SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Bartonella bacilliformis, a Gram-negative, flagellated bacterium, infects human erythrocytes (haematic phase) and endothelial cells (tissue phase), resulting in a biphasic disease. In the tissue phase of disease (verruga peruana), infection leads to infection of endothelial cells and a pronounced proliferation of these cells, resulting in characteristic skin eruptions of papules and nodules. We have studied the properties of endothelial cells infected in vitro. Extensive cytoskeletal remodelling of endothelial cells occurred after infection in vitro with B. bacilliformis. The cells became spindle shaped and contained arrays of actin stress fibres orientated parallel to the long axis of the cell. Cell–cell contacts were disrupted, along with the distribution of the plasma membrane marker protein, PECAM-1, which participates in cell–cell junctions. The prominent stress fibres terminated in an increased number of focal contacts, which were studied using immunofluorescent staining for paxillin, a cytoplasmic protein that localizes in the focal adhesions. These morphological changes are consistent with activation of intracellular Rho by B. bacilliformis. Formation of stress fibres and the increased number of focal adhesions could be prevented by preincubation of the endothelial cells with C3 exoenzyme, which inactivates intracellular Rho by ADP ribosylation. Endothelial cell motility was greatly diminished in infected cells and the cells did not respond effectively to a stimulus that would evoke motility. In addition, infection of endothelial cells interfered with their ability to form networks of capillary tubes when suspended within three-dimensional collagen matrices. If the properties of infected endothelial cells in vivo are similar, the infected cells will probably not participate effectively in angiogenesis.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Infections by Bartonella bacilliformis in humans manifest both a high and a low capacity to overwhelm body defences, in two distinct stages of infection. The initial haematic phase (Oroya fever) is characterized by a severe, often fatal, haemolytic anaemia caused by accelerated splenic and hepatic destruction of parasitized red cells. This initial phase, which may sometimes be absent, is followed by a non-fatal, chronic tissue phase (Carrion's disease), characterized by the appearance of haemangioma-like papules or nodules, called verrugas, which can persist for more than a year.

Histologically, the verrugas are characterized by a marked concentration of B. bacilliformis, both within the cytoplasm of endothelial cells and extracellularly in the surrounding interstitium. The existence of numerous capillaries and pseudocapillaries within verrugas suggests that Bartonella infection of endothelial cells provokes a local angiogenic response (Arias-Stella et al., 1987; Schneider et al., 1993). Verrugas are histologically very similar to the haemangioma-like lesions of bacillary angiomatosis, associated with infection by B. henselae in immunocompromised patients (Tappero et al., 1993), as well as those observed in Kaposi's sarcoma. Bacillary angiomatosis can be successfully treated with antibiotics (whereas Kaposi's sarcoma cannot), demonstrating the importance of the continued presence of the bacteria (Webster et al., 1992). B. henselae produces one or more substances that stimulate growth of endothelial cells in vitro (Garcia et al., 1990; Maeno et al., 1999) and invasion of vascular endothelial cells may represent an early step in the formation of Bartonella-induced vasoproliferative responses (Garcia et al., 1992).

The bacteria in vivo target erythrocytes and endothelial cells exclusively, but in vitro B. bacilliformis can infect a variety of different cell types, including human umbilical vein endothelial cells (HUVECs), HeLa cells, Hep2cells and skin fibroblasts (Hill et al., 1992; Batterman et al., 1995). Therefore, the mechanism used for internalization must be general and not limited to erythrocytes and endothelial cells. However, a stimulatory effect on proliferation has been shown to be specific to vascular endothelial cells only (Maeno et al., 1999). Cytochalasin D added to host cells prior to infection inhibits the internalization of B. bacilliformis (Hill et al., 1992), which suggests an active role of the host cell in the uptake of B. bacilliformis and indicates that actin polymerization and cytoskeletal rearrangements are probably involved in the process. Cytoskeletal rearrangements following infection by B. quintana and B. henselae have been reported (Palmiri et al., 1996; Dehio et al., 1997). Addition of genistein, tyrphostin or staurosporine, inhibitors of tyrosine kinase and protein kinase C, significantly decreased the internalization of B. bacilliformis (Williams-Bouyer and Hill, 1999), suggesting a role for kinase activity and/or tyrosine phosphorylation in the bacterial internalization process.

It seems clear that an angiogenic response must occur in vivo in response to the bacterial infection, but the mechanism by which this is accomplished remains unknown. The bacteria could synthesize some diffusible product that, in analogy with known angiogenic factors, could stimulate endothelial cells to participate in angiogenesis (Garcia et al., 1990; 1992). Alternatively, host angiogenic factors could be produced as a consequence of either the infection of endothelial cells or an immune/macrophage response against the bacteria, or the infected cells themselves may have angiogenic potential.

In the present study, we demonstrate morphological and cytoskeletal alterations – formation of actin stress fibres, increased numbers of focal contacts and loss of cell–cell junctions in infected endothelial cells, which are consistent with activation of intracellular Rho. Inactivation of intracellular Rho with C3 exoenzyme prevented formation of stress fibres terminating in focal contacts and loss of cell–cell junctions. The infected cells appear not to be activated for angiogenesis, but rather to be grossly deficient in the ability to perform various aspects of angiogenesis, as assessed in vitro.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Infection of endothelial cells induces actin stress fibre formation

Infected endothelial cells progressively assumed an elongated spindle-shaped morphology, in contrast to the more rounded shape of the uninfected endothelial cells. To study the cytoskeletal rearrangement responsible for this altered morphology in endothelial cells infected with B. bacilliformis, phalloidin fluorescein was used to visualize the F-actin network. In uninfected cells, fibres of the F-actin network had no particular orientation and were present in a haphazard arrangement (Fig. 1A). Until 12 h post infection, infected cells displayed a F-actin network similar to that observed in uninfected cells. Between 12 and 24 h post infection, prominent, thick F-actin bundles (stress fibres) formed, arranged in parallel arrays and orientated with the long axis of the cell (Fig. 1B). Stress fibres were not formed when the endothelial cells were incubated with heat-killed or gentamycin-treated bacteria.

image

Figure 1. Formation of actin stress fibres and loss of cell–cell contacts in endothelial cells infected with B. bacilliformis. Endothelial cells grown on gelatin-coated coverslips to 80% confluency were infected for 24 h, fixed, permeabilized and stained with F-actin phalloidin-fluorescein and anti-PECAM-1 antibodies.

