis wt) are internalized with most of the bacteria being killed within the first hours after uptake. The surviving smooth LPS organisms then start to replicate, whereas the surviving rough LPS organisms fail to do so. This pattern was also seen with the naturally rough species investigated (data not shown).
|
An obvious question at this point is whether these effects can be induced by LPS alone or need the intact bacterium. As a first step to address this question, we challenged monocytes with wild-type smooth B. suis in the absence or presence of exogenous smooth or rough LPS purified from respective B. abortus strains. CFU counts showed that the presence of either LPS type reduced the number of intracellular viable bacteria to a similar extent, whereas the kinetics remained the same (Fig. 2B) . This is consistent with the view that LPS of Brucella acts as a ligand for phagocytosis-competent host cell receptors but has no obvious effect on the subsequent intracellular survival.
Phagosomes containing rough brucellae become hyperfusiogenic and phagosomes containing smooth brucellae become nonfusiogenic
As the results so far showed that both LPS phenotypes had similar kinetics of binding and uptake—albeit to a different extent—but not replication, we became interested in the intracellular compartments occupied by the different phenotypes. The electron microscopal investigation of PBM infected with brucellae showed that uptake of both smooth and rough organisms resulted initially in individual compartments with the typical ultrastructural morphology of early phagosomes (Fig. 3A
3B
3D
). As shown before by our group [24
], the majority of internalized smooth brucellae remained in these early-type phagosomes (Fig. 3C)
for up to 24 h, which was the longest chasing period studied, whereas some of the phagosomes harboring smooth brucellae matured to late-type phagosomes. Challenge with prekilled smooth brucellae resulted in the slow progressive maturation of all the early-type phagosomes (Fig. 3J) , suggesting that the arrest in phagosome maturation typical of smooth LPS phenotypes depends on the viability of the pathogen.
 View larger version (130K): [in a new window] |
Figure 3. Intracellular compartments occupied by Brucella organisms with rough or smooth LPS in human monocytes. Human monocytes were challenged with brucellae at a MOI of 500:1 or a 1:1 mixture of brucellae and latex beads for 15 min, chased for up to 24 h (hpi = hours post infection) and processed for electron microscopy. The electron micrographs show the intracellular compartments occupied by brucellae with smooth (B. melitensis wt, B. suis wt) or rough (B. melitensis perA, B. suis manB, B. ovis, B. canis) LPS phenotype and latex beads. All bars = 2 µm except for A = 1 µm, C = 0.3 µm and J = 0.5 µm. (A) Brucellae attached to the surface of the monocytes (asterisks) are engulfed by pseudopods and enclosed in individual phagosomes with tightly apposed walls (x). (B, C) Smooth brucellae remain in these tight-fitting phagosomes (circle in B; x in C). (D-I) The initially individual phagosomes containing rough brucellae (D) progressively fuse (circles in E) giving rise to one or few giant communal vacuoles (F- I). Not only prefused phagosomes, but also individual phagosomes merge with these megasomes (arrows in F). Some intraluminal brucellae are abutted to the phagosomal wall (arrows in H). (J, K) Prekilled brucellae of either LPS phenotype are delivered to phagosomes, which mature to phagolysosomes (circle in J and K). (L) Simultaneous uptake of rough brucellae and latex beads (LB) demonstrates that phagosome fusion is restricted to brucellae-containing phagosomes (circles). These features were consistently observed in monocytes from six different donors.
|
 View larger version (28K): [in a new window] |
Figure 4. Rate and kinetics of phagosome fusion in human monocytes upon challenge with various Brucella strains with rough or smooth LPS. Phagosome fusion as demonstrated in Fig. 3
was quantified on electron micrographs using the stereological approach of Mayhew et al. [21
, 62
]. Vacuoles containing three or more individual brucellae were regarded as fused phagosomes. (A) Phagosome fusion is more prominent with the constructed rough mutants (B. melitensis perA, B. suis manB) than with the natural rough species (B. ovis, B. canis). Killing of the mutants reduces phagosome fusion to about the levels observed with the natural smooth parental species (B. melitensis wt, B. suis wt). (B) The relative proportions of the number of fused phagosomes harboring B. suis manB organisms per monocyte (fp/m) over an eight-hour period indicate that the individual phagosomes fuse at several places simultaneously and these prefused phagosomes then coalesce.
