Bartonella henselae causes the vasculoproliferative disorders bacillary angiomatosis and peliosis probably resulting from the release of vasculoendothelial growth factor (VEGF) from infected epithelial or monocytic host cells. Here we demonstrate that B. henselae in addition to VEGF induction was also capable of inhibiting the endogenous sucide programme of monocytic host cells. Our results show that B. henselae inhibits pyrrolidine dithiocarbamate (PDTC)-induced apoptosis in Mono Mac 6 cells. B. henselae was observed to be present in a vacuolic compartment of Mono Mac 6 cells. Direct contact of B. henselae with Mono Mac 6 cells was crucial for inhibition of apoptosis as shown by the use of a two-chamber model. Inhibition of apoptosis was paralleled by diminished caspase-3 activity which was significantly reduced in PDTC-stimulated and B. henselae-infected cells. The anti-apoptotic effect of B. henselae was accompanied by (i) the activation of the transcription factor NF-κB and (ii) the induction of cellular inhibitor of apoptosis proteins-1 and -2 (cIAP-1, -2). Our results suggest a new synergistic mechanism in B. henselae pathogenicity by (i) inhibition of host cell apoptosis via activation of NF-κB and (ii) induction of host cell VEGF secretion.
Microbe–macrophage interactions play a central role in the pathogenesis of many bacterial infections. Several extracellular pathogens induce apoptosis in macrophages. For instance, delivery of invasion plasmid antigen B (IpaB; Shigella flexneri ) and Salmonella invasion protein B (SipB; Salmonella spp.) to the macrophage cytoplasm via a type III secretion process results in the activation of the pro-apoptotic caspase-1 (Navarre and Zychlinsky, 2000). Yersinia spp. induce apoptosis by Yersinia outer protein (Yop) P/J-mediated inhibition of nuclear factor-κB (NF-κB; Ruckdeschel et al., 1998). In contrast, the anti-apoptotic strategies of the intracellular pathogens Chlamydia pneumoniae (Fischer et al., 2001), Brucella suis (Gross et al., 2000), Rickettsia rickettsii (Clifton et al., 1998) and Ehrlichia spp. (Yoshiie et al., 2000) are generally aimed at avoiding the loss of their cellular habitat. However, the molecular mechanisms underlying the inhibition of apoptosis by bacterial pathogens have been analysed to a far less extent.
Bartonella henselae is a fastidious, slow-growing, facultative intracellular bacterium causing cat scratch disease (CSD), bacillary angiomatosis (BA) and peliosis hepatis (PH; Anderson and Neuman, 1997). CSD is usually a benign and self-limiting disease characterized by lymphadenopathy related to a cat scratch or bite. In immunocompromised individuals, B. henselae causes tumorous proliferations of endothelial cells in the skin and internal organs, referred to as BA and PH respectively (Schwartzman, 1992; Slater et al., 1992; Adal et al., 1994). B. henselae is detectable in these vasculoproliferative lesions and antibiotic treatment results in complete regression (Anderson and Neuman, 1997). There are at least three different mechanisms by which Bartonella may cause vasculoproliferative disorders: (i) directly triggering endothelial cell proliferation, probably by proteins located in the bacterial cell wall (Garcia et al., 1990; Conley et al., 1994; Maeno et al., 1999), (ii) inducing the secretion of vasculoproliferative cytokines from infected host cells (Kempf et al., 2001; Resto-Ruiz et al., 2002) and (iii) inhibition of apoptosis of endothelial cells (Kirby and Nekorchuk, 2002).
For all of these reasons, we were interested in whether B. henselae might be able to inhibit apoptosis in monocytes. To investigate this issue, we used the Mono Mac 6 cell line (Ziegler-Heitbrock et al., 1988) and the apoptosis-inducing reagent pyrrolidine dithiocarbamate (PDTC) which inhibits NF-κB mobilization in human monocytes (Ziegler-Heitbrock et al., 1993). Similar models of apoptosis inhibition have been described using the intracellular pathogen Chlamydophila pneumoniae (Wahl et al., 2001). By means of a co-culture infection model, we found that B. henselae inhibits PDTC-induced apoptosis in Mono Mac 6 cells. This inhibition of host cell apoptosis was paralleled by the cellular presence of B. henselae and also by a diminished caspase-3 activity which was significantly reduced in PDTC-stimulated and B. henselae-infected cells. VEGF does not appear to be involved in the inhibition of apoptosis by B. henselae, although Mono Mac 6 cells secrete VEGF upon infection. Furthermore, the anti-apoptotic effect was accompanied by (i) the activation of the transcription factor NF-κB and (ii) induction of cellular inhibitor of apoptosis proteins-1 and -2 (cIAP-1, -2). Our results extend the already proposed two-step mechanism of B. henselae pathogenicity (Kempf et al., 2002) whereby infection with B. henselae leads to the inhibition of apoptosis and a subsequent prolonged synthesis of VEGF by B. henselae-infected monocytes.
