OUP user menu

Staphylococcus saprophyticus ATCC 15305 is internalized into human urinary bladder carcinoma cell line 5637

Florian Szabados , Britta Kleine , Agnes Anders , Martin Kaase , Türkân Sakınç , Inge Schmitz , Sören Gatermann
DOI: http://dx.doi.org/10.1111/j.1574-6968.2008.01218.x 163-169 First published online: 1 August 2008

Abstract

Invasion of bacteria into nonphagocytic host cells is an important pathogenicity factor for escaping the host defence system. Gram-positive organisms, for example Staphylococcus aureus and Listeria monocytogenes, are invasive in nonphagocytic cells, and this mechanism is discussed as an important part of the infection process. Uropathogenic Escherichia coli and Staphylococcus saprophyticus can cause acute and recurrent urinary tract infections as well as bloodstream infections. Staphylococcus saprophyticus shows strong adhesion to human urinary bladder carcinoma and Hep2 cells and expresses the ‘Microbial Surface Components Recognizing Adhesive Matrix molecule’ (MSCRAMM)-protein SdrI with collagen-binding activity. MSCRAMMs are responsible for adhesion and collagen binding in S. aureus and are discussed as an important pathogenicity factor for invasion. To investigate internalization in S. aureus, several fluorescence activated cell sorting (FACS) assays have been described recently. We used a previously described FACS assay, with slight modifications, in addition to an antibiotic protection assay and transmission electron microscopy to show that S. saprophyticus ATCC 15305 and the wild-type strain 7108 were internalized into the human urinary bladder carcinoma cell line 5637. The discovery of the internalization of S. saprophyticus may be an important step for understanding the pathogenicity of recurrent infections caused by this organism.

Keywords
  • internalization
  • FACS
  • gentamicin protection assay
  • 5637
  • Staphylococcus saprophyticus

Introduction

Bacterial invasion into nonphagocytic host cells seems to be an effective mechanism for preventing elimination and maintaining infection. In gram-negative bacteria, this mechanism is well documented. In uropathogenic Escherichia coli (UPEC), internalization into the human urinary bladder carcinoma cell line 5637 (Martinez et al., 2000; Hung et al., 2002; Hunstad et al., 2005) and formation of an intracellular biofilm has been described. This internalization allows the organism to persist in the urinary tract (Anderson et al., 2003; Garofalo et al., 2007). Within the cells, the bacteria are protected from antibiotics and the host defence system.

In gram-positive organisms, internalization is observed in Staphylococcus aureus (Jonsson et al., 1991; Patti et al., 1992; McDevitt et al., 1994; Menzies, 2003) and Listeria monocytogenes (Cabanes et al., 2004), and contributes to persistence. It thus seems to be an integral part of the infection process. In S. aureus, adhesion to fibronectin is discussed as an important prerequisite for internalization (Phonimdaeng et al., 1990; Patti et al., 1992; Switalski et al., 1993; McDevitt et al., 1994; Symersky et al., 1997; Palma et al., 2001). Recently, internalization has also been suspected in a clinical isolate of Staphylococcus epidermidis (Khalil et al., 2007). Staphylococcus saprophyticus, a coagulase-negative staphylococcus (CoNS), is detected in up to 40% of urinary tract infections (UTI) in young female outpatients (Wallmark et al., 1978). Recurrent infections are common (Wallmark et al., 1978; Hovelius et al., 1979), bloodstream infections have been documented but are less common (Lee et al., 1987; Golledge, 1988), and endocarditis due to S. saprophyticus is rare (Singh & Raad, 1990). Staphylococcus saprophyticus expresses several surface proteins such as SdrI (Sakinc et al., 2006), UafA (Kuroda et al., 2005), Ssp (Sakinc et al., 2007), and the fibronectin-binding autolysin Aas (Hell et al., 1998). These proteins may contribute to adhesion and possibly internalization of the bacteria into epithelial cells. We therefore sought to analyse internalization of S. saprophyticus into the human urinary bladder carcinoma cell line 5637 (5637 cells). Although S. saprophyticus shows strong adhesion to various epithelial cell lines, such as renal tubular epithelial cells and HEp2 cells (Gatermann et al., 1988; Hell et al., 1998), internalization of S. saprophyticus has only been suspected through the use of conventional microscopy but not proven in an older publication (Schmidt et al., 1989).