|
The constructed rough mutants and naturally rough species consistently induced a much higher release of the CXC chemokines GRO- (CXCL1) and IL-8 (CXCL8), the CC chemokines MIP-1 (CCL3) and RANTES (CCL5), as well as the proinflammatory cytokine IL-6, as compared with the parental smooth strains (Fig. 5A
). The same pattern was seen for the CC chemokines MCP-1 (CCL1) and MIP-1ß (CCL4) and the anti-inflammatory cytokine IL-10 (data not shown). The amounts induced by the rough mutants were equivalent to those induced by 10 ng/ml LPS from E. coli (data not shown). Except for RANTES and TNF-, the difference between smooth and rough strains was obvious right from the beginning (especially with IL-6) but became more pronounced after 8-24 h (Fig. 5A)
. RANTES showed a somewhat delayed and less pronounced increase, which was enhanced accordingly when using monocyte-derived macrophages or vitamin D3-differentiated THP-1 cells instead of monocytes as host cells (data not shown). The TNF- levels showed the early peak and subsequent slow decline characteristic of this cytokine; here, the difference between the smooth parental strains and rough mutants was slightly overlain by an interspecies difference with B. suis inducing higher levels of TNF- than B. melitensis (Fig. 5A)
. When studying the dose-response relationship of cytokine/chemokine release, it was observed that the gap between the smooth parental strains and rough mutants widened when a MOI higher than 10:1 is was used (Fig. 5B)
.
 View larger version (38K): [in a new window] |
Figure 5. Chemokine and cytokine release by human monocytes upon challenge with Brucella strains with rough or smooth LPS as determined by specific ELISAs. (A) Kinetics of cytokine/chemokine release. Human monocytes were challenged for 60 min either with brucellae displaying smooth or rough LPS at a MOI of 20:1, or with LPS from E. coli serotype 055:B5 as a positive control, or were left untreated as negative control. Culture supernatants were harvested at various intervals for measurements. The data shown are representative for five independent experiments using monocytes from different donors. Results are displayed as means ± SD of triplicate measurements. The constructed rough brucellae mutants (B. melitensis perA and B. suis manB) consistently induce much higher release of GRO-, IL-8, MIP-1, RANTES and IL-6 than the parental smooth strains (B. melitensis wt and B. suis wt). Although this pattern is obvious right from the beginning, it becomes more pronounced after 8-24 h. The same difference is seen for MCP-1, MIP-1ß, and IL-10 (data not shown). The amounts of cytokines and chemokines induced by the rough brucellae mutants are equivalent to those induced by 10 ng/ml LPS from E. coli (data not shown). (B) Dose-response relationship of cytokine/chemokine release. Human monocytes were challenged for 60 min with brucellae displaying smooth or rough LPS at increasing MOI of 5:1 to 500:1. Culture supernatants were harvested after 24 h for measurements. The data shown are representative for two independent experiments using monocytes from different donors. Results are displayed as means ± SD of triplicate measurements. One example each typical of the groups investigated, that is, cytokines (TNF-), CXC chemokines (GRO-), and CC chemokines (MIP-1), is depicted. For each group, the gap between the two LPS phenotypes widens when an MOI higher than 10:1 is used. (C) Effect of exogenous LPS on the release of TNF- by Brucella-challenged monocytes. Human monocytes were challenged for 60 min with smooth B. suis wt in the absence or presence of 0.1, 1, or 10 µg/ml exogenous smooth or rough LPS from B. abortus strains 2308 and 45/20, respectively, or left unchallenged for negative control. Culture supernatants were harvested after 24 h for measurement of TNF-. Presence of exogenous smooth or rough LPS did not alter the amount of TNF- released by Brucella-challenged monocytes.