Bartonella henselae inhibits apoptosis in Mono Mac 6 cells
Inhibition of apoptosis was analysed by treating Mono Mac 6 cells with the NF-κB inhibitor PDTC (200 µM) as described earlier (Della Ragione et al., 2000; Wahl et al., 2001) and/or infecting with B. henselae Marseille (Drancourt et al., 1996). Apoptosis and viability were quantified via TUNEL reaction and MTS viability assays respectively (see Experimental procedures). PDTC treatment of Mono Mac 6 cells led to 75% apoptotic cells as revealed by TUNEL reaction. The number of apoptotic cells was significantly reduced to 24% in the presence of B. henselae(Fig. 1A and B). MTS viability testing revealed consistent results illustrating that PDTC decreased the viability of Mono Mac 6 cells to 52% which was restored to 100% in the presence of B. henselae (Fig. 1C).
The optimal concentration of PDTC was determined to be 200 µM, although B. henselae was also capable of inhibiting apoptosis at concentrations ranging from 50 to 500 µM (Fig. 2A). Inhibition of apoptosis by B. henselae was dose dependent because an increase in the multiplicity of infection (moi; 10, 100, 200) resulted in a lower number of apoptotic (TUNEL-positive) cells and increased cell viability (Fig. 2B and C). Moreover, inhibition of PDTC-triggered apoptosis was achieved by infection not only with B. henselae Marseille, but also with B. henselae Houston-1 and Bartonella quintana Toulouse (Fig.S1).
Expression of Bartonella adhesin A is crucial for inhibition of apoptosis
Expression of B. henselae adhesin A (BadA) is crucial for host cell adherence and triggering of a proangiogenic host cell response (T. M. Riess et al., submitted). We therefore analysed whether B. henselae BadA–, a Tn-903 transposon mutant (Riess et al., 2003), was able to inhibit apoptosis in Mono Mac 6 cells. Inhibition of apoptosis was significantly reduced when Mono Mac 6 cells were infected with B. henselae BadA–(Fig. 3A). BadA expression also correlated with an approximately threefold decrease in adherence to Mono Mac 6 cells 1 h after infection (B. henselae wild type 2.44% versus B. henselae BadA– 0.80% of the inoculum). From this, we conclude that expression of BadA is involved in the inhibition of apoptosis by B. henselae.
To further elucidate whether the apoptosis inhibiting effect of B. henselae was specifically linked to bacterial viability, B. henselae were killed with gentamicin and co-cultivated with PDTC-treated Mono Mac 6 cells for 24 h, and cell viability was subsequently quantified. Results illustrate that gentamicin-killed B. henselae inhibited PDTC-triggered apoptosis similar to viable bacteria (Fig. 3B) indicating that preformed bacterial factors such as surface molecules might be crucial for the inhibition of apoptosis. Data are consistent with those obtained using heat-killed B. henseale showing a diminished but significant inhibition of apoptosis compared with viable bacteria (Fig. 3C).
Inhibition of apoptosis is linked to the cellular presence of B. henselae
To investigate whether the inhibition of apoptosis is directly linked to the cellular presence of B. henselae, colocalization of apoptosis (TUNEL reaction) and B. henselae (immunostaining) was quantified microscopically (Fig. 4A) by counting ≈ 500 PDTC-treated (200 µM) B. henselae-infected Mono Mac 6 cells over at least 20 random microscopic fields. Twenty-four hours after co-cultivation, 86% of all cells were viable and 14% were apoptotic. The vast majority (94%) of the viable cells were infected. Inhibition of apoptosis was not detectable (Fig. 4B) when B. henselae was physically separated from Mono Mac 6 cells by filter dishes (pore size 0.2 µm). From these data, we conclude that inhibition of apoptosis is directly linked to the cellular presence of B. henselae.
To further quantify the interaction of B. henselae with Mono Mac 6 cells, the number of adherent and intracellular B. henselae was assessed. The data revealed that 30 min after infection 10.0 ± 5.8 × 105 bacteria were adherent (≈ 10 bacteria per cell), and 2.3 ± 0.6 × 104 bacteria could still be detected intracellularly 24 h after infection (see also Fig.S2). When cells were treated with increasing concentrations of cytochalasin-D (cyto-D), the number of intracellular bacteria decreased in a dose-dependent manner (Fig. 5A). However, this did not influence the inhibition of apoptosis by B. henselae (Fig. 5B) indicating that intracellular presence is not required for inhibition of apoptosis.
Bartonella henselae is located in a vacuolic compartment in Mono Mac 6 cells
Rickettsia spp. are phylogenetically closely related to Bartonella spp. It is known that Rickettsia spp. inhibit apoptosis of host cells (Clifton et al., 1998) and that once inside the host cell these bacteria lyse the phagosomal membrane to get access to the cytoplasm (Walker, 1984). Therefore, we wanted to elucidate the subcellular location of B. henselae in Mono Mac 6 cells 24 h after infection using transmission electron microscopy (Fig. 6). Consistent with results from TUNEL and MTS assays, apoptotic cells were not detected in the control group. In contrast, morphologic analysis of the cellular architecture of PDTC-treated Mono Mac 6 cells revealed typical apoptotic changes including nuclear condensation, segregation of nuclear chromatin and formation of apoptotic bodies. Apoptotic cells were not detectable when infected with B. henselae. Additionally, the morphological signs of apoptosis were not observed when PDTC-treated Mono Mac 6 cells were co-cultivated with B. henselae. Microscopic enlargements of the subcellular regions revealed that B. henselae is not located in the cytoplasm but was surrounded by a membrane, suggesting that in Mono Mac 6 cells the intracellular habitat of B. henselae is within a vacuolic compartment.