Materials and methods

Bacteria and cell lines

Staphylococcus saprophyticus ATCC 15305, S. aureus Cowan I, Staphylococcus carnosus TM300, S. epidermidis ATCC 12228, S. saprophyticus strain 7108, uropathogenic E. coli (UPEC) UTI89 and CFT73 (clinical isolates), human urinary bladder carcinoma cell line 5637 (ATCC HBF-9), and HEp2 cell line (DSMZ, Braunschweig, Germany) were used throughout this study. Bacteria were grown for 12 h at 37 °C in a brain–heart infusion broth (Oxoid, Basingstoke, UK), diluted with fresh broth, and cultured to mid-logarithmic phase. Bacteria were harvested and washed twice with phosphate-buffered saline (PBS) (140 mM NaCl, 8 mM Na2HPO4, 2 mM KH2PO4, pH 7.4).

Fluorescein isothiocyanate (FITC) staining of staphylococci

Bacterial pellets of 1 × 108 CFU were suspended in PBS containing 100 mg L−1 FITC (Applichem, Darmstadt, Germany) and incubated for 1 h at room temperature. Staphylococci were washed thrice with PBS, centrifuged at 4000 g for 2 min, adjusted to a concentration of 1 × 108 CFU mL−1 with RPMI 1640 (PAA Laboratories, Pasching, Austria), and kept at 4 °C until use. Stained bacteria were used in the assays within 12 h. Preliminary experiments had shown that no significant loss of viability or fluorescence intensity occurred within 48 h.

Cell culture preparation and invasion of cell line

5637 cells and HEp2 cells were cultured in RPMI 1640 with phenol red. Modified RPMI 1640, suitable for our cells, was supplemented with 10% heat-inactivated fetal calf serum (PAA Laboratories) and 1 mg mL−1 pyruvate (Invitrogen, Karlsruhe, Germany); 18 h before inoculation with bacteria, 3 × 105 cells were seeded in each of the wells of the plate (Greiner Bio-one, Frickenhausen, Germany) in modified RPMI 1640 with phenol red. Host cell viability was tested using Trypan blue (Invitrogen) staining. The supernatant of eukaryotic cells was removed. Next, 200 μL of FITC-stained bacterial suspension (2 × 107 CFU) was added to 300 μL of the modified RPMI 1640 cell culture medium containing eukaryotic cells. A bacteria to eukaryotic cell ratio of 66 (2 × 107 bacteria and 3 × 105 eukaryotic cells) was used. Plates were incubated for 2 h.

Gentamicin protection assay

After incubation, the culture supernatant was removed and eukaryotic cells were washed twice with Dulbecco's PBS (PAA Laboratories) to remove loosely adherent bacteria. Fresh modified RPMI 1640 without phenol red, containing 100 μg mL−1 gentamicin (PAA Laboratories) and 18 μg mL−1 lysostaphin (WAK-Chemie, Steinbach, Germany) was added and incubated for 1 h at 37 °C. Then, 100 μL of the supernatant was plated onto Mueller Hinton agar plates (Merck, Darmstadt, Germany) in serial dilutions. No bacteria were cultured in these controls. The adherent eukaryotic cells were harvested from these plates using 500 μL 0.25% trypsin–EDTA (Invitrogen) treatment for 5 min complemented by mechanical detachment with a cell scraper (TPP, Trasadingen, Switzerland). Cells were lysed by the addition of 1000 μL double-distilled water for 30 min. Next, 500 μL of the resultant suspension was plated in serial dilutions on Mueller Hinton agar plates, and the colonies were counted after overnight incubation (18 h at 37 °C).

