Infections were performed with Streptococcus pneumoniae serotype 35A (ATCC11733, Jagau et al., 2019a), since this serotype is less capsulated and is described as a suitable serotype producing the whole subset of virulence traits (Bergmann et al., 2009; Luttge et al., 2012). This strain henceforth is abbreviated as WT. In addition, we use the isogenic pneumolysin-deficient isogenic strain (St35AΔply, Pracht et al., 2005). Pneumococci were grown to mid log phase in Todd Hewitt liquid broth supplemented with 5% yeast extract at 37°C and 5% CO₂. The Δply strain was cultivated in the same medium supplemented with 5.0 μg/ml erythromycin. To infect endothelial cells, bacteria were centrifuged for 4 min at 3,000 × g, washed twice with 10 ml phosphate-buffered saline (PBS), and an amount of 1 × 10⁸ pneumococci per mL was photometrically adjusted by determining the optical density at 600 nm. For heat inactivation of pneumococci, 1 × 10⁹ bacteria were resuspended in 1 ml PBS and incubated at 60°C for 1 h. Heat inactivation was confirmed by plating on blood agar.

Human umbilical vein endothelial cells (HUVEC) were purchased as pooled donor cryostocks from PromoCell (Heidelberg, Germany) and were subcultivated using Accutase solution according to the recommendations of the manufacturer. The primary endothelial cells were cultivated in Endothelial Cell Growth Medium (ECGM, PromoCell) supplemented with the recommended suspension of growth factors and hormones in 75-cm² cell culture flasks at 37°C and 5% CO₂. Subcultured cells were used only up to the 8th passage.

Pneumococcus-specific polyclonal antibodies were generated in rabbit by Pineda. Alexa Fluor 488-conjugated goat anti-rabbit, Alexa Fluor 568-conjugated goat anti-rabbit, and 4′,6-diamidino-2-phenylindole (DAPI) were from Thermo Fisher Scientific (Waltham, MA, United States), paraformaldehyde (PFA) was purchased from Polysciences (Hirschberg, Germany), and mounting medium was from Dako (Jena, Germany). Alexa-488-conjugated phalloidin was purchased from Abcam (Cambridge, MA, United States) and porcine gelatin from Merck (Burlington, MA, United States). Purification of Strep-tagged pneumolysin was described recently (Jahn et al., 2020). Pneumolysin-activity analyses were performed using citrated sheep blood from Fiebig Nährstofftechnik, Germany. In addition, the purity and integrity of the purified pneumolysin protein were also controlled by denaturating SDS-polyacrylamide gel electrophoresis.

The hemolytic activity of pneumococcal lysates was determined essentially as described by Benton et al. (1997). In brief, heterologously produced and purified pneumolysin protein was diluted serially in PBS from 2,500 ng/ml down to 19.5 ng/ml in a 96-well microtiter round-bottom plate. The erythrocyte suspension of citrated sheep blood was washed twice with PBS, and 100 μl of a 1:50 dilution was added to 100 μl of lysis buffer, containing 10 mM DTT and 0.1% bovine serum albumin. The plates were incubated at 37°C for 30 min and then centrifuged at 1,500 × g for 10 min at RT. Erythrocyte-free PBS and lysis buffer were used as negative controls. Experiments were performed in three independent assays, each in duplicates, and the presence or absence of erythrocyte sediment was photographically monitored.