A. F-actin network in the uninfected endothelial cells.

B. Thick actin stress fibres orientated along the long axis of the infected endothelial cells.

C. Uninfected endothelial cells showing intense PECAM-1 staining at the cell–cell junctions, (marked by arrow).

D. Infected endothelial cells showing both loss of PECAM-1 staining and loss of cell–cell contacts (intercellular gaps; marked by arrow). Bar, 10 μm.

Download figure to PowerPoint

Infection of endothelial cells results in loss of cell–cell junctions

Loss of F-actin at cell–cell contacts and formation of gaps between adjacent cells became evident 24 h after infection. Loss of cell–cell junctional staining and formation of intercellular spaces was further observed after staining for these junctions using anti-platelet/endothelial cell adhesion molecule-1 (PECAM-1) antibodies. PECAM-1, a glycoprotein of the Ig superfamily, localizes at cell–cell junctions of endothelial cells, as seen in immunostaining of uninfected endothelial cells (Fig. 1C). Loss of PECAM-1 and F-actin at the junctions indicates that infection results in decreased cell–cell contacts (Fig. 1D).

Infection of endothelial cells causes an increase in cell–substrate focal contacts

Even after prolonged incubation, infected endothelial cells did not detach from monolayer, showing that they remained firmly adherent to the substratum. The increased formation of stress fibres suggested that the number or strength of focal adhesions with the extracellular matrix might also be increased. Immunostaining with anti-paxillin antibodies, a cytoplasmic protein that localizes to focal adhesions, permitted the identification of an increased number of prominent focal adhesions (Fig. 2A and B), located particularly at the distal ends of the longitudinal axis of infected cells. Immunostaining with anti-focal adhesion kinase (FAK) or anti-phosphotyrosine antibodies confirmed the presence of increased focal contacts in infected cells (Fig. 2C–F). When images of fluorescent anti-paxillin and F-actin stains were superimposed, the stress fibres formed during infection were seen to terminate in paxillin-containing focal contacts (Fig. 2G–I). Total endothelial cell lysates infected with B. bacilliformis for 2, 12 or 24 h showed an increase in the expression of paxillin and FAK during the time-course of infection (Fig. 3A). Anti-phosphotyrosine antibody immunocomplexes of infected endothelial cell lysates blotted with anti-paxillin and anti-FAK antibodies showed an increased tyrosine phosphorylation of both the protein targets with the increase in time of infection (Fig. 3B). Increased tyrosine phosphorylation of paxillin as well as FAK was also confirmed by immunoprecipitation with their respective specific antibodies and blotting with anti-phosphotyrosine antibodies (Fig. 3C). Increased focal contacts observed in infected cells could be attributed to both increased expression and tyrosine phosphorylation of paxillin and FAK.

image

Figure 2. Increased focal contacts and termination of stress fibres in focal adhesions. Endothelial cells grown on gelatin-coated coverslips to 80% confluency were infected for 24 h, fixed and permeabilized.

A, C and E. Uninfected endothelial cells showing few focal adhesions after staining with anti-paxillin (A), anti-FAK (C) and anti-phosphotyrosine (E) antibodies.

B, D and F. Infected cells showing prominent and increased focal contacts with anti-paxillin (B), anti-FAK (D) and anti-phosphotyrosine antibodies (F); the arrow in each panel shows focal contacts.

G, H and I. F-actin phalloidin-fluorescein (G), anti-paxillin (H), staining of the same cell and overlay of (G) and (H) showing termination of stress fibres in paxillin-rich focal contacts (I). Bar, 10 μm.

Download figure to PowerPoint

image

Figure 3. Protein levels and tyrosine phosphorylation of paxillin and FAK in infected endothelial cells. Endothelial cells were infected with B. bacilliformis for the different time intervals and cell lysates were prepared and used for immunoprecipitation as described in Experimental procedures.

A. Uninfected and infected (2, 12 and 24 h) endothelial cell lysates were electrophoresed on SDS-PAGE and analysed using Western blotting with anti-paxillin and anti-FAK antibodies.

B. Total lysates of uninfected and infected (2, 12 and 24 h) endothelial cells were immunoprecipitated with anti-phosphotyrosine monoclonal antibodies and, after separation of immunoprecipitates on SDS-polyacrylamide gradient gel (4–20%), proteins were transferred on to the membrane and analysed using Western blotting with anti-paxillin and anti-FAK monoclonal antibodies.

C. Total lysates of uninfected and infected (2, 12 and 24 h) endothelial cells were also immunoprecipitated with specific antibodies to paxillin and FAK and analysed using Western blotting with anti-phosphotyrosine antibodies. WB, Western blot analysis; IP, immunoprecipitation.

Download figure to PowerPoint

B. bacilliformis-induced actin stress fibre formation and increased focal contacts are blocked by C3 exoenzyme

Activation of Rho within cells is known to induce both actin stress fibre and focal contact formation. C3 exoenzyme from Clostridium botulinum was used to specifically inactivate endogenous Rho within cells. 100 μg ml−1 of C3 exoenzyme was found to be sufficient to markedly inactivate Rho in endothelial cells after 8 or 12 h of incubation of endothelial cells with C3 exoenzyme, but not after 4 h of incubation (data not shown). When C3 exoenzyme and bacteria were added simultaneously, no stress fibre formation or loss of cell–cell junctions was observed using F-actin phalloidin rhodamine staining (Fig. 4A and B). As addition of C3 exoenzyme was increasingly delayed relative to infection, there was a steady increase in stress fibre formation (Fig. 4C–H). When C3 exoenzyme was added 16 h after the infection with bacteria, endothelial cells prominently displayed stress fibres (Fig. 4I and J). Anti-paxillin antibody staining demonstrated an increasing number of focal contacts as addition of C3 exoenzyme was delayed.

image

Figure 4. Effect of C3 exoenzyme on B. bacilliformis-induced actin stress fibre formation in infected endothelial cells. Confluent monolayers of endothelial cells infected with B. bacilliformis were incubated with 100 μg ml−1 C3 exoenzyme for different time intervals and stained with F-actin phalloidin rhodamine to visualize the effects of C3 exoenzyme on the F-actin network.

