Transactivated Epidermal Growth Factor Receptor Recruitment of α-actinin-4 From F-actin Contributes to Invasion of Brain Microvascular Endothelial Cells by Meningitic Escherichia coli

Abstract

Bacterial penetration of the blood-brain barrier requires its successful invasion of brain microvascular endothelial cells (BMECs), and host actin cytoskeleton rearrangement in these cells is a key prerequisite for this process. We have reported previously that meningitic Escherichia coli can induce the activation of host's epidermal growth factor receptor (EGFR) to facilitate its invasion of BMECs. However, it is unknown how EGFR specifically functions during this invasion process. Here, we identified an important EGFR-interacting protein, α-actinin-4 (ACTN4), which is involved in maintaining and regulating the actin cytoskeleton. We observed that transactivated-EGFR competitively recruited ACTN4 from intracellular F-actin fibers to disrupt the cytoskeleton, thus facilitating bacterial invasion of BMECs. Strikingly, this mechanism operated not only for meningitic E. coli, but also for infections with Streptococcus suis, a Gram-positive meningitis-causing bacterial pathogen, thus revealing a common mechanism hijacked by these meningitic pathogens where EGFR competitively recruits ACTN4. Ever rising levels of antibiotic-resistant bacteria and the emergence of their extended-spectrum antimicrobial-resistant counterparts remind us that EGFR could act as an alternative non-antibiotic target to better prevent and control bacterial meningitis.

Keywords: bacterial meningitis; cytoskeleton; epidermal growth factor receptor; invasion; α-actinin-4.

Figure 1

Meningitic E. coli PCN033-induced activation of EGFR contributes to its invasion of hBMECs. (A) Effects of the EGFR inhibitor gefitinib, AG1478 and the analog AG825 on PCN033 invasion of hBMECs. Data represent the mean ± SD of three duplications and are presented here as relative invasion compared with the vehicle control. (B) Verification of successful EGFR KO and its influence on bacterial invasion. (C) Tyrosine phosphorylation of EGFR during the infection. (D) EGFR activation in response to both viable PCN033 and heat-inactivated PCN033 strains. (E) Quantitation analysis of different EGFR ligands in response to PCN033 infection. GAPDH was used as the endogenous control. (F) Comparison of the induction of HB-EGF by the viable and inactivated PCN033 over time. (G) Effect of CRM197 treatment on the PCN033-induced EGFR activation. hBMECs were challenged with PCN033 for 3 h. (H) The effect of CRM197 on PCN033 invasion of hBMECs. The results represent the relative invasion percentage compared with the untreated control. ^*p < 0.05, ^**p < 0.01.

Figure 2

EGFR-ErbB3 dimerization induced by meningitic E. coli is important for EGFR activation and bacterial invasion of hBMECs. (A) Identification of the possible dimers involved in the EGFR response to meningitic E. coli infection. Infected cell lysates were immunoprecipitated with the anti-EGFR antibody, and then subjected to western blotting with different antibodies against ErbB. (B) Verification of the association between ErbB3 and EGFR via immunoprecipitation experiments with the anti-ErbB3 antibody. The hBMECs were challenged with PCN033 for 3 h. (C) Knock-down and verifying ErbB3 expression levels in hBMECs via the shRNA approach. β-actin and GAPDH were used as the endogenous controls for western blotting and quantitative PCR. (D) Comparison of the infection-induced EGFR activation in the control cells and the ErbB3 knocked-down cells. (E) Effect of the ErbB3 knock-down on meningitic E. coli (PCN033 and RS218 strains) invasion of hBMECs. ^**p < 0.01.

Figure 3

EGFR activation assists with PCN033-induced cytoskeleton fiber breakdown. The wild-type hBMECs (a–f) and the EGFR-KO cells (clones #4, g,h) were challenged with PCN033 with or without AG1478 (as described in Materials and Methods), and then subjected to confocal fluorescence microscopy. F-actin is labeled with actin-tracker green, and nuclei are stained with DAPI. The (a–f) were biological duplicates in each group, respectively.

