Colibactin in avian pathogenic Escherichia coli contributes to the development of meningitis in a mouse model

ABSTRACT Colibactin is synthesized by a 54-kb genomic island, leads to toxicity in eukaryotic cells, and plays a vital role in many diseases, including neonatal sepsis and meningitis. Avian pathogenic Escherichia coli (APEC) is speculated to be an armory of extraintestinal pathogenic Escherichia coli and can be a potential zoonotic bacterium that threatens human and animal health. In this study, the APEC XM meningitis mouse model was successfully established to investigate the effect of colibactin in in vivo infection. The clbH-deletion mutant strain induced lower γ-H2AX expression, no megalocytosis, and no cell cycle arrest in bEnd.3 cells, which showed that the deletion of clbH decreased the production of colibactin in the APEC XM strain. The deletion of clbH did not affect the APEC XM strain’s ability of adhering to and invading bEnd.3 cells. In vitro, the non-colibactin-producing strain displayed significantly lower serum resistance and it also induced a lower level of cytokine mRNA and few disruptions of tight junction proteins in infected bEnd.3 cells. Meningitis did not occur in APEC ΔclbH-infected mice in vivo, who showed fewer clinical symptoms and fewer lesions on radiological and histopathological analyses. Compared with the APEX XM strain, APEC ΔclbH induced lower bacterial colonization in tissues, lower mRNA expression of cytokines in brain tissues, and slight destruction of the brain blood barrier. These results indicate that clbH is a necessary component for the synthesis of genotoxic colibactin, and colibactin is related to the development of meningitis induced by APEC XM.

Avian pathogenic Escherichia coli (APEC), a principal member of the extraintestinal pathogenic Escherichia coli (ExPEC) group, induces severe respiratory and systemic diseases in poultry and leads to extensive economic losses. NMEC is another important member of the ExPEC group and is the most common secondary cause of central nervous system (CNS) infections in newborns with high morbidity and mortality [6]. Based on genotypic and phylogenetic group studies, both APEC and NMEC showed discernible phylogenetic overlaps and shared some virulence-associated factors, such as type 1 fimbriae, increased serum survival, and salmochelin [7,8]. Furthermore, several studies have demonstrated that APEC induces bacteremia or meningitis in neonatal rat or mouse models [9,10]. Therefore, APEC strains are speculated to be an armory of NMEC and can be potential zoonotic bacteria.
E. coli strains carrying the pks island can be isolated from multiple parts of the human body and confirm a natural transmission from mothers to their offspring [3]. The positive rates of pks + E. coli isolated from the gut are relatively low in healthy adults (19.7% [11] to 32% [12]) and neonates (26.9% to 33%) [13]. However, the percentage of E. coli harboring pks island, increases distinctly in infectious disease isolates. E. coli is considered to be responsible for urosepsis [14], prostatitis [15], septicemia [12], and newborn meningitis [16]. The pks island is also strongly linked with E. coli strains of the phylogroup B2 [11,13] and with several virulence factors (adhesins, hemolysins, toxins, and siderophores) [11]. Interestingly, a majority of the NMEC (67.92% to 78.8%) belong to the B2 phylogenetic group [17,18]. Several virulence factors related to the pks island have also been shown to be involved in infant meningitis [19]. Therefore, the function of the pks island has been investigated in the etiology of neonatal meningitis, such as in colonization in the immature gut, translocation to the bloodstream [20], apoptosis of T lymphocytes, and development of septicemia [21]. Furthermore, pks + Klebsiella pneumoniae has significant tropism toward the brain of BALB/c mice, and colibactin plays a key role in the pathogenic steps that lead to the development of meningitis [5]. However, the function of colibactin in the pathogenesis of E. coli meningitis is still unclear. As mature colibactin is still difficult to extract in the purified form from bacteria to date [22], it is difficult to investigate the role of colibactin in the development of meningitis. clbH belongs to the pks island and is involved in the formation of the genotoxic necessary AM-ACP unit [23]. Therefore, we selected clbH to block the synthesis of colibactin in the APEC XM strain. The E. coli meningitis mouse model was established to evaluate the role of colibactin in the development of meningitis in this study.

Ethics statement
The animal experiments followed the National Institute of Health guidelines for the ethical use of animals in China. All procedures were approved by the Animal Care and Ethics Committee of Yangzhou University. Four-week-old Institute of Cancer Research (ICR) mice were provided by the Comparative Medicine Center of Yangzhou University (License number: SCXK (Su) 2017-0007) and had free access to food and water under a 12 h light/dark cycle with observation twice a day. All manipulations were performed under anesthesia to minimize the suffering of animals. The mice were euthanized with overdose isoflurane exposure and samples were collected for analysis.