|
from monocytes challenged with B. suis in the absence or presence of highly purified smooth or rough LPS from B. abortus. However, the amount of TNF- released was similar in the presence of smooth or rough exogenous LPS (Fig. 5C)
. |
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Influence of the brucellae LPS phenotype on binding and uptake by human monocytes
In the present study, more rough than smooth brucellae bound to human monocytes and were internalized (Fig. 1
2
3)
. This confirms recent data in human monocytes/macrophages in which a smooth B. melitensis wt strain was far less internalized than a constructed rough mutant strain [26
]. As phagocytosis took place in absence of heat-labile complement components and Brucella-specific antibodies (i.e., largely "nonopsonic" conditions), it appears that innate pattern-recognition receptors play a major role in phagocytosis of brucellae. It could be speculated that the absence of the O-chain either unmasks ligands of the core oligosaccharide or enables hydrophobic interactions, thus enhancing the uptake of rough LPS phenotypes. However, the presence of exogenous smooth, as well as rough LPS reduced the uptake of smooth brucellae to a similar extent (Fig. 2B)
, suggesting that the O-chain as well as the core oligosaccharide can bind to host cell receptors, as it is known from other Gram-negative bacteria. For example, Actinobacillus pleuropneumoniae use the O-chain as a major adhesion for lung tissue [27
], whereas Pseudomonas aeruginosa and Salmonella typhi use the core oligosaccharide as a ligand [28
]. Although the host cell receptor(s) in question need to be identified, work by our group [24
] and others [29
] indicates that adhesion molecules of monocytes/macrophages are involved in binding and/or uptake of Brucella.
Influence of the brucellae LPS phenotype on survival and replication in human monocytes
Following a characteristic decline within the first hours after internalization, the number of viable reisolated smooth organisms increased again, whereas the number of viable reisolated rough organisms stayed at that basal level for several days (Fig. 2) . This difference is typical of both the smooth and rough LPS phenotypes of Brucella [11
, 26
]. Although it is well documented that in the extracellular environment, presence or absence of the O-chain is associated with resistance or susceptibility to complement-mediate killing [30
, 31
], the role of LPS for the intracellular survival and replication of Brucella is less clear. Virulent B. abortus were able to counteract the effect of nitric oxide in a murine macrophage line [32
], but nitric oxide does not play a significant role in human monocytes/macrophages. The O-chain obviously protects Brucella against bactericidal cationic peptides [11
, 33
, 34
], but this does not explain differences between strains of the same LPS phenotype. The active inhibition of TNF- production in human monocyte/macrophage-like cells, which has been considered to be a major attribute of Brucella virulence [35
], was observed with smooth, as well as rough strains and still remains enigmatic.
As the numbers of viable intracellular bacteria dropped considerably at a similar degree for both LPS phenotypes, it can be assumed that the LPS phenotype somehow is linked to the ability to replicate rather than to survive. Accordingly, deep rough LPS mutants of Bordatella bronchiseptica have been shown to survive within the host cells as efficiently as the smooth parental strains [36
]. The observation that the perA mutant was less attenuated that the manB mutant (Fig. 2)
could be due to the perA gene product acting further downstream of the manB gene product [37
]. Again, the host cells or species used as experimental system obviously influence the reaction to the pathogen, as the B. melitensis perA mutant, although being attenuated in our system, showed the same survival as the parental strain in a bovine peritoneal macrophage-like cell line [14
].
Influence of the brucellae LPS phenotype on the phagosome trafficking in human monocytes
Upon uptake by monocytes, both LPS phenotypes initially were located in phagosomes with similar morphology, but soon after followed different intracellular pathways (Fig. 3)
. The finding that the further maturation of the phagosomes harboring smooth brucellae was arrested at an early stage confirms our results from a previous study [24
]. It should be stressed here that Brucella obviously follows different routes in professional and nonprofessional phagocytes. Because of the important work of Cheville and co-workers it has long been known that in epithelial cells, brucellae survive and replicate within cisternae of the rough endoplasmic reticulum (rER), reached late after infection via a phagosomal route [38
39
40
]. More recently, Gorvel and co-workers suggested an retrograde autophagocytic pathway for B. abortus in a peculiar HeLa cell subclone [41
, 42
] a finding that awaits confirmation in more physiological cells. In professional phagocytes, however, an association between brucellae and the rER or autophagosomes was observed neither in the present nor previous ultrastructural studies [24
, 43
, 44
]. It is likely that the quite different membrane sorting mechanisms and routes of vesicle trafficking in polarized epithelial cells and nonpolarized professional phagocytes lead to different pathways for Brucella in these two types of host cells.