Figure 4

Identification of ACTN4 as an EGFR interacting protein in hBMECs upon meningitic E. coli infection. (A) SDS-PAGE comparison of the immunoprecipitation products from uninfected and infected cells. Immunoprecipitation was conducted with an anti-EGFR antibody. (B,C) Immunoprecipitation and western blotting verification of the EGFR–ACTN4 interaction in response to infection with meningitic E. coli PCN033 (B) and RS218 (C). (D) Confocal immunofluorescence microscopy showed the meningitic E. coli co-located with EGFR and ACTN4. E. coli was labeled with Cy3, ACTN4 was labeled with FITC, and EGFR was labeled with Dylight 405. Cells were mounted and visualized using the Zeiss LSM 880 confocal system. The scale indicates 20 μm.

Figure 5

Meningitic E. coli-induced EGFR recruitment of ACTN4 is important for bacterial invasion. (A,B) The meningitic E. coli PCN033- (A) or RS218-induced (B) activation of EGFR results in increased recruitment of ACTN4 to EGFR and decreased binding of ACTN4 to F-actin, while inhibition of EGFR by AG1478 reversed this alteration. The densitometry results represent the ratio of EGFR to ACTN4 as well as F-actin to ACTN4. (C) Quantitative PCR and western blotting verification after ACTN4 genetic manipulation via CRISPR/Cas9. (D) Effect of the ACTN4 knock-down on meningitic E. coli PCN033 or RS218 invasion. Results are presented as the relative invasion percentage compared with that of the wild-type cells. (E) Effect of the ACTN4 knock-down on cell cytoskeleton morphology via confocal fluorescence microscopy. ^*p < 0.05, ^**p < 0.01.

Figure 6

S. suis SC19 infection can activate EGFR and competitively recruit ACTN4. (A) Confocal fluorescence microscopy showing the contribution of EGFR activity to S. suis SC19-induced F-actin breakdown and cytoskeleton rearrangement. F-actin was labeled with actin-tracker green, and nuclei were stained with DAPI. (B) Immunoprecipitation and western blotting revealed the increased recruitment of ACTN4 by EGFR over time during the infection. Anti-EGFR antibody was used for the immunoprecipitation experiments. (C) Immunoprecipitation and western blots showing the increased EGFR association with ACTN4, and the decreased association of F-actin with ACTN4 during the challenge infection with SC19. Immunoprecipitation was carried out using an anti-ACTN4 antibody.

Figure 7

Confocal fluorescence microscopy showing activated EGFR competitively recruiting ACTN4 from intracellular F-actin in response to meningitis-causing bacteria. The meningitic E. coli and S. suis infections were conducted with or without AG1478 pretreatment of the cells as follows. (A) Normal hBMECs without challenge, (B) The cells infected with PCN033, (C) The cells treated with AG1478 prior to PCN033 infection, (D) The cells infected with SC19, (E) The cells treated with AG1478 prior to SC19 infection. ACTN4 was labeled with Dylight 405, EGFR was labeled with Cy3, and F-actin was stained with actin-tracker green. The arrow heads indicate the co-location of ACTN4 and F-actin, and the arrows indicate the co-location of ACTN4 and EGFR. The cells were mounted and then visualized using the Zeiss LSM 880 confocal system.

Figure 8

Schematic representation of the mechanism involving EGFR in meningitic bacterial penetration of the BBB. Bacterial infection of hBMECs induces the ligand (HB-EGF)-dependent transactivation of EGFR, which requires the cleavage activity of ADAM17. Activated EGFR dimerizes with its heterogenous partner ErbB3, and then competitively recruits and disassembles ACTN4 from the intracellular F-actin fibers, leading to breakdown as well as reorganization of the actin cytoskeleton, which eventually facilitates bacterial invasion.