A and B. C3 exoenzyme added simultaneously with the bacteria to endothelial cells.

C and D. C3 exoenzyme added after 4 h of infection.

E and F. C3 exoenzyme added after 8 h of infection.

G and H. C3 exoenzyme added after 12 h of infection.

I and J. C3 exoenzyme added after 16 h of infection.

The same field was visualized for GFP bacteria (A, C, E, G and I) and for F-actin phalloidin rhodamine (B, D, F, H and J) staining respectively. As the C3 exoenzyme treatment was delayed with respect to time of infection, there was a steady increase in the formation of stress fibres. Bar, 30 μm.

Download figure to PowerPoint

When endothelial cells were preincubated with C3 exoenzyme for 4, 8 or 12 h, no stress fibre or increased focal contacts were seen; however, an intense F-actin peripheral staining and thick PECAM-1 junctional staining were seen (data not shown). Preincubation with C3 exoenzyme under these conditions however, also prevents infection of endothelial cells by B. bacilliformis.

Endothelial cells pretreated with Y-27632 (a specific inhibitor of Rho-kinase, which is a downstream effector of Rho in the signalling cascade leading to cytoskeletal rearrangement) at a concentration of 100 μM for 1 h followed by infection with B. bacilliformis for 24 h showed no stress fibre formation after infection. Furthermore, inactivation of Rho-kinase had no effect on the internalization of B. bacilliformis in endothelial cells.

Reduced rate of migration of infected endothelial cells in in vitro migration assays

To determine whether B. bacilliformis infection of endothelial cells alters cell behaviour to mimic that of endothelial cells involved in angiogenic responses, cell migration assays were used to compare infected and uninfected cells. Confluent uninfected and infected (24 h, 50–100 added bacteria per cell) monolayers were scratched and monitored for endothelial cell migration into the scratch (wound-healing) over a period of 144 h. Uninfected endothelial cells migrated within 6 h and filled the gap within 18–24 h (Fig. 5A). With infected cells, no migration was observed up to 20 h. After 24 h, some migration was evident, but the scratch could nearly be filled only after incubation for 144 h (Fig. 5B). Cells infected with 10 added bacteria per endothelial cell showed a reduced rate of migration and the infected cells were able to fill the scratch by 96 h. The residual migration observed with infected endothelial cells could be as a result of uninfected or infected cells. To confirm this, migration assays with infected endothelial cells [green fluorescent protein (GFP)-expressing bacteria] were also performed on the coverslips. After fixation at various time-points, visualization of coverslips under the fluorescent microscope revealed the presence of few endothelial cells containing GFP-expressing bacteria in the scratch, indicating some residual migration of infected endothelial cells.

image

Figure 5. Reduced rate of migration of infected endothelial cells.

A and B. Confluent monolayers of uninfected and 24 h-infected endothelial cells were scratched and phase-contrast photographs of the same field at various time-points over 144 h were taken.

A. Uninfected cells.

B. Infected cells.

C. Single uninfected and infected endothelial cells were photographed for 48 h to assess spontaneous migration.

Download figure to PowerPoint

To examine cell motility in more detail, uninfected versus infected cells were compared in random motility assays of endothelial cells plated at low densities. Individual fields were repeatedly photographed over 48 h to examine endothelial cell migratory behavior. There was little or no net movement of the infected cells. In contrast, clear migration activity was observed from the control uninfected cells (Fig. 5C).

Infection inhibits endothelial cell morphogenesis in three-dimensional collagen matrices

To determine if infection with B. bacilliformis alters the ability of endothelial cells to form capillary tubes in three-dimensional extracellular matrix environments, infected versus uninfected endothelial cells were assayed for the ability to undergo endothelial cell morphogenesis. Endothelial cells infected for 4, 8, 12 or 24 h were resuspended in type I collagen gels using a previously described endothelial cell morphogenesis system (Davis and Camarillo, 1996). In control experiments, uninfected endothelial cells formed vacuoles within a few hours of incubation. The vacuoles, subsequently coalesced to form lumens by 24 h (Fig. 6A). After 24 h, branching of lumenal structures was observed, which progressed to form networks of endothelial cell-lined tubes over a 48–72-h period (Fig. 6C).

image

Figure 6. Uninfected and infected endothelial cells in the three-dimensional collagen matrix.

Confluent monolayers of endothelial cells were infected at a multiplicity of 1:100 with B. bacilliformis.

A and C. Uninfected endothelial cells resuspended in a three-dimensional collagen gel system.

B and D. Infected endothelial cells resuspended in a three-dimensional collagen gel system.

The cells were fixed at 24 h (A and B) and 48 h (C and D), stained with toluidine blue and photographed. Arrows in A and B indicate vacuoles, whereas arrows in C indicate endothelial cells lining the lumenal structures. Bar, 100 μm.

Download figure to PowerPoint

Endothelial cells infected for either 12 or 24 h contained numerous small vacuoles in the cytoplasm (Fig. 6B), but the cells did not appear to progress to form appropriate-sized lumenal structures compared with uninfected cells. In addition, there was no indication of branching or networking of the developing infected endothelial cells. After prolonged incubations (48 and 72 h), infected endothelial cells in the collagen gels began to die (Fig. 6D).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Pathogenic bacteria, especially intracellular pathogens, utilize host signalling and response mechanisms as an indispensable component of the infectious process. As a consequence, major changes can occur in the organization of intracellular actin, cell–cell interactions, cell mobility and morphogenesis. In vivo, both B. bacilliformis and B. henselae invade endothelial cells and the infection is associated with both proliferation of endothelial cells and neovascularization. Here, we show that infection with B. bacilliformis results in significant morphological and functional changes in infected endothelial cells in vitro.