Bartonella henselae does not inhibit apoptosis via autocrine or paracrine VEGF effects
There is increasing evidence that VEGF might be involved in the inhibition of apoptosis even in the presence of bacterial compounds such as lipopolysaccharides (Dias et al., 2002; Munshi et al., 2002). As it is known that B. henselae is capable of inducing VEGF in host cells (Kempf et al., 2001; Resto-Ruiz et al., 2002), we wanted to elucidate whether B. henselae-triggered VEGF production might be involved in the inhibition of apoptosis via autocrine or paracrine mechanisms. For this purpose, we tested whether B. henselae is able to trigger VEGF synthesis in Mono Mac 6 cells. The data reveal that B. henselae induces VEGF secretion in Mono Mac 6 cells and that in the presence of PDTC, B. henselae-infected Mono Mac 6 cells still secreted significantly higher amounts of VEGF compared with non-infected control cells (Fig. 7A). The use of VEGF-neutralizing antibodies did not influence the inhibition of PDTC-triggered apoptosis by B. henselae (Fig. 7B) indicating that VEGF is not involved in inhibition of apoptosis by B. henselae.
Inhibition of PDTC-induced apoptosis by B. henselae is accompanied by activation of NF-κB
Infection of Mono Mac 6 cells with B. henselae resulted in an infection rate of ≈ 50% and only a low percentage of apoptotic cells (Fig. 4). A constitutive level of NF-κB binding activity was detected in non-infected Mono Mac 6 cells (Fig. 8, lane 1), in agreement with previous studies (Frankenberger et al., 1994; Wahl et al., 2001). Pretreatment of Mono Mac 6 cells with the apoptosis inducer and NF-κB inhibitor PDTC decreased NF-κB binding activity (Fig. 8, lane 4), and was accompanied by a large number of apoptotic monocytes. B. henselae-induced activation of NF-κB (Fig. 8, lane 3) was not blocked by PDTC (Fig. 8, lane 5), suggesting that the remaining NF-κB binding activity still protects infected monocytes from PDTC-induced cell death.
Bartonella henselae infection inhibits PDTC-induced procaspase-3 processing and DEVDase activity
To examine whether inhibition of PDTC-induced apoptosis by B. henselae was also reflected on a molecular level, procaspase-3 processing was monitored by Western blot analysis because caspase-3 is the central executioner caspase of the apoptotic cell death programme (Porter and Janicke, 1999). As shown in Fig. 9A, the proteolytically processed p17 fragment of caspase-3 was detected in PDTC-treated Mono Mac 6 cells, whereas lysates of the control cells displayed only the p32 proform. Infection with B. henselae was not associated with any procaspase-3 processing but completely inhibited the PDTC-induced procaspase-3 cleavage so that the p17 fragment was no longer detectable.
To confirm that procaspase processing was associated with caspase activation, caspase-3-like activity was measured in a fluorometric assay using the fluorogenic peptide substrate Ac-DEVD-AMC (Fig. 9B). Treatment of Mono Mac 6 cells with PDTC induced almost a fivefold increase of DEVDase activity compared with the untreated control. Infection with B. henselae slightly enhanced the basal DEVDase activity (approximately twofold) and significantly diminished the PDTC-induced DEVDase activity to that of B. henselae-infected controls. These data clearly demonstrate that infection of Mono Mac 6 cells with B. henselae efficiently inhibits PDTC-induced procaspase-3 processing and activation.
Bartonella henselae-infection induces cIAP-2 in Mono Mac 6 cells
It was recently described that Neisseria gonorrhoeae and C. pneumoniae protect host cells from apoptosis (Binnicker et al., 2003; Wahl et al., 2003). cIAPs, in particular cIAP-2, play an important role in the inhibition of apoptosis in these particular models; therefore, we were interested in whether infection with B. henselae also modulates cIAP transcription and expression. As shown in Fig. 10A, infection with B. henselae led to increased levels of cIAP-1 and -2 mRNA 4 h after infection. PDTC treatment significantly reduced cIAP-1 and -2 transcription; however, this effect was overcome by infection with B. henselae. In agreement with mRNA expression, cIAP-1 and -2 protein levels were raised in B. henselae-infected cells as shown by immunoblotting. PDTC reduced cIAP-1 and -2 expression, which was increased when cells were infected with B. henselae (Fig. 10B). Using the NF-κB inhibitor MG-132, expression of cIAP-2 was almost completely abolished; however, this effect was again partially overcome when cells were infected with B. henselae(Fig. 11).