Flow cytometry

5637 cells were coincubated with bacteria and harvested as described in Gentamicin protection assay, but suspensions were diluted with 1000 μL RPMI 1640 without phenol red instead of double-distilled water. The suspension was submitted to a FACSCalibur® flowcytometer (BD Bioscience, Bedford). Data were acquired in logarithmic mode for the forward scatter (FSC), side scatter (SSC), and green fluorescence channel FL1-H (e.g. FITC). Bacteria are smaller than eukaryotic cells; hence FSC and SSC differentiation of bacteria and eukaryotic cells was possible. Intact 5637 cells with distinct FSC and SSC attributes were gated to discriminate intact eukaryotic cells from fragments. Data are presented as the distribution of FL1-H to the FSC attributes of the gated eukaryotic cells. Staphylococcus aureus Cowan I and the UPEC strains UTI89 and CFT73 were used as positive controls for this assay, and S. carnosus TM300 was used as a negative control. FACS experiments were performed on different days, and 2 × 104–4 × 104 events were counted. Staphylococcus saprophyticus ATCC 15305, S. saprophyticus strain 7108, and S. epidermidis ATCC 12228 were compared with the controls. The mean fluorescence of FL1-H (e.g. FITC) given by the FACS machine (arbitrary units) was counted as a surrogate marker for internalization. Trypan blue quenching (0.2%) was used to reduce extracellular signals and increase the specificity of internalized bacteria (Pils et al., 2006).

Microscopy

To investigate the efficacy of the treatment in the FACS internalization assay, 5637 cells were investigated by conventional BX 40 fluorescence microscopy (Olympus, Hamburg, Germany). FITC-stained bacteria were incubated in RPMI 1640 containing gentamicin and lysostaphin for 1 h. Trypsin–EDTA (0.25%) was added for 30 s and bacteria were submitted to microscopy using BD-Falcon® Culture slides (BD Bioscience). For comparison, the experiments were repeated without gentamicin, lysostaphin, and trypsin–EDTA treatment. One thousand eukaryotic cells were observed randomly in four independent experiments and classified into four groups based on the number of adherent and internalized bacteria. These groups contained no bacteria, one to five bacteria, six to 20 bacteria, and more than 20 bacteria in projection with one eukaryotic cell.

Transmission electron microscopy (TEM)

A suspension of 5637 cells from the gentamicin protection assay was used. The lysis step was omitted; cells were centrifuged gently (1000 g) for 60 s, transferred into 500 μL Dulbecco's PBS, and fixed with 500 μL glutaraldehyde 2.5%.

Cells were fixed in Dalton solution (1% osmiumtetroxide), dehydrated (ethanol series: 30–100%), and embedded in epoxy resin. Ultrathin sections were stained with Ultrostain I and Ultrostain II (Leica, Solms, Germany), and examined with an electron microscope EM 900 (Zeiss, Oberkochem, Germany).

Statistical analysis

Experimental data were assessed with an unpaired Student's t-test for comparison between means. P values <0.05 were considered to be statistically significant.