The chamber separation cell migration analysis (CSMA) was established for determination of the speed and closure efficiency of a defined gap of 500 μm in diameter that is generated between two confluently grown endothelial cell layers. For this purpose, a three-chambered silicone inlet from ibidi^® was placed on a 35-mm petri dish suitable for microscopic visualization. Each chamber covers a growth area of 0.22 cm² with a median width of 500 μm. The bottom of three adjacent inlet chambers was coated with 2% porcine gelatin for 1 h at 37°C. Excess of gelatin was removed by two wash steps with 0.1 M PBS (pH 7.4). Further, 70 μl of 3 × 10⁵/ml HUVEC suspension was seeded into each of the three adjacent inlet chambers and cultivated for 18 h at 37°C and 5% CO₂ to reach confluence. Cell migration analyses were started by removing the three-chambered silicone inlet. The dish was washed twice with ECGM to remove unattached HUVEC, and cellular gap closure was analyzed microscopically using a confocal laser scanning microscope (CLSM) at different time points of cultivation. The microscope setting was adjusted using the bright-field mode of the CLSM (Leica, SP8, DMI8) with the × 20/0.75 IMM objective and a zoom factor of 0.75. For evaluation of wound healing parameters, both cell borders were microscopically visualized and snap shots were taken at three different representative sights of view. The imaged region size constantly covered 775 μm × 775 μm. For calculation of the cell-free gap area per field of view, the cell borders of the cell-free gaps were marked using the LAS X software tool. At each observation time point, the size of the remaining cell-free area was determined via the area determination tool of the LAS X microscope software. The values were put in relation to the size of the total cell-free area determined at time t = 0 h, which was normalized to 0% gap closure. The percentage of gap closure (i) was calculated according to the following formula:

Additionally, the average velocity of the reduction of the cell-free area within 24 h of monitoring was calculated based on the difference between the values of the remaining cell-free area between each monitored period of time of gap closure. This value is divided by 24 h, which resulted in the average rate of gap closure per hour in μm²/h. Assuming that the cell-free area reduces linearly over time, the following formula according to Bobadilla and colleague’s “monolayer edge velocimetry” method was applied for calculation of the average gap closure rate [U_w] (ii, Bobadilla et al., 2019).

(ii)

To determine the impact of pneumococci on the velocity and efficiency of endothelial wound healing, wild-type serotype 35A pneumococci and a pneumolysin-deficient (Δply) derivative were added to the cell layers using a multiplicity of infection (MOI) of 10, after the three-chambered inlet was removed. The culture medium was exchanged after 1 h and additionally after 6 h of cell cultivation. The microscopic visualization and evaluation of gap closure were performed as described above. Assays were performed in replicates in three independent analyses and statistically evaluated, as described below. Of note, with the aim to prevent unspecific effects due to increasing amounts of bacteria during 24 h of CSMA, uncontrolled bacterial growth throughout the CSMA was prevented by medium exchange at specific time points during CSMA. The same experimental setup was performed with heat-inactivated pneumococci. In addition, recombinantly produced pneumolysin was employed in the CSMA. Based on the results of toxic activity against erythrocytes, CSMA experiments were performed as described above in the presence of 50, 250, 375, and 500 ng/ml pneumolysin protein.

For immunofluorescence staining, HUVEC were seeded in three-chambered silicone inlets, which were placed onto gelatin-coated glass coverslips into a 24-well plate. The CSMA was performed with and without bacteria, as described above. After each time point, the cell culture medium was removed, and after two wash steps with PBS, cell growth was stopped by fixation with 3% paraformaldehyde (PFA) in PBS at 4°C o/n. Antibody-mediated fluorescent staining was performed in a dark chamber and was started by blocking for 1 h at RT with 5% bovine serum albumin (BSA). All staining steps were performed in a volume of 50 μl and were separated by intensive washing in PBS reservoirs. Probes without bacteria were incubated with Alexa 488-conjugated phalloidin (1:1,000 dilution) and with DAPI (in 1:1,000 dilution). For differential staining of attached and internalized pneumococci, the bacteria-incubated probes were additionally incubated with primary pneumococcus-specific antibodies (1:100 dilution) followed by an Alexa 488-conjugated secondary anti-rabbit antibody (1:1,000 dilution). After cell permeabilization using 0.1 Triton X-100 for 5 min, the internalized pneumococci were detected with the polyclonal pneumococcus-specific antibody (1:100 dilution) followed by incubation with an Alexa 568-conjugated secondary antibody (1:1,000 dilution). Internalized bacteria inside the endothelial cells appear in red and bacteria attached to the cell surface appear yellow due to the overlay of both secondary conjugated antibodies. Prior to microscopic visualization, the stained probes were mounted onto glass slides using a mounting medium from Dako. A microscopical visualization of the fluorescence signals of the CSMA samples was performed using the × 20/0.75 IMM objective of the CLSM (Leica, Sp8, DMI8) and a zoom factor of 0.75 to maintain a constant size of the field of view. For visualization of attached and internalized pneumococci, the CLSM mode with the × 63/0.75 oil-immersion objective was used and results are shown as representative images of the merged fluorescence channels. The scale bars are defined in each figure legend and indicated sizes of 100 μm and of 10 μm in magnified pictures. Contrast and brightness were only slightly optimized using Adobe Photoshop CS5 (version 12.0.) without changing any results. A time-lapse movie was generated using the software “Microsoft Photos” (Microsoft corporation, version: 2021.21090.10007.0). The movie is based on eight images at a speed of one image per second and is included in the Supplementary Material.