During the complex interaction between an infectious agent and the host organism, induction or prevention of apoptosis could be a critical determinant in the outcome of infection. Cell death by apoptosis is a common response of mammalian cells to bacterial infection. A wide variety of pathogens induce apoptosis including Mycobacteria spp., Legionella spp., Shigella spp., Salmonella spp., Pseudomonas spp. and Listeria spp. (Zychlinsky and Sansonetti, 1997). The benefit of induction of apoptosis might be understood as a host cell deletion (e.g. macrophages by Shigella spp. and Salmonella spp.) to enable bacterial survival within the host. However, some pathogens have been found to inhibit apoptosis (Hacker and Fischer, 2002; Lax and Thomas, 2002). The anti-apoptotic properties of the intracellular pathogen C. pneumoniae aimed at avoiding the loss of their cellular habitat have been extensively analysed (Fischer et al., 2001; Rajalingam et al., 2001; Wahl et al., 2001; Airenne et al., 2002). Interestingly, similar anti-apoptotic effects are known for a large group of α2-proteobacteria to which B. henselae also belongs. In particular, Brucella suis (Gross et al. 2000), R. rickettsii (Clifton et al., 1998) and Ehrlichia spp. (Yoshiie et al., 2000) are known to exhibit anti-apoptotic properties. For these obligate and facultative intracellular bacteria, it can be suggested that within a dead host cell bacterial replication is impossible. B. henselae displays a facultative intracellular lifestyle (Andersson and Dehio, 2000) and therefore it is not surprising that the pathogen inhibits apoptosis in endothelial cells (Kirby and Nekorchuk, 2002) which represents one of the assumed bacterial habitats.
Activated monocytes or macrophages typically infiltrate BA lesions (LeBoit et al., 1989; Monteil et al., 1994). As it is known that: (i) a human monocytic cell line (THP-1) shows an increased VEGF response to B. henselae infection (Resto-Ruiz et al., 2002), (ii) Bartonella spp. are able to survive and to replicate within host cells (Brouqui and Raoult, 1996; Kempf et al., 2000) and (iii) bacterial infections with intracellular pathogens result in the inhibition of apoptosis, we analysed whether B. henselae is capable of inhibiting apoptosis in a monocytic cell line. TUNEL and MTS viability assays clearly revealed that B. henselae inhibits PDTC-induced apoptosis in Mono Mac 6 cells. In all experiments, the viability of PDTC-treated cells upon infection was restored to the level of non-infected control cells illustrating an efficient and effective anti-apoptotic capacity of B. henselae.
Viability of B. henselae does not appear to be crucial for this inhibition process (Fig. 3) shown by the use of gentamicin-killed bacteria, indicating that preformed factors might be involved in the process of apoptosis inhibition. This is consistent with our observations that inhibition of apoptosis was also achieved, although to a less extent, by co-cultivating PDTC-treated Mono Mac 6 cells with heat-killed B. henselae (Fig. 3). Moreover, we demonstrated that expression of BadA is involved in the inhibition of apoptosis as the B. henselae BadA– lost the capacity to inhibit apoptosis. This might be linked to the ability of B. henselae to adhere to Mono Mac 6 cells via BadA as B. henselae BadA showed a significantly diminished adherence to host cells. It could be speculated that BadA-mediated host cell adhesion represents the functional basis for further interaction of B. henselae with Mono Mac 6 cells, e.g. via the T4SS. It was recently shown that the T4SS of B. henselae is crucial for the inhibition of apoptosis in endothelial cells (Schmid et al., 2004). Accordingly, the T4SS of B. suis is essential for intracellular survival and intracellular presence is linked to the inhibition of apoptosis (O’Callaghan et al., 1999). Our suggestion is supported by data from the transwell-filter experiments (Fig. 4) which revealed that direct interaction of the pathogens with the host cells is crucial for the inhibition of apoptosis.
In our experiments, inhibition of apoptosis was linked to the colocalization of B. henselae and Mono Mac 6 cells as confirmed by confocal laser scanning microscopy, a two-chamber model and electron microscopy (Figs 4 and 5). This is in contrast to previously published results which described a soluble bacterial factor present in conditioned media responsible for mediating anti-apoptotic activity of B. henselae-infected endothelial cells (Kirby and Nekorchuk, 2002). However, the authors did not consider a potential role of VEGF in their experiments (Kempf et al., 2001; Resto-Ruiz et al., 2002). VEGF is a soluble compound produced by host cells upon infection with B. henselae and is secreted into the culture supernatant. Host cell-derived VEGF is known to possess strong anti-apoptotic activity for endothelial cells (Spyridopoulos et al., 1997), and as this was not assessed these authors can not exclude that their observed anti-apoptotic effect is caused by VEGF. Based on this knowledge, we wanted to elucidate whether inhibition of apoptosis in B. henselae-infected Mono Mac 6 cells is a VEGF-independent process. Mono Mac 6 cells secreted VEGF upon infection with B. henselae similar to monocytic THP-1 cells (Resto-Ruiz et al., 2002) even when treated with PDTC (Fig. 7), and inhibition of apoptosis was not abolished by the use of VEGF-neutralizing antibodies. Therefore, the inhibition of apoptosis in Mono Mac 6 cells by B. henselae is a VEGF-independent process and depends on the cellular presence of B. henselae.
