A shear stress micromodel of urinary tract infection by the Escherichia coli producing Dr adhesin

Abstract

In this study, we established a dynamic micromodel of urinary tract infection to analyze the impact of UT-segment-specific urinary outflow on the persistence of E. coli colonization. We found that the adherence of Dr+ E. coli to bladder T24 transitional cells and type IV collagen is maximal at lowest shear stress and is reduced by any increase in flow velocity. The analyzed adherence was effective in the whole spectrum of physiological shear stress and was almost irreversible over the entire range of generated shear force. Once Dr+ E. coli bound to host cells or collagen, they did not detach even in the presence of elevated shear stress or of chloramphenicol, a competitive inhibitor of binding. Investigating the role of epithelial surface architecture, we showed that the presence of budding cells-a model microarchitectural obstacle-promotes colonization of the urinary tract by E. coli. We report a previously undescribed phenomenon of epithelial cell "rolling-shedding" colonization, in which the detached epithelial cells reattach to the underlying cell line through a layer of adherent Dr+ E. coli. This rolling-shedding colonization progressed continuously due to "refilling" induced by the flow-perturbing obstacle. The shear stress of fluid containing free-floating bacteria fueled the rolling, while providing an uninterrupted supply of new bacteria to be trapped by the rolling cell. The progressive rolling allows for transfer of briefly attached bacteria onto the underlying monolayer in a repeating cascading event.

Fig 1. Dr+ E. coli adhere to bladder T24 cells through interaction with DAF.

(A) Surface expression of DAF and CEACAM receptors in T24 bladder carcinoma cells. FACS histograms of a representative labeling of T24 cells with antibodies against human DAF (red curve) or CEACAM (blue curve) conjugated with FITC are shown in comparison to labeling with the isotype control antibody (green line). T24 cells constitutively express DAF, but not CEA receptor. Data is representative of three repeats. (B,C) Immunofluorescence analysis of T24 cells using polyclonal anti-DAF antibodies. The arrowhead marks a single cell that overexpresses the DAF receptor. (D-F) T24 cells incubated with Dr+ E. coli in static conditions. Unbound bacteria were washed off, and bound ones were visualized via endogenous GFP fluorescence (D) or immunolabeling with anti-Dr antibodies (E). Panel (F) was generated using phase-contrast microscopy. (G-H) T24 cells incubated with (H) or without (G,I) polyclonal anti-DAF antibodies and washed with Dr- (G) or Dr+ (H,I) E. coli. Unbound bacteria were washed off, and remaining ones were visualized using GFP fluorescence. Pictures (B-I) are representative of four independent replicates. Scale bars correspond to 50 μm.

Fig 2. Adherence of Dr+ E. coli to T24 cells is inhibited by increasing shear stress.

(A) Accumulation of Dr+ E. coli over 20 minutes of flow at different values of shear stress. One representative dataset out of 12 replicates is shown. (B) Attachment rate (open circles) calculated as the total count of binding events divided by total time, and accumulation rate (full circles) calculated from the accumulation curves in panel A. The dashed line represents accumulation rate for Dr- (non-fimbriated) E. coli. Inset: adherence efficiency of Dr+ bacteria at individual shear stresses, defined as the ratio of accumulation to attachment rates. (C) Classification of attachment events according to the ultimate fate of the bound bacterial cell: red–total attachment events (recorded at one-frame resolution over 12 minutes); blue–detachment events within 1 second from the initial binding; green points–detachment events more than one second from the initial binding; black–persistently bound bacteria. The dashed line represents the number of accumulated Dr- E. coli after 12 minutes. Inset: adherence efficiency of Dr+ bacteria, here defined as the ratio of bacterial accumulation after 12 minutes to the attachment events count during the 12 minutes. The error bars in (B) and (C) give 95% confidence intervals of 12 replicates. (D) Black line–results of the detachment experiment. ca. 1350 initially bound bacteria were tracked for eventual detachment by the increasing flow of medium, generating shear stress ranging from 0.06 to 18.46 pN μm^-2 over 28 minutes. The flow was increased stepwise without pausing between subsequent flow values, as denoted in the figure. Red line–representative result of the detachment experiment with 1150 bacteria initially bound. Each flow step was separated from the next by a 2-minute pause (not included in the time scale), totaling 42 minutes. One representative experiment out of 12 replicates is shown. The error bars give 95% confidence intervals of 12 replicates, and correspond to a fraction of initially bound bacteria that remain bound at the end of flow (at a shear stress of 4.62 pN μm^-2). S2 Fig shows statistics of bacteria that remain bound at the end of each flow step relative to the beginning of the same step, at flows ranging from 0.06 to 4.62 pN μm^-2.