The proportion of cell proliferation and cell migration during gap closure was microscopically visualized and quantified using the EdU Click assay from Carl Roth, Germany, according to the instructions of the manufacturer. In brief, CSMA was performed as described above in the presence of 10 μM 5-ethynyl-2′-desoxyuridin (EdU), which is incorporated into newly synthesized DNA of proliferation endothelial cells during cell growth. The culture medium was exchanged after 1 h and after 6 h of cultivation followed by supplementation of the required EdU substances. After each cultivation time point, migrated cell samples were stopped by fixation with 3% PFA, as described above. For detection of cell proliferation, the cells were permeabilized using 0.1% Triton X-100 and cells were incubated with 5-TAMRA-PEG3-Azide solution. Microscopic imaging using the CLSM detected red fluorescent cell nuclei at an emission peak at 579 nm. For counterstaining of migrating endothelial cells, DAPI stain was performed as described above. The proportion of proliferating cells was microscopically quantified by counting the red fluorescent and blue fluorescent cell nuclei using the × 20/0.75 IMM objective of the CLSM (Leica, Sp8, DMI8) and a zoom factor of 0.75. For comparative quantification, the number of nuclei was determined in randomly chosen sights of view within the whole gap area of 500 nm in diameter at different time points of gap closure analyses. All experiments were performed in at least four independent assays, each at least in duplicates, and the data were expressed as mean value including standard deviation.

For analysis of cell migration under defined shear stress conditions, the CSMA in combination with flow cultivation was established using the microfluidic system of ibidi^®. In brief, primary HUVEC were seeded into gelatin-coated two-chambered silicone inlets (ibidi^®), which were placed onto ibiTreat coverslips (ibidi^®), as described above. The cells were cultivated for at least 18 h at 37°C and 5% CO₂ to reach confluence. The silicone inlet was removed after a gentle wash step with ECGM. Immediately after that, a sticky 0.4-mm microscope slide with Luer adaptors was placed on the ibiTreat coverslip and 150 μl of ECGM was filled into the Luer adaptor reservoirs. The coverslip was pressed onto the sticky slide with a special clamp (ibidi^®) for 1 min at 37°C. The slide was connected to a degassed perfusion set with tubings of 50 cm in length and 1.6 mm in diameter. Cell adaptation to the flow was started at 4 dyn/cm² for 30 min followed by an increase to 1 dyn/cm² every half an hour until the final shear stress of 10 dyn/cm² was reached. The cells were cultivated in continuous flow of 10 dyn/cm² for up to 48 h using a total circulating volume of 13.6 ml ECGM. Gap closure was microscopically visualized and was calculated as described above at time points 0, 6, 24, and after 48 h. The same experimental setting was applied to illustrate cell movement and gap closure by fluorescence visualization of the actin cytoskeleton and the nuclei, as described above. A schematic workflow of the combined technique using CSMA and microfluidic cultivation is shown in Figure 5A. The microfluidic slide containing the HUVEC layer with a defined cell-free area was also incubated with Serotype 35A pneumococci (WT), grown to mid log phase, which were supplemented to the circulating cell culture medium at an MOI of 10. In accordance with the infection routine of the static assays, the cell culture medium was exchanged after 1, 6, 24, and 30 h of bacterial circulation in the microfluidic system. The effect of bacteria on the speed of gap closure was determined microscopically after fixation of the microfluidic infection and immunofluorescence staining, as described above.