Between 16 and 24 h after addition of B. bacilliformis to endothelial cells, the infected cells became spindle-shaped, indicating that a major change in cytoskeletal arrangement occurred during this time. Between 16 and 24 h, the unoriented and diffuse F-actin network of the uninfected cell was replaced with stress fibres, thick F-actin bundles arranged in parallel arrays, as demonstrated by visualization with fluorescein-labelled F-actin phalloidin staining. Prominent and increased numbers of focal adhesions, particularly at the distal ends of the longitudinal axis of the now spindle-shaped cells, were visualized using anti-paxillin (a cytoplasmic protein that localizes in focal adhesions), anti-FAK and anti-phosphotyrosine antibodies. The stress fibres terminate in these focal contacts, as demonstrated by overlay of the phalloidin–fluorescein and the anti-paxillin antibody-stained images (see Fig. 2G–H). It seems probable that the observed change in cell shape is a consequence of the formation of the temporally associated parallel arrays of stress fibres.

Formation of stress fibres in cells is known to be under the control of the signalling protein, Rho (Paterson et al., 1990; Ridley and Hall, 1992; 1994). Inactivation of Rho by C3 exoenzyme from C. botulinum has been widely used to study the function of Rho in various cell types, including endothelial cells (Mohr et al., 1992; Aepfelbacher et al., 1997). We have demonstrated that the formation of stress fibres within the infected cells can be blocked by simultaneous addition of C3 exoenzyme and reduced by delayed addition of C3 exoenzyme (see Fig. 4A–J). This finding indicates that formation of stress fibres in infected cells requires the normal Rho-dependent pathway. Furthermore, inhibition of stress fibre formation in infected endothelial cells by inactivation of Rho-kinase with its specific inhibitor Y-27632 (Uehata et al., 1997) did not inhibit internalization of bacteria, indicating that stress fibre formation is solely an outcome of the infection process.

Activated Rho stimulates the tyrosine phosphorylation of a number of proteins known to localize to focal contacts, FAK, paxillin and p130 (Flinn and Ridley, 1996). Increased tyrosine phosphorylation of FAK and paxillin observed in endothelial cells infected with B. bacilliformis has further substantiated the possibility of activation and involvement of Rho in the infection process.

In addition, the infected cells display loss of cell–cell contacts and loss of a membrane-localized protein, PECAM-1 (see Fig. 1D), which participates in the formation and maintenance of the cell–cell contacts. Geometrically, it may be difficult for an elongated cell to participate effectively in cell–cell contacts with a rounded cell, so the appearance of gaps and the loss of PECAM could be a secondary or mechanical consequence of the change in cell shape. Alternatively, these changes may reflect a more direct alteration in intracellular function.

Infected endothelial cells display diminished spontaneous motility and diminished ability to respond to a signal that would normally induce motility. Wound healing, filling of a scratch on a confluent monolayer (Coomber and Gotlieb, 1990; Aepfelbacher et al., 1997), was used to study migration. Wound healing is entirely dependent on migration of cells up to 30–36 h, as proliferation of endothelial cells in response to the scratch begins only after 36 h and primarily involves cells located distal to the wound (Romer et al., 1994). An early step in induced migration is loss of cell–cell contacts, as was observed for infected endothelial cells. However, infected endothelial cells were found to have a greatly diminished ability to fill a gap by migration.

The impaired migratory ability of infected endothelial cells might result from formation of thick robust stress fibres (see Fig. 1B), which are associated with diminished motility (Burridge and Chrzanowska-Wodnicka, 1996). Formation of stress fibres could interfere with the dynamic reorganization of the actin cytoskeleton needed for protrusion at the front and retraction at the rear of migrating cells (Nobes and Hall, 1999). The increased number of focal adhesions would be expected to have the effect of anchoring the cell, unless the stress fibres and focal adhesions are formed and reformed dynamically. Nobes and Hall (1999) demonstrated that focal adhesions and stress fibres were not required for cell movement; on the contrary, microinjection of constitutively activated Rho into fibroblasts severely inhibited wound closure.

In vivo, infection with either B. bacilliformis or B. henselae is associated with angiogenic skin lesions. Electron microscopic observations in bacillary angiomatosis (Kostianovsky and Greco, 1994) have visualized endothelial cell sprouting, intercellular canalization, an interconnected network of albuminal processes and primitive capillaries with abortive lumina, similar to the morphological events described in primary and secondary angiogenesis (Wagner, 1980). Infected cells in the experiments reported here were studied utilizing a well-characterized in vitro capillary morphogenesis system, in which endothelial cells suspended in collagen gel differentiate to form vacuoles, lumens and a capillary network (Davis and Camarillo, 1996). B. bacilliformis infection of endothelial cells led to the early death of these cells in this system and, as a consequence, eliminated their ability to undergo morphogenesis when placed in the three-dimensional collagen matrices. In vivo, the cells that effectively participate in angiogenesis are probably uninfected endothelial cells, responding to angiogenesis signals present in the verruga.

What purpose could these alterations induced by B. bacilliformis in endothelial cells serve during the actual disease process? The presumed infectious cycle of B. bacilliformis involves introduction of bacteria into the blood from the infected proboscis of a sand fly, multiplication in the red blood cells of the host, transference of the focus of infection from the blood to skin endothelial cells in the form of a verruga, with transmission of the bacteria back to sand flies during the haematic and verrugal phase. As the verruga can persist for up to 1 year, this persistent focus of bacterial infection provides for a long-term carrier state and for reinoculation of sand flies.

The verruga must be characterized by local immunological incompetence to account for its longevity, especially taking into consideration the fact that survivors of B. bacilliformis infection have life-long immunity and that B. bacilliformis is highly immunogenic. Bacteria are found both inside and outside the endothelial cells, so that an intracellular location is probably not the sole protective mechanism available to B. bacilliformis. The verruga is also characterized by a great stimulation of endothelial cell growth and formation of neocapillaries, some functional, some not. The functional capillaries would provide B. bacilliformis with a location suitable for reinfecting a sand fly and also a route from the verruga into the general circulation. The non-functional capillaries could provide B. bacilliformis with a closed, privileged sanctuary that, in some ways, could provide more protection than an intracellular location.