Results

Staphylococcal internalization in a FACS assay

Stained bacteria showed a strong FL1-H and a weak FSC signal, characteristic of small, fluorescent particles. In contrast, 5637 cells were associated with strong FSC and SSC signals but weak FL1-H signals, and could thus be gated and selected. Adhesive and internalized bacteria were measured as a strong FL1-H signal associated with FSC and SCC attributes of 5637 cells (Fig. 1). Extracellular adherent bacteria were lysed by gentamicin and lysostaphin and detached by trypsin treatment. Trypsin treatment has been shown to strongly reduce adherent bacteria (Chhatwal et al., 1987). An FL1-H signal associated with a strong FSC signal thus represented predominantly internalized bacteria. The FL1-H signal of noninvasive S. carnosus TM 300 (Sinha et al., 1999) was used as a negative cut-off value of the internalization assay. The data were presented as mean fluorescence units (MFU) given by the machine, and were gated to FSC and SSC attributes of 5637 cells (Fig. 2). The mean fluorescence of S. saprophyticus ATCC 15305 and wild-type strain 7108 was significantly higher than that of S. carnosus TM300. The mean fluorescence of S. epidermidis ATCC 12228 was not different from that of S. carnosus TM300 (Fig. 3). The mean fluorescence of S. aureus Cowan I, E. coli UTI89, and E. coli CFT73 was significantly higher than that of S. carnosus TM 300 (Table. 1). After Trypan blue quenching, the mean fluorescence in S. saprophyticus ATCC 15305 and S. saprophyticus strain 7108 was still significantly higher than that of S. carnosus TM 300. After Trypan blue quenching, the mean fluorescence of S. aureus Cowan I, E. coli UTI89, and E. coli CFT73 was still significantly higher than that of S. carnosus TM 300 (Table 1). Interestingly, the mean fluorescence of S. aureus was significantly reduced after Trypan blue quenching (Fig. 3). Furthermore, no significant internalization of S. saprophyticus ATCC 15305 and wild-type strain 7108 into the HEp2 cell line was documented in this FACS assay (32 MFU; n=6) compared with S. carnosus TM300 (15 MFU; n=4). Internalization of S. aureus Cowan I into HEp2 cells was still significantly higher (143 MFU; n=4) compared with S. carnosus TM300.

1

Internalization in the FACS assay: The distribution of FL1-H-positive events (e.g. FITC) to the eukaryotic cell population defined by strong FSC-H signal was used as a surrogate marker for internalization (left side). Internalized bacteria of Staphylococcus saprophyticus (a) and Staphylococcus aureus (b) were measured as FL1-H-positive events associated with the FSC-H of eukaryotic cells. The upper limit of the FL1-H distribution of Staphylococcus carnosus TM300 (c) was used to establish a cutoff, horizontal line. Internalization of bacteria into eukaryotic cells causes shift of the distribution to higher FL1-H values (left side). For analytical comparison the means of the FL1-H distribution given by the FACS machine were used (right side). Internalization of S. aureus and S. saprophyticus was significantly higher compared with S. carnosus TM300.

2

Internalization into human urinary bladder carcinoma cell line 5637: The MFU of FL1-H given by the FACS machine was measured quantitatively. Staphylococcus aureus Cowan I (n=18) and clinical isolates of uropathogenic Escherichia coli strains UTI 89 (n=8) and CFT 73 (n=8) were used as positive controls. The internalization was significantly higher in Staphylococcus saprophyticus ATCC 15305 (n=18) and wild-type strain 7108 (n=12) compared with Staphylococcus epidermidis ATCC 12228 (n=8) and non invasive Staphylococcus carnosus TM300 (n=18). Error bars ±SD; *significant difference (P<0.05) concerning S. saprophyticus.

3

Internalization into human urinary bladder carcinoma cell line 5637 in the presence of trypan blue: The MFU of FL1-H was measured quantitatively after trypan blue treatment to reduce extracellular signals. Staphylococcus aureus Cowan I (n=8) and clinical isolates of uropathogenic Escherichia coli strains UTI 89 (n=8) and CFT 73 (n=8) were used as positive controls. Mean fluorescence were the same for Staphylococcus saprophyticus, uropathogenic E. coli, Staphylococcus epidermidis ATCC 12228 and Staphylococcus carnosus TM 300 as without trypan blue. However, mean fluorescence was significantly reduced in S. aureus Cowan I compared to experiments without trypan blue (compare Fig. 2). Error bars±SD; *significant difference (P<0.05) concerning S. saprophyticus.

View this table:
1

MFU of internalized bacteria: Internalization into human bladder carcinoma cell line. The MFU of FL1-H given by the FACS machine was measured quantitatively. Mean fluorescence±SD and number of experiments were presented with and without extracellular quenching.