The characteristic close apposition between the enclosed brucellae and the phagosomal wall has been observed in other systems as well [45 46 47 48 49 50 ] and is thought to be crucial for bacterial secretion systems to deliver specific effector molecules across the narrow intraphagosomal space into the host cell. Nothing is yet known about the molecular characteristics of the tight phagosome harboring Brucella, whereas for Mycobacteria and Legionella, it has been demonstrated that their tight phagosomes selectively accumulate, as well as exclude host cell factors regulating membrane trafficking [46 , 51 ]. Rough brucellae, in contrast, induced the progressive fusion of their individual phagosomes, resulting in one or few giant communal vacuoles (Fig. 3) . Similar compartments termed "inclusions", "parasitophorous vacuoles", and "megasomes" are induced by Chlamydia [52 ], Leishmania [53 ], and Helicobacter pylori [25 ], respectively. The origin of these giant compartments is attributed to pathogen-induced active modification of the phagosomes to become fusiogenic. Our findings that the homotypic fusiogenicity is neither observed with phagosomes harboring killed rough brucellae nor extend to latex-bearing phagosomes, support this view.
The molecular mechanisms for this opposite effect of smooth and rough LPS phenotypes of Brucella on the trafficking of their phagosomes is not clear. The principal difference between the two LPS phenotypes is the presence or absence, respectively, of the O-chain, and it is difficult to imagine by which means absence of the O-chain may induce homotypic fusiogenicity of the phagosomes. As phenotypically rough Brucella have alterations in biosynthetic pathways which may not only affect synthesis or transport of complete O-linked polysaccharides, a rough LPS phenotype may just be indicative of altered upstream mechanisms which influence the host cell in a profusiogenic way. Alternatively, the presence or absence of the O-chain may be linked to different receptors possibly sorting for different intracellular compartments with different fusiogenicity. Thus, recent studies have shown that, in murine macrophages, uptake of virulent smooth brucellae but not of smooth nonvirulent mutants or rough mutants was associated with lipid rafts [54 55 56 57 ]. This lipid raft-mediated uptake was believed to be macropinocytosis [56 , 57 ], but no ultrastructural evidence was given, and it should be kept in mind that lipid rafts are part of several cellular activities, including phagocytosis as well as macropinocytosis [58 ]. As to what extent lipid rafts are involved in the genesis of the survival-permitting tight phagosomes induced by mycobacteria [59 ] and smooth brucellae [24 ] has yet to be shown. As for the uptake of Streptococcus pyogenes, the blockade of lipid rafts inhibited the sorting into spacious compartments but not into tight compartments [60 ].
Influence of the brucellae LPS phenotype on cytokine/chemokine release by human monocytes
Although there is a large body of literature on the topic of brucellae-induced cytokine release, especially in vaccinated animals, surprisingly few data are available concerning chemokine production in humans during Brucella infection. Two recent reports mentioned that cultured human mononuclear phagocytes can produce MCP-1, MIP-1, and MIP-1ß upon challenge with smooth LPS [13
, 61
]. The present study is the first one that systematically compares the effect of the two LPS phenotypes on a large number of cytokines/chemokines released by cultured human mononuclear phagocytes. The results demonstrate that rough brucellae generally induce a considerably higher amount of all of the cytokines and chemokines investigated as smooth brucellae (Fig. 5A
5B)
, an effect that depends on the intact organisms and cannot be mimicked by purified LPS (Fig. 5C)
. Although the situation in patients will be influenced by a variety of complex regulatory systems in addition to the clinical status of the patients and the stage of disease, our results suggest that an infection with rough brucellae will lead to a more pronounced inflammatory host response and attract higher number of immune cells, as compared with smooth brucellae. However, the findings that challenge with B. suis induced higher levels of TNF- than B. melitensis (Fig. 5A)
and that exogenous smooth and rough LPS had similar effects support the view that species-specific bacterial components other than the LPS phenotype contribute to the cytokine/chemokine response of the infected host.