Some of the cellular changes seen after infection of endothelial cells, including an elongated spindle shape, appearance of coalesced vacuoles and loss of cell–cell contacts, are also seen in activated endothelial cells actively involved in angiogenesis. However, it is difficult to believe that infected cells in vivo could be directly forming neocapillaries if they have properties similar to the infected endothelial cells studied here. Infected endothelial cells have increased focal contacts, reduced spontaneous motility and a reduced ability to respond to a stimulus, which would normally induce motility. In addition, they are completely defective in the ability to form capillary tubes in a three-dimensional collagen matrix, but the significance of this is not clear, as this impairment appears in part to be related to decreased viability of the infected cells in the three-dimensional collagen matrix. If these alterations in endothelial cells occur in vivo as well, it seems more probable that the infected cells, instead of directly participating in formation of capillaries, might, directly or indirectly, trigger production of angiogenic factors that would induce normal, uninfected endothelial cells to form new capillaries. Addition of apoptosis-inducing agents to endothelial cells is known to stimulate synthesis of vascular endothelial growth factor (VEGF) which, in turn, protects the cells from apoptosis (Marx et al., 1999). Stimulation of the endothelial cells with VEGF results in tyrosine phosphorylation and recruitment of paxillin and FAK to new focal contacts (Abedi and Zachary, 1997). We have observed that infection with B. bacilliformis leads to increased levels of VEGF mRNA (unpublished observations) and increased tyrosine phosphorylation of paxillin and FAK. Synthesis of increased levels of VEGF might prevent the infected cells from undergoing apoptosis prematurely or render infected endothelial cells more resistant to immunological mechanisms for inducing apoptosis. However, using VEGF-specific ELISA, we could not detect VEGF in infected endothelial cell extracts or in the growth medium of infected endothelial cells, so the significance of increased levels of VEGF mRNA remains unclear. Other groups (Maeno et al., 1999) have also failed to detect VEGF in cell extracts and culture supernatants of B. henselae.

Stages of angiogenesis, including endothelial cell migration and proliferation, lumen formation, synthesis of matrix components and remodelling of matrix to provide a framework for tube formation, are tightly regulated and involve more than one angiogenic factor at each stage, with additional modulation by many other factors. Many pathogenic bacteria, including Chlamydia pneumoniae, induce synthesis of soluble factors or cytokines by endothelial cells that stimulate smooth muscle proliferation (Coombes and Mahony, 1999), procoagulant activity (Fryer et al., 1997) and haematopoietic growth functions (Koenig et al., 1994). In addition, probably to redirect the host response along a pathway favourable to the bacteria, bacterial infection can lead to activation of signal transduction pathways (Krull et al., 1999) and increased expression of endothelial cell membrane proteins, including E-selectin, ICAM-I and VCAM-1 (Kaukoranta-Tolvanen et al., 1996). In the case of endothelial cells infected with B. bacilliformis, one could speculate that the infected endothelial cells release signals that stimulate normal endothelial cells to participate in capillary formation. The presence of other inflammatory cells in the lesions also suggests the possible involvement of factors secreted by macrophages and lymphocytes in response to the infection. In this way, the bacteria could ensure themselves a supply of infectable cells, as well as the capillary network needed to re-establish contact with the sand fly vector.

Bartonella spp. can be regarded as emerging pathogens in the sense that they are probably more dangerous to humans in the future than they have been in the past. However, it appears that the pathogenic mechanisms utilized by these organisms to replicate in endothelial cells and transmit themselves are already quite complex and sophisticated. Further investigation of the mechanism by which Bartonella induces angiogenic responses should lead to new insights into the molecular regulation of the angiogenic response.

Experimental procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Bacterial strains

B. bacilliformis was obtained as previously described (Benson et al., 1986) and routinely grown at 28°C in PBS (phosphate-buffered saline) over brain–heart infusion (BHI) agar plates containing 10% defibrinated sheep blood (Dickinson Microbiology System). Escherichia coliβ2150 carrying an expression vector pCD353 (containing GFP as an expression marker) was kindly provided by Dr Christoph Dehio (Max-Plank-Institut für Biologie). E. coli expressing C3 exoenzyme gene from Clostridium botulinum type D strain 1873 cloned into pGEX expression vector system was kindly provided by Dr Bradley McIntyre (University of Texas, M.D. Anderson Cancer Center).

Antibodies and reagents

Mouse monoclonal antibodies directed against paxillin, FAK and phosphotyrosine (PY20) were from Transduction laboratories (A Becton Dickinson company). Polyclonal antibodies against PECAM-1 were from Santa Cruz Biotechnology, Fluoroscein isothiocyanate (FITC)-labelled F-actin phalloidin was from Molecular Probes. VEGF and bFGF (basic fibroblast growth factor) were from Upstate Biotechnology (Lake Placid). Rho-kinase inhibitor Y-27632 was a gift from Yoshitomi Pharmaceutical industries.

Transfer of pCD353 into B. bacilliformis by conjugation

Conjugation to generate fluorescent bacteria was performed essentially as described (Dehio & Meyer, 1997). Briefly, three parental matings were performed. Plasmid pCD353 from E. coli was transferred to B. bacilliformis with the helper plasmid pRK2013 present in E. coliβ2150. E. coli strains carrying an expression vector pCD353 (containing GFP gene as an expression marker) and pRK2013 were grown overnight in 10 ml of Luria–Bertani (LB) medium containing kanamycin (30 μg ml−1) and diaminopimelic acid (DAP) at 37°C with shaking. E. coli cultures were harvested and cells were resuspended in 100 μl of PBS; a 2 d-old culture of B. bacilliformis was also harvested and resuspended in 100 μl of PBS. After mixing, the bacteria were dotted on nitrocellulose discs on a BHI blood agar plate overlaid with DAP and incubated for 8 h at 30°C. After conjugation, bacterial cells were collected from nitrocellulose discs by scrapping and washing the discs with PBS. The suspension was plated on the BHI blood agar plates overlaid with kanamycin (60 μg ml−1) and isopropyl-β-d-thiogalactopyranoside (IPTG) (250 μM). Plates were incubated for at least 7–10 d at 28°C to obtain transconjugants.