Samples were fixed in 5% formaldehyde and 2% glutaraldehyde in 0.1 M HEPES buffer (HEPES 0.1 M, 0.09 M sucrose, 10 mM CaCl₂, 10 mM MgCl₂, pH 6.9) at 4°C and washed twice with TE buffer, pH 7.0 (20 mM Tris–HCl, 1.0 mM EDTA), before dehydrating in a graded series of ethanol for 15 min at each step. Samples were then subjected to critical-point drying with liquid CO₂ (CPD 300, Leica Microsystems, Wetzlar, Germany) and sputter coated with a gold–palladium film (SCD 500, Bal-Tec, Liechtenstein) before examination in a field-emission scanning-electron microscope Zeiss Merlin (Oberkochen, Germany) using the Everhart–Thornley SE detector and the SE in-lens detector at a 75:25 ratio with an acceleration voltage of 5 kV.

Statistical significances were analyzed by the two-tailed unpaired T-test with unequal variance (Welsh test, Rasch et al., 2011) or in case of equal variances with Student’s T-test, as noted in figure legends. Data are based on technical replicates that derived from at least three independent experimental analyses. In case of microscopic evaluation, data from at least three different sights of view were combined for mean value calculation. The graphs display mean values and calculated standard deviations. P-values of < 0.05 were considered to be statistically significant and were marked by asterisks or declared in figure legends.

To model the effects of pneumococcal infection on endothelial wound healing, we established a two-chambered separation cell migration assay (CSMA), which enabled real-time microscopic visualization and reproducible quantification of the pneumococcal effect on endothelial tissue gap closure. HUVEC were seeded into a chambered cell culture inlet, which after removal generates a defined gap between the adjacent confluently grown monolayers. HUVEC typically serve as a well-characterized and widely used model cell type for functional analyses of the vascular endothelium. The cell growth within the defined gap area of the CSMA samples was microscopically monitored for up to 24 h.

In Figure 1A, representative images depict the fields of view, which are used for quantification of the cell-free area for CSMA without bacteria (untreated) and in the presence of wild-type pneumococci (+WT) at indicated time periods of cell cultivation. The cell borders were marked with black dashed lines using the LAS X software (Figure 1A). The bright-field images clearly visualize the reduction in the size of the cell-free area due to cell growth in untreated samples, until the gap is totally closed after 24 h (Figure 1A, untreated). The generated time-lapse movie displays closure of the cell-free gap within 24 h (Supplementary Movie). In the presence of pneumococci, overall cell growth was significantly reduced at each visualized time point and a substantial cell-free gap area was still detected after 24 h (Figure 1A, +WT). Already after 6 h of cell cultivation and longer, significant differences in gap closure between the untreated samples and the CSMA samples with bacteria were observed and provide evidence that the presence of pneumococci significantly delayed the gap closure of the HUVEC layer (Figure 1B). Quantification of the cell-free area revealed that 90% ± 5% of the gap area was closed already after 18 h of cultivation and a total gap closure (100%) was achieved after 24 h at the latest in untreated samples without bacteria (Figures 1A,B, untreated). Fluorescence staining of the actin cytoskeleton and the cell nuclei confirmed the stepwise reduction of the remaining cell-free gap area until the cell-free area was completely closed after 24 h of incubation without bacteria (Figure 1C, untreated). The cells remained attached to the gelatinated coverslip surface during the whole cultivation period, and the actin cytoskeleton stain further revealed a cobblestone-like typical cell morphology within the growing cell layer. Moreover, already 1 h after starting the CSMA cell cultivation, cells with protruded cell evaginations and extended filopodia were detected at the outer cell border directing toward the center of the gap-free area via fluorescence imaging and also via electron microscopic visualization (Figure 1D, untreated, white arrows). In contrast, after 12 h only 24 ± 1% of the gap area was closed and even 72 ± 3% of the gap still remained cell free after 24 h of incubation in the presence of pneumococci at an MOI of 10 (Figures 1B,C +WT). After differential immune fluorescence staining, microscopic magnification of a cell region within the cell layer beyond the cell-free gap displays numerous regions of cell lesions and adjacent cells generate long and thin stress fibers (Figure 1D, +WT, green arrows). This cell morphology predominantly appears in the presence of pneumococci. As shown in Figure 1D, yellow arrows point to attached pneumococci and internalized bacteria appear in red, as shown in Figure 1D (+WT). The electron microscopic magnification illustrates the attachment of pneumococci to HUVEC (yellow arrows in Figure 1D, +WT). The cells lack the typical cell morphology and appear as diffuse cell clusters with undefined cell borders.