Fig 3. Chloramphenicol blocks initial bacterial adherence and does not detach pre-bound bacteria.

(A) Accumulation curves of Dr+ E. coli pre-incubated with Cm at the indicated concentrations at shear stress of 0.01 pN μm^-2. Real-time accumulation was monitored using GFP fluorescence. One representative experiment out of 12 replicates is shown. (B) Attachment (open circles) and accumulation (full circles) rates of bacteria pre-incubated with Cm, calculated identically as in Fig 2. Inset: adherence efficiency as a function of Cm concentration. (C) Classification of attachment events according to the fate of the bacterial cell pre-incubated with Cm: red–total attachment events (recorded at one-frame resolution over 12 minutes); blue–detachment events within 1 second from the initial binding; green points–detachment events more than one second from the initial binding; black–persistently bound bacteria. Inset: adherence efficiency as a function of Cm concentration The error bars in (B) and (C) give 95% confidence intervals of 12 replicates (D) Results of the detachment experiments performed at three different Cm concentrations. Ca. 1100–1600 initially accumulated bacteria were tracked for eventual detachment during flow of medium supplemented with Cm at a given concentration. The flow was increased stepwise, generating shear stress ranging from 0.06 to 9.23 pN μm^-2 as denoted in the figure. For each Cm concentration, one representative experiment out of 12 replicates is shown. The error bars give 95% confidence intervals of 12 replicates, and correspond to a fraction of initially bound bacteria that remain bound at the end of the flow step (at a shear stress of 4.62 pN μm^-2). S3 Fig shows statistics of bacteria that remain bound at the end of each flow step relative to the beginning of the same step, at flows ranging from 0.06 to 4.62 pN μm^-2.

Fig 4. Shear stress inhibits adherence of Dr+ bacteria to type IV collagen by reducing success rate of initial binding.

(A) Accumulation of Dr+ E. coli washed through the flow chamber coated with 20, 2 or 0.2 μg ml^-1 human type IV collagen, at shear stress ranging from 0.01 to 1.15 pN μm^-2. Real-time accumulation was monitored using GFP fluorescence. One representative experiment out of 12 replicates is shown. (B) Classification of attachment events according to the ultimate fate of the bacterial cell bound to type IV collagen: red–total attachment events (recorded at one-frame resolution over 12 minutes); blue–detachment events within 1 second from the initial binding; green points–detachment events more than one second from the initial binding; black–persistently bound bacteria. The dashed line represents the number of Dr- E. coli accumulated after 12 minutes of flow. Inset: adherence efficiency, defined as the ratio of bacterial accumulation after 12 minutes to the attachment events count during the 12 minutes. The error bars give 95% confidence intervals of 12 replicates. (C) Results of the detachment experiment using type IV collagen. Ca. 100–850 initially accumulated bacteria were individually tracked in stepwise increasing shear stress, ranging from 0.58 to 9.23 pN μm^-2 over 16 minutes, as shown in the figure. For each collagen concentration, one representative experiment out of 12 replicates is shown. Collagen concentrations in panels (B) and (C) are as indicated in panel (A). The error bars give 95% confidence intervals of 12 replicates, and correspond to a fraction of initially bound bacteria that remain bound at the end of flow (at a shear stress of 9.23 pN μm^-2. S4 Fig shows statistics of bacteria that remain bound at the end of each flow step relative to the beginning of the same step. (D) Accumulation of Dr+ E. coli pre-incubated with Cm or Kn washed through the flow chamber coated with 20 μg ml^-1 human type IV collagen, at shear stress of 0.01 pN μm^-2. Real-time accumulation was monitored using GFP fluorescence. One representative experiment out of 12 replicates is shown. (E) The number of bound bacteria in the field of view in the accumulation experiments described in panel (D). The error bars give 95% confidence intervals of 12 replicates.

Fig 5. Shear stress induces moving of Dr+ bacteria on the surface of the T24 cell layer.

Representative 1-dimensional trajectories of individual bacteria bound to the confluent layer of T24 cells at shear stress of 0.14, 0.28 and 0.58 pN μm^-2. An asterisk (*) indicates detachment event. Trajectories were obtained from time-lapse videos of Dr+ E. coli bound to T24 cells, as described in Fig 2A.