Strains, growth conditions and plasmids
The APEC XM strain (O2:K1) was isolated from the brain of a duck with symptoms of septicemia and meningitis (donated by Dr. Guoqiang Zhu, Yangzhou University). It grew aerobically on Luria-Bertani (LB) plates or in LB broth with agitation (180 rpm) at 37°C. When necessary, antibiotics were added with the following concentrations: ampicillin (100 μg/mL) or chloramphenicol (34 μg/mL). The strains and plasmids used in this study are listed in Table 1.

Construction of the clbH deletion mutant and complemented mutant
The deletion of clbH in the chromosome of APEC XM strain was achieved using bacteriophage λ Red recombinase system with primers and plasmid pKD3, pKD46, and pCP20 as described previously [24] (Table 1). For the construction of complemented mutant, the coding sequences of clbH gene were amplified from the APEC XM genome and cloned into plasmid pBR322. Polymerase Chain Reaction (PCR) and DNA sequencing confirmed modified genotypes in mutant strains. All primers used in this study are listed in Table 1.

Growth curves
Briefly, 200 μL (1 × 10 8 CFU/mL) bacteria in the exponential phase were inoculated in 20 mL LB medium with or without ampicillin in a 37°C shaking incubator at 180 rpm for 22 hours. The absorbance of bacterial culture was recorded per hour by spectrophotometer at 630 nm. The above experiments were repeated independently three times. The growth curves were drawn by GraphPad Prism 5.0 software (GraphPad Software).

Colibactin cytotoxicity assays
In the present study, bEnd.3 cells (American Type Culture Collection, ATCC CRL-2299) were used to demonstrate the cytotoxic effect of colibactin on eukaryotic cells. The cells were cultured in the DMEM (Gibco, 12800-017), supplemented with 10% heatinactivated fetal bovine serum (FBS; Gibco, 16140-071) at 37°C in a humidified 5% CO 2 atmosphere. The bEnd.3 cells (about 75% confluence) were infected with bacteria in the exponential phase with a multiplicity of infection (MOI) of 100. After 4 h infection, the cells were washed three times with PBS, and further incubated in DMEM with 10% FBS containing gentamicin (100 μg/mL) for the following analysis.
The bEnd.3 cells were observed for megalocytosis at 72 hours post-incubation (hpi) [25]. The cells were fixed with 4% paraformaldehyde for 20 min, and then stained with 0.1% methylene blue for 20 min.
The megalocytosis of cells were observed by an inverted microscope. Cytotoxic effects of colibactin produced by APEC XM, APEC ΔclbH, or APEC ΔclbH/pclbH were quantified by measurement of absorbance at 630 nm using a microplate reader. The expression of γ-H2AX in bEnd.3 cells were detected at 0 and 72 hpi. After washing 3 times, the cells were fixed with 4% paraformaldehyde for 20 min, permeabilized with 0.1% Triton X-100 for 20 min and processed for immunofluorescence following a standard protocol [26]. The primary antibody was a monoclonal rabbit anti phosphorylated H2AX (Cell Signaling Technology, #9718). The secondary antibody was a goat-anti-rabbit IgG (H + L) Alexa Fluor Plus 488 (ThermoFisher Scientific, A-21070). Then, the cells were stained with 4ʹ, 6-diamidino-2-phenylindole (DAPI; Beyotime Biotechnology, C1002). Finally, the coverslips were fixed using a fluorescence mounting medium. The GFP fluorescence was detected and photographed by a fluorescence microscope (Leica, Germany).
Four hours after the infection, the cell cycle of bEnd.3 cells were measured at 48 hpi [1,25]. The cells were collected, centrifuged at 400 g for 5 min at 4°C, washed with PBS, and resuspended in 70% ice-old ethanol for fixation at 4°C overnight. The cells were then centrifuged at 800 g for 10 min at 4°C, washed with PBS, and stained with FxCycle™ PI/RNase staining solution (Thermo Fisher Scientific, F10797) at room temperature for 15 min. The cell cycle was monitored on the BD LSRFortessa flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA) with 10,000 events/determination and analyzed with Flowjo software (Tree Star Inc.). The experiments were repeated three times independently.