The pore-forming toxin pneumolysin is described as the main inducer of cytolytic pore formation during pneumococcus infection and in particular as substance causing damages to the brain endothelium (Zysk et al., 2001). With the aim to clarify the impact of cytotoxic effects of pneumolysin on the endothelial gap closure, a pneumolysin-deficient isogenic derivative of the serotype 35A pneumococcus strain (Δply) was included in CSMA experiments.

Quantification of the cell-free area confirmed that the gap closure is significantly reduced in the presence of Δply pneumococci with an MOI of 10 reaching a maximum of 57.14 ± 11.5% gap closure after 24 h of cell cultivation (Figure 2A, dashed line). This retardation in gap closure is also illustrated by representative microscopic bright-field images (Figure 2C, Δply). Without bacteria, the gap is closed after 24 h of incubation (Figure 2A, black line and Figure 2C, untreated). Compared to the lowest gap closure of 28.69 ± 2.52% after 24 h of cell cultivation in the presence of the wild-type strain (Figure 2C, WT, black line with triangles and Figure 2C, +WT), the percentage of gap closure in the presence of Δply bacteria (57.14 ± 11.5%) is rather doubled. According to the curve slope between 6 and 24 h of cell cultivation, the velocity of gap closure in the presence of Δply pneumococci is strongly reduced compared to the speed of gap closure without bacteria (Figure 2A). With the aim to assess whether metabolic active bacteria are required to cause the monitored effects on endothelial cell growth, the same amount of heat-inactivated pneumococci was employed in the CSMA. In contrast to the lack of gap closure in the presence of wild-type pneumococci, a complete gap closure is reached after 24 h of incubation in the presence of heat-inactivated pneumococci, indicating that metabolic active pneumococci are required to interfere with cell growth (Figures 2A,C, +heat-inactivated WT, dotted line). Calculation of the velocity of gap closure for the whole incubation period of 24 h according to the method of monolayer edge velocimetry revealed the highest values for the speed of gap closure of HUVEC without pneumococci (Figure 2B, untreated, Bobadilla et al., 2019). Cell growth was significantly slowed down to one-third of the speed of HUVEC without bacteria in the presence of wild-type bacteria (Figure 2B, untreated, +WT). Without the cytotoxic effects due to pneumolysin production, the velocity of gap closure was only 1.7-fold less compared to untreated HUVEC (Figure 2B, +Δply). Fluorescence microscopy revealed that the endothelial cell morphology of HUVEC with pneumolysin-deficient pneumococci is similar to untreated HUVEC (Figure 2D, +Δply). Of note, despite the high amount of Δply bacterial attachment, which is visualized in the magnified image (Figure 2D, zoom in, 6 h), the actin stain revealed no substantial changes in endothelial cell morphology. At 200-fold magnification, attached pneumococci are visualized in yellow due to the fluorescence overlay after differential immunofluorescence staining (Figure 2D, +Δply, zoom in 6 h, yellow arrows). Thus, no stress fibers and no areas of cell lesions are detected in cell layers incubated with pneumolysin-deficient pneumococci (Figure 2D, +Δply, zoom in 6 h). In sum, despite the lack of substantial changes in cell morphology, the velocity of cell growth and the efficacy of gap closure are significantly reduced in samples with Δply bacteria even after 24 h. Therefore, the impact of the pneumolysin toxin itself on endothelial gap closure is further investigated in more detail.