Without extracellular quenchingTrypan blue quenching
MFUSDnMFUSDn
S. saprophyticus ATCC 1530587.721.71875.629.84
S. saprophyticus strain 710886.134.81885.425.28
Uropathogenic E. coli UTI8972.827.81269.724.28
Uropathogenic E. coli CFT73120.726.91260.625.98
S. aureus Cowan I252.399.318123.838.38
S. carnosus TM 30019.16.21812.34.18
S. epidermidis ATCC 1222812.46.81213.56.58

Staphylococcal internalization in gentamicin protection assay

Staphylococcus saprophyticus was significantly internalized in the classical gentamicin protection assay compared with S. carnosus TM 300 and S. epidermidis ATCC 12228 (Fig. 4). The initial bacterial count (2 × 107) was compared with plated bacteria from lysed eukaryotic cells. Using S. saprophyticus ATCC 15305, 17 000 bacteria were counted (P=0.03; n=5), and the count corresponded to 1.2% of the initial bacterial amount. Using S. saprophyticus strain 7108, 24 000 bacteria (P=0.02; n=3) were counted, and the levels corresponded to 1.4% of the initial bacterial amount. These data were significantly different when compared with S. carnosus TM 300 (1300 counts) and S. epidermidis ATCC 12228 (2000 counts). In S. aureus Cowan I, 133 000 bacteria were counted (n=5), and the count corresponded to 6.6% of the initial bacterial amount.

4

Internalization in gentamicin protection assay: Bacteria (2 x 107 CFU) were incubated with human bladder carcinoma cell line 5637 at a multiplicity of infection of 66. Extracellular bacteria were killed using lysostaphin and gentamicin. Eukaryotic cells containing bacteria were lysed and bacteria were liberated and counted. Bacterial absolute count of Staphylococcus aureus Cowan I (n=5), Staphylococcus saprophyticus ATCC 15305 (n=5), S. saprophyticus strain 7108 (n=3) were significantly higher compared with Staphylococcus carnosus TM300 (n=5) and Staphylococcus epidermidis ATCC 12228 (n=3) absolute counts. Error bars±SD; *significant difference (P<0.05) concerning S. saprophyticus.

Internalization in microscopy

Fluorescence microscopy

The majority (960 of 1000) of counted eukaryotic cells were found in the groups with very few bacteria. Five hundred and one eukaryotic cells were counted without bacteria, 459 eukaryotic cells with one to five bacteria. Only 22 eukaryotic cells with six to 20 bacteria and 18 eukaryotic cells with more than 20 bacteria were counted per eukaryotic cell in four independent experiments. For comparison, nontreated eukaryotic cells (RPMI 1640 without gentamicin, lysostaphin, and trypsin) were counted. In the majority of nontreated eukaryotic cells (981 of 1000), more than six bacteria per eukaryotic cell were counted, including 909 eukaryotic cells with more than 20 bacteria. Only in 19 of the eukaryotic cells one to five bacteria or less were counted.

TEM

Internalization of S. saprophyticus ATCC 15305 into 5637 cells was documented by TEM (Fig. 5). Up to 30 bacteria per cell were found in selected eukaryotic cells.

5

Internalization in microscopy: Transmission electron micrograph of internalized Staphylococcus saprophyticus ATCC 15305. The scale bar indicates 1 μm.

Discussion

Investigation of invasion is important for understanding many clinically important infectious diseases.