To analyse the subcellular location of B. henselae in more detail, transmission electron microscopy was performed and revealed that the bacteria are located in intracellular membrane-bound vacuoles similar to J774 murine macrophages (Musso et al., 2001). Bacterial morphology appeared unaffected and viable bacteria were recoverable even 18 h after infection, suggesting that killing of B. henselae in these organelles was insufficient. One may assume that these vacuoles are part of the endocytic pathway; however, the exact nature of these vacuoles remains unclear. It is unlikely that B. henselae modulates apoptosis of host cells from this intracelullar compartment because the addition of cyto-D abolished the amount of intracellular bacteria whereas inhibition of apoptosis remained unaffected (Fig. 5).
The molecular mechanisms of host cells that are elicited upon a B. henselae infection are largely unknown. It has been shown that B. henselae induces NF-κB activation in endothelial cells (Fuhrmann et al., 2001). NF-κB plays a key role in apoptosis (Karin and Lin, 2002) and is constitutively expressed in Mono Mac 6 cells (Frankenberger et al., 1994), and macrophages require constitutive NF-κB activation for viability (Pagliari et al., 2000). Therefore, we analysed the potential role of NF-κB in the inhibition of apoptosis by B. henselae using PDTC, a known inducer of apoptosis in Mono Mac 6 cells (Ziegler-Heitbrock et al., 1993) and several other monocytic cell lines (Della Ragione et al., 2000; Hida et al., 2000; Pagliari et al., 2000). We suggest that activation of NF-κB mediates B. henselae-triggered inhibition of apoptosis in Mono Mac 6 cells as demonstrated by the fact that B. henselae still induced NF-κB activation in the presence of PDTC. These results are consistent with previous reports that inhibition of apoptosis by C. pneumoniae and R. rickettsii is NF-κB dependent (Clifton et al., 1998; Wahl et al., 2001). Moreover, we demonstrated that the PDTC-induced activation of the effector caspase-3 is completely abolished when cells were infected with B. henselae shown by Western blotting and quantification of caspase-3 activity (Fig. 9) according to results obtained with endothelial cells (Kirby and Nekorchuk, 2002). The fact that the slightly increased DEVDase activity of B. henselae-infected Mono Mac 6 cells does not correlate with a Western blot-detectable p17 fragment of caspase-3 is explained by the much higher sensitivity of the fluorometric caspase activity assay.
The inhibitors of apoptosis protein (IAPs) have been shown to effectively suppress apoptotic cell death. Human IAPs were first described as being homologous to the baculoviral IAP family which allow the survival and propagation of the virus. Human IAPs include cIAP-1, cIAP-2, x-linked IAP (XIAP), neuronal IAP (NIAP) and survivin. IAPs have been identified as potent inhibitors of active caspase-3, caspase-7 and procaspase-9 (LaCasse et al., 1998; Deveraux and Reed, 1999). Transcriptional regulation of IAPs is mediated by NF-κB (Chu et al., 1997; Stehlik et al., 1998). Our results clearly show that infection with B. henselae induces cIAP-1 and -2 at protein levels even in the presence of PDTC (Fig. 10). Similar results were recently published for N. gonorrhoeae and C. pneumoniae which also inhibit host cell apoptosis (Binnicker et al., 2003; Wahl et al., 2003). These data are supported by the observation that a B. henselae infection also leads to cIAP-2 gene induction and expression when MG-132, a well-described inhibitor of NF-κB, was added (Fig. 11). Considering that: (i) activation of NF-κB accompanies inhibition of apoptosis in B. henselae-infected Mono Mac 6 cells, (ii) B. henselae infection induces NF-κB activation in endothelial cells (Fuhrmann et al., 2001) and (iii) B. henselae inhibits host cell apoptosis in endothelial cells by inactivation of caspase-3 (Kirby and Nekorchuk, 2002), it can be hypothesized that the induction of cIAP-2 via NF-κB might be responsible for the inhibition of host cell apoptosis by B. henselae.
We interpret the results described herein as an extension of the formerly published paracrine VEGF-loop model operating in B. henselae infections (Kempf et al., 2001; Resto-Ruiz et al., 2002). We speculate that the activation of NF-κB and cIAP-2 leads to the inhibition of apoptosis in B. henselae-infected monocytes and to a prolonged secretion of VEGF which is responsible for establishing vasculoproliferations. In line with our hypothesis is the report that an elevated and prolonged VEGF production by tumour cells in mice leads to characteristic features of PH (Wong et al., 2001). Taken together, we have shown that B. henselae inhibits apoptosis in Mono Mac 6 cells. This inhibition of apoptosis is related to the activation of NF-κB and cIAP proteins. Inhibition of apoptosis may lead to a prolonged production of VEGF by B. henselae-infected monocytes presumably involved in triggering the vasculoproliferative disorders BA and PH. Our results underline that the inhibition of apoptosis in monocytes by B. henselae might be crucial in understanding the mechanisms leading to the vasculoproliferative disorders BA and PH.