Infection of human endothelial cells with B. bacilliformis

HUVECs (Clonetics) were propagated between passages 2 and 7 as described (Davis and Camarillo, 1996). HUVECs were seeded on glass coverslips, tissue culture flasks and in six-well plates coated with gelatin (1 mg ml−1), in M199 containing 20% fetal bovine serum, heparin and bovine hypothalamic extract (Macaig et al., 1979) for infection and/or immunofluorescent staining. For infection, transconjugated fluorescent B. bacilliformis from a 2 d-old culture induced with 250 μM IPTG was used throughout the study. Cultures were harvested and washed several times with PBS and resuspended in plain M199. The cells were infected with 100 bacteria per cell. Infected monolayers were maintained at 37°C with 5% CO2.

Immunofluorescence staining of endothelial cells

Infection was terminated at the indicated time-points by several washes of PBS and, subsequently, cells were fixed with 4% paraformaldehyde in PBS for 30 min at room temperature. Cells were washed several times with PBS, post fixation. Uninfected and infected monolayers on glass coverslips were blocked with two washes of Tris-glycine buffer and the cells were permeabilized with 0.5% Triton X-100 for 15 min. Coverslips were then blocked overnight at 4°C in PBS containing 0.1% Triton X-100, 1% BSA, 1% normal rabbit serum and 0.2% sodium azide. After overnight incubation, primary antibody was added to the same solution and incubated for 1 h at room temperature. Coverslips were washed three times with PBS and further incubated with secondary antibody (Rabbit anti-mouse IgG-TRITC). Coverslips were again washed with PBS and mounted on the slides using a fluorescent mounting medium (DAKO); fluorescent images were captured using an Axiophot II digital imaging system.

Immunoprecipitation

Uninfected and infected endothelial cells (2, 12 or 24 h after the infection with B. bacilliformis) were lysed in 1 ml of ice-cold lysis buffer containing 10 mM Tris-Cl pH 7.6, 50 mM NaCl, 5 mM EDTA, 30 mM sodium pyrophosphate, 1 mM orthovandate, 1% NP-40, 50 mM NaF, 10 μg ml−1 leupeptin, 1 μg ml−1 pepstatin and 1 mM phenylmethylsulphonyl fluoride. Lysates were clarified by centrifugation at 16 000 g for 10 min. Immunoprecipitation was carried out using Dynabeads Protein A (Dynal), according to the manufacturer's instructions. Mouse monoclonal antibodies to paxillin, FAK and phosphotyrosine were immobilized to protein A on the magnetic beads and cross-linked using DMP (dimethyl pimelimidate) in 0.2 M triethanolamine pH 9.0. Cross-linked antibodies to the protein A magnetic beads were used to capture the antigens from the endothelial cell lysates. Immunoprecipitates were washed several times with PBS, extracted in 1× sodium dodecyl sulphate (SDS) buffer, then separated on SDS-polyacrylamide gel and analysed using Western blot analysis. Anti-phosphotyrosine antibody was used for immunoblotting with the immunoprecipitates of anti-paxillin and FAK, whereas anti-paxillin and anti-FAK antibodies were used to probe anti-phosphotyrosine immunoprecipitates.

C3 exoenzyme purification and ADP ribosylation

The gene for C3 exoenzyme from C. botulinum type D strain 1873 (Popoff et al., 1991) was cloned into the pGEX expression vector system (Pharmacia Biotech) to generate a glutathione-S-transferase fusion protein (pGEX2T-C3) (Dillon and Feig, 1995). C3 exoenzyme was purified as described (Udagawa and McIntyre, 1996). Briefly, a single colony of E. coli (strain JM109) transformed with pGEX2T-C3 DNA was inoculated into 200 ml of LB medium containing 100 μg ml−1 ampicillin and grown overnight at 37°C on a shaker. This culture was then added to 1800 ml of LB containing ampicillin, shaken at 37°C for 1 h, induced with IPTG (100 μg ml−1) and the culture grown at 37°C on a shaker for an additional 7 h. The bacteria were pelleted at 5000 g and the cells were lysed in 80 ml of ice-cold PBS containing 1 mg ml−1 lysozyme, 1% Triton X-100, 25% sucrose, 1 mM EDTA, 5 mM β-mercaptoethanol and 1 mM PMSF for 30 min on ice, with occasional shaking. The slurry was then passed through a French press and pancreatic DNase I (Sigma) was added to a final concentration of 100 μg ml−1, stirred for an additional 20 min at 4°C and centrifuged at 10 000 g for 10 min at 4°C.

The lysate was added to 2 ml of washed glutathione-agarose beads (Pharmacia) and mixed gently at 4°C for 1 h. The beads were centrifuged at 500 g for 5 min, washed four times with 50 ml of PBS containing 1% Triton X-100 and washed three times with 25 ml of 50 mM Tris-HCl (pH 7.5), 150 mM NaCl and 2.5 mM CaCl2; the beads were then resuspended in 1 ml of Tris buffer. Purified alpha thrombin (60 National Institutes of Health units) was added and incubated for 16 h at 4°C to cleave the fusion protein. The beads were centrifuged at 1000 g for 2 min, washed twice with 1 ml of the Tris solution and concentrated using a Centricon-10 (Amicon). The product ran as a single band at approximately 25 kDa on a 12% SDS-polyacrylamide gel. To determine the concentration required for in vivo inactivation of Rho, endothelial cells were treated with 10, 20, 50 and 100 μg ml−1 C3 exoenzyme and the extent of ADP ribosylation of Rho in vivo in endothelial cells was determined by subsequent in vitro[32P]-ADP ribosylation. A concentration of 100 μg ml−1 C3 exoenzyme was selected for the study. Cells were incubated with 100 μg ml−1 C3 exoenzyme for 0 (simultaneous addition of bacteria and C3 exoenzyme), 4, 8, 12 or 16 h after infection with B. bacilliformis.