To clarify the impact of pneumolysin-mediated cytotoxicity on the velocity and efficacy of cellular gap closure, we included pneumolysin in the CSMA. First, the activity of the purified pneumolysin protein was determined by standardized hemolysis analysis (Figure 3A). Incubation of the erythrocyte solution with a serial dilution of pneumolysin protein ranging from 2,500 ng/ml down to 19.5 ng/ml revealed that lysis of blood cells occurred in the presence of 156 ng/ml pneumolysin and higher protein amounts (Figure 3A). Based on the results of the lytic activity toward sheep erythrocytes, CSMA experiments were performed in the presence of 50, 250, 375, and 500 ng/ml pneumolysin protein (Figures 3B,C). In contrast to 100% gap closure, which is constantly reached by untreated HUVEC without toxin after 24 h, the lowest amount of 50 ng/ml pneumolysin reduced the mean value of gap closure to 96.58 ± 5.73% (Figure 3B, dashed line with squares). Only 83.73 ± 9.14% gap closure was achieved after 24 h in the presence of 250 ng/ml pneumolysin (Figure 3B, dashed line with crosses). At this toxin concentration, which already led to lysis of sheep erythrocytes, a substantial area remained cell free, although several endothelial cells remained attached to the slide surface (Figure 3C). A significantly less gap closure was monitored after HUVEC incubation with 375 ng/ml pneumolysin reaching values of only 39.14 ± 12.76% gap closure after 24 h (Figure 3B, dashed line with dots). Within the first 4 h, a maximal retardation in cell growth is detected in these samples, leaving the size of the cell-free area rather unchanged (Figure 3C). After longer incubation times, massive cell detachment occurred in the presence of 375 ng/ml pneumolysin, which effectively prevents any further cell growth (Figure 3C). Incubation of HUVEC with 500 ng/ml pneumolysin instantly induced massive cell damage and cell detachment and therefore gap closure remained unattainable (Figures 3A,B). Representative bright-field visualization confirmed the increasing cell damage and loss of cells already after 2 h of incubation with 500 ng/ml pneumolysin (Figure 3C). These data confirmed the substantial effect of the cytotoxin pneumolysin to gap closure of wounded endothelial cell layers.

Gap closure of the cell-free area might be achieved by two different cell growth mechanisms: cell migration and cell proliferation. We aimed to decipher the proportion of cell proliferation from total gap closure by applying the EdU-Click technique in the CSMA. This technique enables to determine the proportion of proliferating cells from the total amount of migrating cells at a specific time point. Cells with DAPI-stained blue fluorescent nuclei, which were present within the defined gap after 24 h of cell cultivation, were defined and counted as migrating but not proliferating cells. The incorporation of 5-ethynyl-2′-deoxythymidine into newly synthesized DNA enabled the microscopic detection of red fluorescent nuclei of proliferating cells within the defined cell-free area at 0, 6, and 24 h. In untreated samples, the amount of proliferating cells with red fluorescent nuclei is significantly increasing in the defined gap area over time, reaching an average amount of 374 proliferating cells (Figures 4A,B, untreated, red bars). In contrast, rather no cell proliferation is displayed in the presence of pneumococci (mean of 14 cells) and only an average amount of 50 red nuclei are detected in CSMA samples incubated with Δply–pneumococci, which is rather 7.5-fold less and significantly different from untreated samples (Figures 4A,B, +WT, +Δply). A mean value of 635 migrating but not proliferating cells with blue fluorescent nuclei was counted within the defined area of the closed gap after 24 h of HUVEC cultivation (Figure 4B, untreated, blue bar). Significantly less non-proliferating cells were detected in the presence of wild-type bacteria reaching only a maximum of 175 migrating cells within the gap area. This is rather four-fold less cell migration compared to the HUVEC sample without bacteria. Interestingly, in the presence of the pneumolysin-deficient strain, a mean value of 414 non-proliferating cells migrated into the gap area, which is significantly more than two-fold cell migration compared to the migration in the presence of the wild-type bacteria (Figure 4B, blue bars). In all samples with and without bacteria, the amount of cell proliferation was lower than the amount of cell migration. The data provide strong evidence that the incubation with pneumococci significantly affects both cell migration and cell proliferation of HUVEC, showing stronger effects on cell proliferation even in the absence of the cytotoxic effects of pneumolysin.