The classical internalization experiment is the gentamicin protection assay, in which internalization is detected by growing bacteria from the intracellular compartment. Although this assay is straightforward, the procedure is time consuming and the variability of results is high. Using a FACS device for measurement allows high throughput of cells and a more differentiated investigation of cells and bacteria. Staphylococcal internalization often leads to destruction of the eukaryotic cells (Krut et al., 2003) and invaded bacteria are liberated into the growth medium. Hence, the gentamicin protection assay may underestimate the real percentage of invaded eukaryotic cells. In the FACS assay, intact eukaryotic cells can be gated and distinguished from destroyed cells and liberated bacteria. Our FACS internalization assay has only slight modifications to previously described FACS internalization assays (Sinha et al., 1999; Krut et al., 2003). Internalization of UPEC strains UTI89, CFT73, and S. aureus Cowan I was confirmed according to the literature (Sinha et al., 1999; Martinez et al., 2000). Staphylococcus epidermidis ATCC 12228 was not internalized into 5637 cells and HEp2 cells. Recently, internalization of a clinical isolate of S. epidermidis into the human bone cell line MG63 was described (Khalil et al., 2007), indicating that internalization may be an important pathogenicity factor in CoNS as well. Because only a few adherent S. saprophyticus were found in our assay, as shown by fluorescence microscopy and extracellular Trypan blue quenching (Pils et al., 2006), we conclude that adherence does not influence the internalization measurement in S. saprophyticus in our assay. After Trypan blue quenching, internalization into the human bladder carcinoma cell line was significantly reduced in S. aureus Cowan I. This provides evidence that adherence plays an important role in S. aureus Cowan I, and internalization measured by FACS assays without extracellular quenching may overestimate internalization. We therefore conclude that extracellular quenching should not be omitted in FACS internalization assays. The lack of internalization of S. saprophyticus into HEp2 cells is further evidence that our FACS assay clearly differentiates between adherence and internalization, because a strong adherence of S. saprophyticus to HEp2 cells has been described earlier (Gatermann et al., 1988; Hell et al., 1998). The internalization of S. aureus was fourfold higher in quantitative FACS measurement and sixfold higher in the gentamicin protection assay as compared with S. saprophyticus. Because staphylococci grow in clusters, classical plate counts will rather represent the number of bacterial clusters whereas the FITC signal in the FACS assay is proportional to the number of bacteria. FACS results and the classical assay therefore yield similar results that need not correlate numerically. Using electron microscopy, S. saprophyticus ATCC 15305 was documented to be inside 5637 cells (Fig. 5). These pictures bear striking similarities to internalized UPEC strains (Martinez et al., 2000). Recently, the ability of S. aureus to infect and survive in professional phagocytes and nonphagocytic cells has been described (Jonsson et al., 1991; Patti et al., 1992; McDevitt et al., 1994; Menzies, 2003). The survival of bacteria in nonphagocytic host cells, such as urothelial cells, may have important implications in UTI (Martinez et al., 2000; Mulvey et al., 2001; Anderson et al., 2003; Hunstad et al., 2005; Garofalo et al., 2007). UPEC and S. saprophyticus both cause UTI and bloodstream infections. Certain UPEC strains, for example UTI89 and CFT73, are known to be internalized into 5637 cells (Martinez et al., 2000; Mulvey et al., 2001). Also, it was shown that UPECs are internalized into urothelial cells in a mouse model, leading to persistence of the bacteria in the urinary tract (Garofalo et al., 2007). Using this FACS assay, live S. saprophyticus was internalized into the 5637 cells. Formaldehyde treatment significantly reduced the internalization of S. saprophyticus into 5637 cells, indicating that formaldehyde modifies a factor that is essential for the internalization process in S. saprophyticus. In S. aureus, internalization of the bacteria was not significantly modified by formaldehyde pretreatment (data not shown). These data were in line with previous findings and the reason as to why usually formaldehyde-inactivated S. aureus was used in FACS internalization assays (Sinha et al., 2000; Krut et al., 2003). These data suggest that the mechanism of internalization in S. saprophyticus and S. aureus may be different, and thus further experiments need to be conducted.

In conclusion, this is the first description of S. saprophyticus internalization into the human bladder carcinoma cell line 5637. This observation might lead to further insights into the pathogenicity of recurrent UTIs caused by this species.

Acknowledgements

We thank Anke Albrecht (Bochum, Germany) for technical assistance. Also, we thank K. Hantke (Tübingen, Germany) and U. Dobrindt (Würzburg, Germany) for providing the clinical isolates of UPEC strains.

Footnotes

  • Editor: Jan-Ingmar Flock

References

View Abstract