Fig 6. Budding cells act as obstacles that induce intensive adherence of Dr+ bacteria at elevated shear stress.

(A) Formation of an increased adherence zone behind a host cell budding from cell layer (white arrow). The representative picture was taken after 8 minutes of flow, at shear stress of 0.42 pN μm^-2. Video S10 represents time-lapse videos of zone formation presented in this panel. (B) Representative picture of Dr- E. coli washed through the flow chamber with an overgrown layer of T24 cells with budding cells, at shear stress of 0.42 pN μm^-2. Only several bacteria were accumulated in the field of view after 20 minutes of flow. (C) Representative picture of Dr+ E. coli washed through the flow chamber with a confluent layer of T24 cells (no budding cells) at shear stress of 0.42 pN μm^-2. About 320 bacteria were accumulated in the field of view after 20 minutes of flow. (D) Representative picture of Dr+ E. coli washed through the flow chamber with an overgrown layer of T24 cells with ca. 30 budding cells, at a shear stress of 0.42 pN μm^-2. About 1800 bacteria were accumulated in the field of view after 20 minutes of flow. In panels (A-D) red arrows indicate the direction of flow, bars correspond to 50 μm. (E) Effective cross-sections (blue) calculated for the half-spherical obstacles O₁, O₂ and O₃ (black contour). Cross-section area S, volume coefficient V and the maximal deposition rate R_max are shown for each simulated system. (F) The number of bound bacteria in the field of view in the accumulation experiments described in panels (B), (C) and (D). In (B) and (D), there were 24 ± 8 budding cells. The error bars give 95% confidence intervals of 12 replicates. (G) Schematic depiction of the postulated mechanism of enhanced bacterial capture by adhesive obstacles. Among flow lines that start beyond the 1 μm-thick capturing layer (shown in light red), a fraction will enter the capturing layer near the top of the obstacle (red lines) and then roll over the cell surface, while others will avoid the obstacle (green lines). The region from which the productive (red) trajectories emerge is defined as the effective cross-section (light blue). Other factors such as vorticity might additionally enhance accumulation behind the obstacle.

Fig 7. Rolling host cells induce intensive adherence of Dr producing bacteria at elevated shear stress.

(A-C) Formation of an increased adherence zone due to rolling of a detached host cell at shear stress of 1.15 pN μm^-2. White arrows mark the positions of the rolling cell at a given time. Video S11 represents time-lapse videos of zone formation presented in panels (A-C). (D) Representative picture of Dr- E. coli washed through the flow chamber with an overgrown layer of T24 cells with budding and rolling cells at shear stress of 1.15 pN μm^-2. Only several bacteria were accumulated in the field of view after 20 minutes of flow. (E) Representative picture of Dr+ E. coli washed through the flow chamber with a confluent layer of T24 cells (i.e., without budding and rolling cells) at shear stress of 1.15 pN μm^-2. About 65 bacteria were accumulated in the field of view after 20 minutes of flow. (F) Representative picture of Dr+ E. coli washed through the flow chamber with overgrown layer of T24 cells and about 20 rolling host cells at shear stress of 1.15 pN μm^-2. About 2200 bacteria were accumulated in field of view after 20 minutes of flow. In all pictures red arrows indicate the direction of flow, bars correspond to 50 μm. (G) The number of bound bacteria in the field of view in the accumulation experiments described in panels (D), (E) and (F). In panels (D) and (F), 23 ± 8 rolling host cells moved through the field of view. The error bars give 95% confidence intervals of 12 replicates. (H) Schematic depiction of possible outcomes of cell budding/exfoliation in the flow of bacteria-rich medium. The initially detached host cell remains attached to the cell line with a layer of adherent bacteria, and rolls along the surface pushed by the flow. It captures the incoming bacteria, acting as an obstacle, and deposits them on the cell layer surface as it rolls. Eventually, high flow or low bacterial concentration might lead to full detachment, or the cell can come to a full stop in low flow conditions, accumulating more bacteria as a typical obstacle.

Fig 8. Model of adherence of Dr producing E. coli cells to T24 bladder cell lines at shear stress generated by fluid flow.

Formation of DraE-DAF contacts upon initial contact between the bacterial and human cells. Transiently bound bacteria are anchored by few such interactions and are hence prone to dissociation in high shear stress conditions. Within seconds, the formation of further adhesin-receptor complexes reinforces the adherence, leaving the bacteria almost permanently attached to the cell line.