In conclusion, the results of the present study demonstrate that in isolated human monocytes the smooth vs. rough LPS phenotype of Brucella accounts for considerable differences in infection, both quantitatively (regarding the number of bound and internalized bacteria as well as the amounts of chemokines and cytokines released) and qualitatively (regarding the intracellular compartments occupied and the trafficking of these phagosomes). Although the endotoxicity of lipid A is certainly of great significance in many disease processes, the saccharide domains of naturally occurring LPSs obviously are also of great significance in affecting the pathophysiology of Gram-negative infections. Thus, our data demonstrate that rough LPS phenotypes of Brucella attract and infect monocytes more effectively than smooth LPS phenotypes, but only the latter are able to establish a host cell compartment permitting successful parasitism. The O-chain as the outermost part of the LPS molecule is not only the major antigen targeted by host antibody responses but is also recognized by the innate arm of the immune system, which has effects on the extracellular survival of brucellae in the host. Our novel findings additionally link brucellae LPS phenotype and virulence for humans at the level of the infected host cell. Whether this is due to a direct effect of the LPS molecules or to upstream bacterial mechanisms remains to be established.
Received January 14, 2003; revised April 11, 2003; accepted July 15, 2003.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. V. Delpino, C. A. Fossati, and P. C. Baldi Proinflammatory Response of Human Osteoblastic Cell Lines and Osteoblast-Monocyte Interaction upon Infection with Brucella spp. Infect. Immun., March 1, 2009; 77(3): 984 - 995. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Rajashekara, J. Covert, E. Petersen, L. Eskra, and G. Splitter Genomic Island 2 of Brucella melitensis Is a Major Virulence Determinant: Functional Analyses of Genomic Islands J. Bacteriol., September 15, 2008; 190(18): 6243 - 6252. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Pei, Q. Wu, M. Kahl-McDonagh, and T. A. Ficht Cytotoxicity in Macrophages Infected with Rough Brucella Mutants Is Type IV Secretion System Dependent Infect. Immun., January 1, 2008; 76(1): 30 - 37. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Billard, J. Dornand, and A. Gross Interaction of Brucella suis and Brucella abortus Rough Strains with Human Dendritic Cells Infect. Immun., December 1, 2007; 75(12): 5916 - 5923. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. M. Kahl-McDonagh and T. A. Ficht Evaluation of Protection Afforded by Brucella abortus and Brucella melitensis Unmarked Deletion Mutants Exhibiting Different Rates of Clearance in BALB/c Mice Infect. Immun., July 1, 2006; 74(7): 4048 - 4057. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Pei, J. E. Turse, Q. Wu, and T. A. Ficht Brucella abortus Rough Mutants Induce Macrophage Oncosis That Requires Bacterial Protein Synthesis and Direct Interaction with the Macrophage. Infect. Immun., May 1, 2006; 74(5): 2667 - 2675. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Rajashekara, D. A. Glover, M. Banai, D. O'Callaghan, and G. A. Splitter Attenuated Bioluminescent Brucella melitensis Mutants GR019 (virB4), GR024 (galE), and GR026 (BMEI1090-BMEI1091) Confer Protection in Mice. Infect. Immun., May 1, 2006; 74(5): 2925 - 2936. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Billard, C. Cazevieille, J. Dornand, and A. Gross High Susceptibility of Human Dendritic Cells to Invasion by the Intracellular Pathogens Brucella suis, B. abortus, and B. melitensis Infect. Immun., December 1, 2005; 73(12): 8418 - 8424. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. P. Jimenez de Bagues, A. Gross, A. Terraza, and J. Dornand Regulation of the Mitogen-Activated Protein Kinases by Brucella spp. Expressing a Smooth and Rough Phenotype: Relationship to Pathogen Invasiveness Infect. Immun., May 1, 2005; 73(5): 3178 - 3183. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R.-A. Boigegrain, I. Salhi, M.-T. Alvarez-Martinez, J. Machold, Y. Fedon, M. Arpagaus, C. Weise, M. Rittig, and B. Rouot Release of Periplasmic Proteins of Brucella suis upon Acidic Shock Involves the Outer Membrane Protein Omp25 Infect. Immun., October 1, 2004; 72(10): 5693 - 5703. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. P. Jimenez de Bagues, A. Terraza, A. Gross, and J. Dornand Different Responses of Macrophages to Smooth and Rough Brucella spp.: Relationship to Virulence Infect. Immun., April 1, 2004; 72(4): 2429 - 2433. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