Flow-dependent vascular shear stress has substantial impact on cell proliferation and migration and is crucial for the maintenance of endothelial barrier integrity (Rousseau et al., 2000; Chistiakov et al., 2017). With the aim to establish a model system, which enables the microscopic visualization and quantification of the effect of pneumococcal virulence factors on endothelial migration and wound healing under physiological shear stress conditions, we combined the CSMA with the microfluidic pump system. As depicted in Figure 5A, the CSMA was initiated as described above by using two-chambered silicone inlets for standardized generation of a defined cell-free gap between two confluently grown endothelial cell layers. After covering the CSMA sample with a sticky slide, the whole slide was air free connected with the microfluidic pump system via Luer adaptors. The cells were further cultivated under defined shear stress of up to 10 dyn/cm². The slides are suitable for microscopic live cell imaging and for long-term flow cultivation of endothelial cells. Cell growth and gap closure were determined by microscopic monitoring. Flow speed and shear stress parameters are set and controlled via a special Pump software and mimicked the desired blood vessel situation.

Representative fluorescence images of endothelial gap closure in continuous and defined flow of 10 dyn/cm² indicated that HUVEC require 48 h to close the cell-free area, which doubled the time required for gap closure without flow (Figure 5B, untreated). Staining of the actin cytoskeleton with green fluorescent phalloidin illustrates an increasing cell density reaching a tightly packed cell layer after 24 h. Instead of the complete gap closure, which is reached by untreated HUVEC after 24 h of flow incubation (Figure 5B, untreated), the gap is not closed in the presence of pneumococci even after 48 h of constant flow incubation. In fact, the size of the cell-free area is substantially increased in the presence of pneumococci (Figure 5B, +WT, 24 h). The visual evaluation of the microscopic images indicates a reduction in the cell-free area after 24 h of incubation in flow even in the presence of wild-type pneumococci (Figure 5B, untreated and +WT). Moreover, the magnified images visualize representative cells of the cell area within the white squares, which are oriented in line with the flow direction (Figure 5B, untreated and +WT). After 6 h of flow incubation with circulating pneumococci, the cells appear in typical morphological shape, whereas after 24 h of flow cultivation, the phalloidin stain of the cellular actin cytoskeleton visualizes damaged cells at the wound borders, many lesions in the cell layer, and stress fibers (Figure 5B, +WT, 24 h). Finally, a massive cell detachment is microscopically detected after 48 h of bacterial incubation (Figure 5B, +WT, 48 h). These results demonstrate the substantial impact of pneumococci on endothelial growth under defined shear stress.

In addition to the static cell migration analyses, the combination of endothelial cell migration slides with a microfluidic pump system also enabled the simulation of tissue regeneration under physiological vascular shear stress of defined values. The combined technique serves as suitable model to analyze the effect of bacterial adherence on activated and damaged endothelial cells during sepsis in blood circulation.