Bartonella henselae Marseille (Drancourt et al., 1996), B. henselae Houston-1 (Regnery et al., 1992) and B. quintana Toulouse (collection de l’Institut Pasteur, CIP 103739, Paris, France; kindly provided by A. Sander, Freiburg, Germany) were grown on Columbia agar plates supplemented with 5% defibrinated sheep blood (Becton Dickinson) in a humidified atmosphere at 37°C and 5% CO2. For production of bacterial stock suspensions, bacteria were harvested from agar plates after 4 days of culture, resuspended in Luria–Bertani (LB) medium containing 20% glycerol and stored at −80°C. For infection experiments, bacterial stocks were thawed, washed, suspended in cell culture medium and adjusted to the appropriate concentration. The actual inoculum for each experiment was determined by plating serial dilutions of the suspension and calculating the number of colony-forming units (cfu). Heat-killed B. henselae were generated by incubation at 60°C for 40 min and gentamicin-killed B. henselae were generated by incubation with gentamicin (100 µg ml−1) for 3 h; both were subsequently cultivated onto Columbia agar plates for 4 weeks to reveal no viable bacteria. B. henselae BadA– was generated by transposon mutagenesis (Riess et al., 2003; T. M. Riess et al., submitted) and expression of BadA was determined by transmission electron microscopy.
Cell culture and infection procedures
The permanent and differentiated human monocytic cell line Mono Mac 6 (Ziegler-Heitbrock et al., 1988) was used to study the interaction of B. henselae with human monocytic cells. Mono Mac 6 cells were maintained in RPMI 1640 medium (Biochrom) supplemented with 10% heat-inactivated fetal calf serum (FCS; Sigma), non-essential amino acids (Biochrom), l-glutamine (Gibco), 10 µg of streptomycin per ml and 100 units penicillin (Biochrom) and OPI media supplement (Sigma) containing 0.15 mg of oxalacetate, 0.5 g of pyruvate and 8.2 mg of bovine insulin adjusted to 1 l of medium. For infection experiments, cells were washed and maintained for 1 h in the described medium without antibiotics, and the bacteria were then sedimented onto the cultured cells by centrifugation for 5 min at 300 g. PDTC (Sigma) was dissolved in distilled water and added together with the bacteria at a final concentration of 50–500 µM. MG-132 (Sigma) was dissolved in dimethylsulphoxide (DMSO; Roth) and used at a final concentration of 5 µg ml−1. Tumour necrosis factor-α (TNF-α; Sigma) was used at a concentration of 25 ng ml−1 and cyto-D (Sigma) was used at concentrations of 25, 50 and 100 nM.
VEGF ELISA and use of VEGF-neutralizing antibodies
Determination of VEGF induction upon B. henselae infection was performed without antibiotics and FCS to avoid unspecific VEGF secretion. Supernatants were taken when indicated, centrifuged to remove insoluble particles and frozen at −20°C. VEGF concentration in culture medium was quantified using a human VEGF165-ELISA kit (Quantikine; R & D Systems) according to the manufacturer's instructions.
To exclude potential autocrine anti-apoptotic effects, VEGF activity was blocked using VEGF-neutralizing antibodies (10.0 µg ml−1; R & D Systems) as previously described (Kempf et al., 2001). The neutralizing effect was controlled by cultivating Mono Mac 6 cells in the presence of recombinant human VEGF (10 ng ml−1).
Bacterial adhesion and invasion assays
Bacterial adhesion was quantified 30 min after infection by osmotic lysis of host cells and plating the lysate on Columbia agar plates. For this purpose, cells were resuspended, centrifuged for 5 min at 300 g and washed three times in supplemented RPMI medium. Osmotic lysis was performed to determine the total number of bacteria as previously described (Kempf et al., 2000). Briefly, 900 µl of sterile water was added and the cells were passaged 10 times using a 25-gauge needle. Osmotic lysis was overcome by the addition of 100 µl of 10× phosphate-buffered saline (PBS) to the cell lysates and bacterial number was subsequently determined by spread plating 10-fold serial dilutions onto Columbia agar. Intracellular presence (18 h) was determined by a gentamicin protection assay (Kempf et al., 2000) by adding gentamicin (100 µg ml−1) for 3 h to kill extracellular bacteria. After this time cells were extensively washed to remove gentamicin and osmotic lysis was performed as described above.
TUNEL assay, immunostaining and confocal laser scanning microscopy
Nuclear changes associated with early apoptosis were detected by the TUNEL (terminal deoxynucleotidyl transferase-mediated dUTP nick end labelling) method using an in situ cell death detection kit (Roche Diagnostics). Briefly, Mono Mac 6 cells were first washed with PBS, and 1 × 105 cells were centrifuged onto glass slides using a Labofuge 400 (Heraeus). Cells were fixed in 3.75% PBS-buffered paraformaldehyde solution (pH 7.4) and permeabilized with 0.1% Triton X-100 dissolved in 0.1% sodium citrate. Enzymatic incorporation of fluoresceinated nucleotides was performed according to the manufacturer's instructions.
For immunostaining, cells were blocked for 15 min with 0.1% bovine serum albumin (Biomol) dissolved in PBS. Rabbit polyclonal antibodies were raised against viable B. henselae Marseille. Immunostaining was performed by sequentially incubating the cells with a rabbit anti-B. henselae primary antibody (1 h), followed by incubation with a tetramethylrhodamine-isothiocyanate (TRITC)-conjugated goat anti-rabbit IgG secondary antibody (Dianova).