Bacterial resistance to normal mouse serum
Serum resistance assay was performed in a 96-well plate as described previously [27]. Briefly, specific-pathogenfree (SPF) mouse serum was diluted to 50% with PBS. APEC XM, APEC ΔclbH, and APEC ΔclbH/pclbH strain grown to exponential phase were collected and washed twice with PBS. A dose of 10 μL culture suspension (OD 600 = 1.0) was inoculated into a 96-well plate containing 190 μL of 50% and 100% serum. After incubation for 0.5 h at 37°C, bacterial numbers were calculated using LB plates. The assay was performed in triplicate with three independent experiments.

Adhesion and invasion assay
For adhesion and invasion assay, the strains were grown in LB medium with or without ampicillin in a 37°C shaking incubator at 180 rpm until the optical density at 600 nm reached 1.0 (1 × 10 8 CFU/mL) in exponential phase. The bacteria were collected by centrifugation (3,500 rpm, 8 min), washed twice with phosphate-buffered saline (PBS), and resuspended in FBS-free DMEM. Then, bEnd.3 cells were infected with the APEC XM, APEC ΔclbH, or APEC ΔclbH/ pclbH strain at a MOI of 100 for 4 h at 37°C in 5% CO 2 . The mock-infection cells were cultured in FBSfree DMEM as the control. The bEnd.3 cells were gently washed with PBS three times to remove any nonadherent bacteria, and then lysed with 0.5% Triton X-100 for 30 min at 37°C. The suspensions were collected, serially diluted 10-fold, and plated on LB plates. After incubation overnight at 37°C, the number of CFUs was calculated.

Relative mRNA expression of cytokines and tight junction proteins in vivo and in vitro infection with qRT-PCR
In vitro infection, bEnd.3 cells were infected with the APEC XM, APEC ΔclbH, or APEC ΔclbH/pclbH strain at a MOI of 100 for 4 h at 37°C in 5% CO 2 . The mockinfection cells were cultured in FBS-free DMEM as control. The bEnd.3 cells were gently washed with PBS three times and the total RNA was extracted with TRIzol solution (Invitrogen, 15596-018). In vivo infection, the left hemisphere of brain was homogenized in TRIzol reagent and total RNA was extracted with TRIzol solution. The 900 ng of high-quality RNA was converted into cDNA by PrimeScript RT reagent Kit with gDNA Eraser (Takara, RR047A). qRT-PCR was performed on a CFX CONNECT Real-time PCR machine (Bio-Rad, CFX CONNECT, USA) using ChamQ SYBR qRT-PCR Master Mix (2×) (Vazyme, Q311-02) according to the manufacturer's instructions. The amplification cycles were performed as follows: 95°C for 10 min, followed by 40 cycles of 95°C for 30 s, 60°C for 30 s, and 72°C for 30 s, and a final extension at 72°C for 10 minutes. The 2 −ΔΔCt method was used to analyze the gene expression. The primer sequences of cytokines and tight junction proteins are shown in Table 1.

Expression of tight junction proteins examined in vivo and in vitro infection
In vitro infection, bEnd.3 cells were infected with the APEC XM, APEC ΔclbH, or APEC ΔclbH/pclbH strain at a MOI of 100 for 4 h at 37°C in 5% CO 2 . The mockinfection cells were cultured in FBS-free DMEM as control. The bEnd.3 cells were gently washed with PBS three times and total proteins were extracted from bEnd.3 cells using RIPA Lysate Buffer (Beyotime Biotechnology, P0013B). In vivo infection, total proteins were extracted from the brains using RIPA Lysate Buffer. The concentrations were determined with a bicinchoninic acid protein assay kit (Beyotime Biotechnology, P0010). After SDS-PAGE separated the total proteins, the proteins were transferred to polyvinyl difluoride membranes (Millipore, ISEQ00010). The membranes were incubated with 5% skim milk for 1 h. And then, the membranes were cultured with primary antibodies overnight at 4°C, including ZO-1 (1:1000; Invitrogen, Cat#61-7300), occludin (1:500; Invitrogen, Cat#71-1500), claudin-5 (1:50; Invitrogen, Cat#35-2500), and GAPDH (1:1000; Cell Signaling Technology, Cat#2118). The membranes were washed with Tris-buffered saline/ Tween (TBS-T) buffer and incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies (all at 1:10,000 dilution in 5% nonfat milk) at room temperature for 1 h. After washed with TBS-T, the membranes were incubated with enhanced chemiluminescence (Clinx Science Instruments, 1800212) for 30 s and detected by a chemiluminescence imaging system (Clinx Science Instruments, ChemiScope 5300, China). The band intensity was analyzed using a chemiluminescence imaging system (ChemiScope 5300; Clinx Science Instruments).