Wound-healing and spontaneous migration assays

For migration experiments, endothelial cells were subcultured in six-well culture dishes 24 h prior to the experiment. Monolayers with 80–90% confluence were infected with B. bacilliformis for 24 h. Uninfected and infected monolayers were scratched by dragging a sterile yellow pipette tip across the well to create a cell-free path. Scratched monolayers were washed several times to remove floating cells and further incubated at 37°C. The same field was photographed using phase-contrast microscope at different time intervals to monitor the migration of endothelial cells. For spontaneous migration, uninfected and 24 h-infected endothelial cells were trypsinized and plated at a very low density to achieve a single-cell attachment. The same fields were photographed after attachment at different time intervals and the migration pattern of individual cells was traced and compared.

Three-dimensional collagen assay for capillary morphogenesis

The three-dimensional collagen assay for capillary lumen formation was performed as described (Davis and Camarillo, 1996). Briefly, endothelial cells infected for 12 or 24 h with bacteria and uninfected cells were trypsinized and allowed to recover for 20 min with M199 containing 10% serum at room temperature. The cells were washed to remove serum and resuspended in serum-free medium at a concentration of 1 × 106 cells ml−1. Purified rat-tail collagen was added to tubes to which 10× concentrated M199 medium and NaOH were added as a mixture at 0°C. After thorough mixing, the cells were added at 1 × 106 ml−1 with a final collagen concentration of 3.75 mg ml−1. The cell–collagen mix was then added at 25 μl per well in 4.5 mm diameter microwells. The collagen was allowed to gel and was equilibrated for 30 min at 37°C in a CO2 incubator prior to the addition of culture medium. The serum-free culture medium consisted of Medium 199 with reduced serum supplement II, bFGF (40 ng ml−1), VEGF (40 ng ml−1), phorbol ester (50 ng ml−1) and 50 μg ml−1 ascorbic acid in a 100 μl volume per 4.5 mm well, allowed to incubate for 72 h to allow for capillary lumen formation. The gels were fixed at 24, 48 or 72 h and stained with toluidine blue and photographed.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