Cellular fluorescence was evaluated using a Leica DM IRE 2 confocal laser scanning microscope (Leica). Two different fluorochromes were detected simultaneously representing the green (fluorescein isothiocyanate, FITC) and red (TRITC) channels. Phase contrast was visualized on a third channel. The corresponding images were digitally processed with photoshop 7.0 (Adobe Systems). The percentage of TUNEL-positive or B. henselae-infected cells was determined by counting ≈ 500 Mono Mac 6 cells over at least 20 random microscopic fields.
MTS viability assay and measurement of cell death
A CellTiter-96 AQueous non-radioactive cell viability assay (Promega) was used to assess cell viability. The assay is composed of the tetrazolium compound MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulphophenyl)-2H-tetrazolium] and an electron coupling reagent phenazine methosulphate (PMS). MTS is reduced by viable cells to formazan which can be measured using a spectrophotometer (absorbance at 490 nm). Previous studies using other cell lines clearly showed that the amount of formazan produced is directly proportional to the number of viable cells (Spyridopoulos et al., 1997). Mono Mac 6 cells (5 × 104 cells per well) were seeded in 96-well plates at a volume of 100 µl. After incubation with PDTC or B. henselae for the indicated time, 20 µl of the MTS–PMS mixture (1:0.05) was added to each well and cells incubated for a further 2 h before measuring the absorbance at 490 nm on a TECAN Sunrise photometer (TECAN). Background absorbance from the control wells (medium without cells, medium without cells but with B. henselae, medium containing PDTC or medium containing PDTC and B. henselae respectively) was subtracted. Viability of control cells was set to 100%. Every calculated value of viability and cell death is the average of three to six wells per group.
Transmission electron microscopy
Bartonella henselae-infected and/or PDTC-treated Mono Mac 6 cells were investigated using transmission electron microscopy as described earlier (Kempf et al., 2000). Briefly, cell pellets were fixed at room temperature for 24 h in a 4% glutaraldehyde phosphate (0.05 M)-buffered solution containing 0.15 M NaCl, pH 7.3. Post-fixation was based on 1% osmium tetroxide containing 1% potassium dichromate in 0.85% NaCl at pH 7.3 for 45 min. After embedding in glycide ether the blocks containing cells were cut using an ultra microtome (Ultracut; Reichert). Ultra-thin sections (80 nm) were stained (Ultrastainer) with 0.5% uranyl acetate for 10 min at 30°C, and 2.7% lead citrate for 5 min at 20°C. Grids were examined using a Zeiss EM 902 transmission electron microscope (Zeiss) operating at 80 kV at magnifications between 2000 and 50 000.
Electrophoretic mobility shift assay (EMSA)
Mono Mac 6 cells (5 × 106) were infected as described above and 60 min after infection, nuclear extracts were prepared as previously described (Schreiber et al., 1989). Aliquots of supernatant containing nuclear proteins were stored at −80°C. Protein concentrations were determined using the Bradford assay (Bio-Rad). Oligonucleotide probes were labelled with [γ-32P]-ATP (Amersham Biosciences) using T4-polynucleotide-kinase (New England Biolabs) and purified on a NucTrap probe purification column (Stratagene). The following oligonucleotides were used: NF-κB consensus (NF-κBc 5′-AGT TGA GGG GAC TTT CCC AGG C-3′; SantaCruz) and NF-κB mutant (NF-κBm 5′-AGT TGA GGC GAC TTT CCC AGG C-3′; SantaCruz). Nuclear extracts (3–6 µg) were incubated with 30 000 cpm of the 32P-labelled oligonucleotide probe for 30 min on ice in a buffer containing 5% glycerol, 80 mM NaCl, 1 mM dithiothreitol (DTT), 1 mM ethylenediaminetetraacetic acid (EDTA) pH 8.0, 10 mM Tris-HCl pH 7.2 and 1 µg poly(dI-dC) for NF-κB shifts. Antibodies against p50 (sc-1190X), p65 (sc-372X) and c-Rel (sc-6955X; all Santa Cruz) were included in the binding reaction for supershift analyses. Samples were resolved on a 5% non-denaturing polyacrylamide gel using 0.5× TBE (25 mM Tris-HCl, 25 mM boric acid, 0.5 mM EDTA) as running buffer. Gels were transferred to Whatman 3M paper (Schleicher and Schüll) and dried under vacuum. Protein binding was assessed via autoradiography.
Mono Mac 6 cells were cultivated in 24-well dishes. Total RNA was isolated using the RNeasy mini kit (Qiagen) according to the manufacturer's instructions. cDNA synthesis was performed with 1 µg of total RNA as previously described (Kampik et al., 2000). cDNA products were amplified by polymerase chain reaction (PCR) in 50 µl of mixture containing 10 mM Tris (pH 8.3), 50 mM KCl, 2.5 mM MgCl2 and 200 mM each of dATP, dCTP, dGTP and dTTP in the presence of 25 pmol each of 5′ and 3′ primers and 2.0 U of Taq polymerase (Roche). The following specific primers and conditions were used: β-actin forward, 5′-TAG AAG CAT TGC GGT GGA CGA TGG AGG G-3′; β-actin reverse, 5′-TGA CGG GGT CAC CCA CAC TGT GCC CAT CTA-3′ (1 min of denaturation at 94°C, 2.5 min of annealing and extension at 72°C for 22 cycles; amplicon size: 600 bp); cIAP-1 forward, 5′-CAC ATG CAG CTC GAA TGA GAA C-3′; cIAP-1 reverse, 5′-TCC AAG GCA GAT TTA ACC ACA GG-3′; cIAP-2 forward, 5′-CAT GCA GCC CGC TTT AAA ACA TTC-3′; cIAP-2 reverse, 5′-CAT TTC CAC GGC AGC ATT AAT CAC-3′ (1 min of denaturation at 94°C, 0.5 min of annealing at 58°C and 0.5 min of extension at 72°C for 26 cycles; amplicon size: 400 bp). PCR products were run on 2% agarose gels and stained with ethidium bromide.
Caspase-3 immunoblotting and fluorimetric assay of caspase activity (DEVDase assay)
Proteolytic processing of procaspase-3 was detected by immunoblotting as described (Lauber et al., 2001). Cell extracts were prepared in a lysis buffer containing 0.5% Nonidet P-40, 20 mM Hepes (pH 7.4), 84 mM KCl, 10 mM MgCl2, 0.2 mM EDTA, 0.2 mM ethylene-glycol-bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid (EGTA), 1 mM DTT, 5 µg ml−1 aprotinin, 1 µg ml−1 leupeptin, 1 µg ml−1 pepstatin and 1 mM phenylmethylsulphonyl fluoride. Lysates (350 µg of protein) were subsequently separated under reducing conditions on a 6–15% SDS-polyacrylamide gel and electroblotted to a polyvinylidene difluoride membrane (Amersham Biosciences). The membrane was blocked for 1 h with 5% non-fat dry milk in 25 mM Tris (pH 7.5), 0.15 M NaCl and 0.05% Tween 20 (Sigma) and incubated for 1 h with murine anti-caspase-3 antibodies (BD Biosciences). The membrane was washed four times and incubated for 1 h with a peroxidase-conjugated affinity-purified anti-mouse secondary antibody (Bio-Rad). After extensive washing, the reaction was developed using enhanced chemiluminescence (ECL) reagents (Amersham).
Cytosolic extracts were prepared in the same lysis buffer used for immunoblotting. Caspase activity was determined by incubation of cell lysates (100 µg of protein determined via Bradford assay, Bio-Rad) with 50 µM fluorogenic substrate N-acetyl-Asp-Glu-Val-Asp-aminomethylcoumarin (Ac-DEVD-AMC; Biomol) in 200 µl of buffer containing 50 mM Hepes (pH 7.3), 100 mM NaCl, 10% sucrose, 0.1% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulphonate (CHAPS) and 10 mM DTT. The release of aminomethylcoumarin (AMC) was measured kinetically by spectrofluorimetry using an excitation wavelength of 360 nm and an emission wavelength of 475 nm. DEVDase activity was determined as the slope of the resulting linear regressions and expressed as x-fold values of the control. Measurements were performed in duplicates and displayed as mean values.
cIAP-1 and -2 immunoblotting
Mono Mac 6 cells (1 × 106) were cultured in six-well plates and infected at an moi of 100 to examine the cellular presence of cIAP-1 and -2. At the indicated time points, cells were washed twice with PBS and resuspended in 60 µl of cell lysis buffer (see above). For Western blot analysis, equal amounts of cell extract proteins were subjected to 12% SDS-polyacrylamide gel electrophoresis and transferred onto polyvinylideneflouride membranes. Membranes were blocked with 5% non-fat dry milk in 25 mM Tris (pH 7.5), 0.15 M NaCl and 0.05% Tween 20 (Sigma) and incubated with cIAP-1 and cIAP-2 antibodies (R and D Systems) according to the manufacturer's instructions. After extensive washing, binding of antibodies was detected using a secondary antibody coupled to horseradish peroxidase (Cell Signaling). An actin-specific antibody (Chemicon) was used as an internal control. Signals were visualized with an ECL kit (Amersham).
All experiments were performed at least three times and revealed comparable results. Differences between mean values of experimental and control groups were analysed by student's t-test. A P-value of < 0.05 was considered significant.
We thank G. Häcker (Munich, Germany) and I. Spyridopoulos (Frankfurt a. M., Germany) for helpful discussions and Andrea Schaefer for excellent technical assistance. This work was supported by grants from the Deutsche Forschungsgemeinschaft to V.K. and to S.W., the German Bundesministerium für Bildung und Forschung (Hep-Net) to S.W., the Fortuene Program of the University of Tuebingen to K.L. and the ‘Landesforschungsschwerpunktprogramm’ of the Ministry of Science, Research and Arts Baden-Württemberg to V.K. and to S.W. (Kapitel 1423 Tit.Gr. 74).
Fig.S1. Inhibition of PDTC-triggered apoptosis in Mono Mac 6 cells by different B. henselae isolates (B. henselae Marseille, B. henseale Houston-1, B. quintana Toulouse; moi 100). Cell viability was determined by MTS assay. Viability of control cells was set to 100%. Each calculated value is the average of three samples per group. *significant difference compared with the PDTC group (P < 0.05).
Fig.S2. Numbers of viable intracellular B. henselae over a period of 24 h. Intracellular bacteria were quantified via gentamicin protection assay. Values are given as a percentage of the original inoculum (moi 100).