This research was supported by the Tom and Jean McMullin Chair in Genetics (G. M. Ihler) and by NIH HL grant 59373 (G. E. Davis). We thank Christoph Dehio for providing bacterial strains and plasmids and Yoshitomi Pharmaceutical Industries, Japan for providing Rho-kinase inhibitor Y-27632. We are indebted to Clay and Rola Barhoumi for help in confocal microscopy and Axiophot II digital image analysis. Support and encouragement provided by other members of the Ihler and Davis laboratories during this study is gratefully acknowledged. Portions of these results were presented at the 1st international conference on ′Bartonella as emerging pathogens', March 5–7, 1999, at Tubingen, Germany, and ′A cell biology approach to microbial pathogenesis', April 25–28th 1999, at Portland, Oregon, USA.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • Abedi, H. & Zachary, I. (1997) Vascular endothelial growth factor stimulates tyrosine phosphorylation and recruitment to new focal adhesions of focal adhesion kinase and paxillin in endothelial cells. J Biol Chem 272: 1544215451.
  • Aepfelbacher, M., Essler, M., Huber, E., Sugai, E.M. & Weber, P.C. (1997) Bacterial toxins block endothelial wound repair. Evidence that Rho GTPases control cytoskeletal rearrangements in migrating endothelial cells. Arterioscler Thromb Vasc Biol 17: 16231629.
  • Arias-Stella, J., Lieberman, P.H., Garcia-Caceres, U., Erlandson, R.A., Kruger, H. & Arias-Stella, J., Jr (1987) Verruga peruana mimicking malignant neoplasms. Am J Dermatopathol 9: 279291.
  • Batterman, H.J., Peek, J.A., Loutit, J.S., Falkow, S. & Tompkins, L.S. (1995) Bartonella henselae and Bartonella quintana adherence to and entry into cultured human epithelial cells. Infect Immun 63: 45534556.
  • Benson, L.A., Kar, S., McLaughlin, G. & Ihler, G.M. (1986) Entry of Bartonella bacilliformis into erythrocytes. Infect Immun 54: 347353.
  • Burridge, K. & Chrzanowska-Wodnicka, M. (1996) Focal adhesions, contractility, and signaling. Annu Rev Cell Dev Biol 12: 463518.
  • Coomber, B.L. & Gotlieb, A.I. (1990) In vitro endothelial wound repair. Interaction of cell migration and proliferation. Arteriosclerosis 10: 215222.
  • Coombes, B.K. & Mahony, J.B. (1999) Chlamydia pneumoniae infection of human endothelial cells induces proliferation of smooth muscle cells via an endothelial cell-derived soluble factor (s). Infect Immun 67: 29092915.
  • Davis, G.E. & Camarillo, C.W. (1996) An alpha 2 beta 1 integrin-dependent pinocytic mechanism involving intracellular vacuole formation and coalescence regulates capillary lumen and tube formation in three-dimensional collagen matrix. Exp Cell Res 224: 3951.DOI: 10.1006/excr.1996.0109
  • Dehio, C. & Meyer, M. (1997) Maintenance of broad-host-range incompatibility group P and group Q plasmids and transposition of Tn5 in Bartonella henselae following conjugal plasmid transfer from Escherichia coli. J Bacteriol 179: 538540.
  • Dehio, C., Meyer, M., Berger, J., Schwarz, H. & Lanz, C. (1997) Interaction of Bartonella henselae with endothelial cells results in bacterial aggregation on the cell surface and the subsequent engulfment and internalization of the bacterial aggregate by a unique structure, the invasome. J Cell Sci 110: 21412154.
  • Dillon, S.T. & Feig, L.A. (1995) Purification and assay of recombinant C3 transferase. Methods Enzymol 256: 174184.
  • Flinn, H.M. & Ridley, A.J. (1996) Rho stimulates tyrosine phosphorylation of focal adhesion kinase, p130 and paxillin. J Cell Sci 109: 11331141.
  • Fryer, R.H., Schwobe, E.P., Woods, M.L. & Rodgers, G.M. (1997) Chlamydia species infect human vascular endothelial cells and induce procoagulant activity. J Investig Med 45: 168174.
  • Garcia, F.U., Wojta, J., Broadlay, K.N., Davidson, J.M. & Hoover, R.L. (1990) Bartonella bacilliformis stimulates endothelial cells in vitro and is angiogenic in vivo. Am J Pathol 136: 11251135.
  • Garcia, F.U., Wojta, J. & Hoover, R.L. (1992) Interactions between live Bartonella bacilliformis and endothelial cells. J Infect Dis 165: 11381141.
  • Hill, E.M., Raji, A., Valenzuela, M.S., Garcia, F. & Hoover, R.L. (1992) Adhesion to and invasion of cultured human cells by Bartonella bacilliformis. Infect Immun 60: 40514058.
  • Kaukoranta-Tolvanen, S.S., Ronni, T., Leinonen, M., Saikku, P. & Laitinen, K. (1996) Expression of adhesion molecules on endothelial cells stimulated by Chlamydia pneumoniae. Microb Pathog 21: 407411.DOI: 10.1006/mpat.1996.0071
  • Koenig, A., Yakisan, E., Reuter, M., Huang, M., Sykora, K.W., Corbacioglu, S. & Welte, K. (1994) Differential regulation of stem cell factor mRNA expression in human endothelial cells by bacterial pathogens: an in vitro model of inflammation. Blood 83: 28362843.
  • Kostianovsky, M. & Greco, M.A. (1994) Angiogenic process in bacillary angiomatosis. Ultrastruct Pathol 18: 349355.
  • Krull, M., Klucken, A.C., Wuppermann, F.N., Fuhrmann, O., Magerl, C. & Seybold, S. et al. (1999) Signal transduction pathways activated in endothelial cells following infection with Chlamydia pneumoniae. Immunol 162: 48344841.
  • Maciag, T., Cerundolo, J., Ilsley, S., Kelley, P.R. & Forand, R. (1979) An endothelial cell growth factor from bovine hypothalamus: identification and partial characterization. Proc Natl Acad Sci USA 76: 56745678.
  • Maeno, N., Oda, H., Yoshiie, K., Wahid, M.R., Fujimura, T. & Matayoshi, S. (1999) Live Bartonella henselae enhances endothelial cell proliferation without direct contact. Microbial Pathog 27: 419427.
  • Marx, P.T., Mulder, A.B., Van Den Bergh, F.A., Overbeeke, R., Haanen, C. & Vermes, I. (1999) Apoptosis inducers endotoxin and Fas-ligation enhance the expression of vascular endothelial growth factor in human endothelial cells. Endothelium 6: 335340.
  • Mohr, C., Koch, G., Just, I. & Aktories, K. (1992) ADP-ribosylation by Clostridium botulinum C3 exoenzyme increases steady-state GTPase activities of recombinant rhoA and rhoB proteins. FEBS Lett 297: 9599.
  • Nobes, C.D. & Hall, A. (1999) Rho GTPases control polarity, protrusion, and adhesion during cell movement. J Cell Biol 144: 12351244.
  • Palmiri, J., Teysseire, N., Dussert, C. & Raoult, D. (1996) Image cytometry and topographical analysis of proliferation of endothelial cells in vitro during Bartonella (Rochalimaea) infection. Anal Cell Path 11: 1330.
  • Paterson, H.F., Self, A.J., Garrett, M.D., Just, I., Aktories, K. & Hall, A. (1990) Microinjection of recombinant p21rho induces rapid changes in cell morphology. J Cell Biol 111: 10011007.
  • Popoff, M.R., Hauser, D., Boquet, P., Eklund, M.W. & Gill, D.M. (1991) Characterization of the C3 gene of Clostridium botulinum types C and D and expression in Escherichia coli. Infect Immun 59: 36733679.
  • Ridley, A.J. & Hall, A. (1992) The small GTP-binding protein Rho regulates the assembly of focal adhesions and actin stress fibres in response to growth factors. Cell 70: 389399.
  • Ridley, A.J. & Hall, A. (1994) Signal transduction pathways regulating Rho-mediated stress fibre formation: requirement for a tyrosine kinase. EMBO J 13: 26002610.
  • Romer, L.H., McLean, N., Turner, C.E. & Burridge, K. (1994) Tyrosine kinase activity, cytoskeletal organization, and motility in human vascular endothelial cells. Mol Biol Cell 5: 349361.
  • Schneider, T., Ullrich, R., Schmitt-Graff, A., Bergs, C., Reiterer, L. & Dissmann, T. et al. (1993) Bacillary angiomatosis in German patient with AIDS. Clin Investig 72: 5054.
  • Tappero, J.W., Koehler, J.E., Berger, T.G., Cockerell, C.J., Lee, T.H. & Busch, M.P. et al. (1993) Bacillary angiomatosis and bacillary splenitis in immunocompetent adults. Ann Inter Med 118: 363365.
  • Udagawa, T. & McIntyre, B.W. (1996) ADP-ribosylation of the G protein Rho inhibits integrin regulation of tumor cell growth. J Biol Chem 271: 1254212548.
  • Uehata, M., Ishizaki, T., Satoh, H., Ono, T., Kawahara, T. & Morishita, T. et al. (1997) Calcium sensitization of smooth muscle mediated by a Rho-associated protein kinase in hypertension. Nature 389: 990994.
  • Wagner, R.D. (1980) Endothelial cell embryology and growth. Adv Microcirc 9: 4575.
  • Webster, G.F., Cockerell, C.J. & Friedman-Kien, A.E. (1992) The clinical spectrum of bacillary angiomatosis. Br J Dermatol 126: 535541.
  • Williams-Bouyer, N.M. & Hill, E.M. (1999) Involvement of host cell tyrosine phosphorylation in the invasion of Hep2 cells by Bartonella bacilliformis. FEMS Microbiol Lett 171: 191201.DOI: 10.1016/s0378-1097(98)00605-3