Construction of mouse meningitis model infected by E. coli
Animal infection experiments were carried out to determine the infection rate and colonization ability in the brain, blood, and lung. Briefly, forty 4-week-old ICR mice were randomly divided into four groups. Each mouse was intraperitoneally injected with a dose of 10 7 CFU in 100 μL normal saline or 100 μL sterilization normal saline [28]. After 8h post of infection (poi), clinical symptoms were observed per hour. Cerebrospinal fluid (CSF) samples were obtained by cisterna magna puncture with isoflurane inhalation anesthesia at 12 h poi. The whole blood samples were collected and treated with dipotassium ethylenediaminetetraacetic acid (K 2 -EDTA). Complete blood count test was performed using an automatic blood cell analyzer (Mindray, BC-1900, China). The brains tissues were collected, frozen Instantly in liquid nitrogen, and then stored at −80°C until used for detecting proinflammatory cytokines and tight junction proteins.

Evans blue (EB) permeability assay
At 30 min before euthanasia, the mice were injected with 2% Evans blue solution (100 µL per mouse) into the caudal vein. Afterward, the mice were anesthetized and perfused with 50 mL of ice-cold PBS. Brain tissues were homogenized in 1100 µL pre-cool PBS, and then centrifuged at 15,000 g for 30 min at 4°C [29]. Each 500 µL supernatant was added with an equal amount of 50% trichloroacetic acid. After 12 h incubation at 4°C, the mixtures were centrifuged at 15,000 g for 30 min at 4°C to separate the supernatants. The absorbance was measured at 630 nm using a spectrophotometer.

Bacterial loadings of blood, brain, lung, and CSF
The right hemisphere of the brain, lung, blood, and CSF samples were aseptically harvested and homogenized with sterile pre-cool PBS. After serial 10-fold dilutions in sterile PBS, 10 μL dilution was plated on MacConkey plates and cultured at 37°C. The bacterial loadings were calculated by CFU per gram of organs or per microliter of blood.

Brain histopathology
At 12 h poi, the brains were carefully collected and immediately fixed in 4% paraformaldehyde. After 48 h, the tissues were dehydrated by serial gradient alcohol and xylene, and then embedded in paraffin. The embedded tissues were cut into 4 µm paraffin sections by an automated microtome (Leica, Germany) and stained with hematoxylin and eosin afterward. The brain sections were observed and analyzed by microscope (Nikon, Eclipse 80i, Japan).

Magnetic resonance imaging (MRI) scan
The MRI scanning was performed on a 7.0-T MRI scanner (Bruker Corporation, BRUKER BIOSPEC 70/ 30, Germany). The mice were anaesthetized by isoflurane inhalation. Then, the heads were fixed with two flat head plastic thumbscrews and the mice were placed on a heating pad for maintaining body temperature within 36.5°C to 37.5°C. The mouse was monitored during the scanning, including saturation of pulse oxygen, heart rate, respiratory rate, and rectal temperature. The MRI sequences used in this study were T1-weighted imaging and meglumine gadopentetate-enhanced T1-weighted imaging.

Statistical analysis
Statistical analysis was performed using SPSS 16.0 (SPSS Inc.). Data showed as mean ± standard error of the mean from triplicate independent experiments. P-values were calculated using one-way ANOVA test. A p-value of less than 0.05 was considered statistically significant.

Results
clbH has no effect on APEC growth DNA sequencing and PCR (Figure 1 A) confirmed the deletion (APEC ΔclbH) and complemented (APEC (A) Identification of mutant strains using PCR. The APEC XM, APEC ΔclbH, and APEC ΔclbH/pclbH strains were amplified using primers P1/P2. (B) Growth curves of the three strains, APEC XM, APEC ΔclbH, and APEC ΔclbH/pclbH, grown in LB broth at 37°C under aerobic conditions. The optical density (OD) of the culture was measured at OD 630 . Each curve was presented as the average of the three independent experiments. There were no differences among the three strains.
ΔclbH/pclbH) strains. The DNA sequencing results also showed that the deletion and complemented strains were stable without any spurious mutations in the LB medium after 30 generations. As shown in Figure 1b, the growth curves of the APEC XM, APEC ΔclbH, and APEC ΔclbH/pclbH strains were recorded and depicted during the exponential growth and stationary phases. No significant differences were observed among the three strains. Therefore, clbH deletion and complemented strains were constructed, and clbH did not affect the ability of reproduction in APEC XM.

clbH is involved in colibactin production and elicits in vitro genotoxic effects
Cytotoxic effects of colibactin produced by the three strains were determined by quantification of H2AX phosphorylation, megalocytosis, and cell-cycle distribution. The percentage of γ-H2AX-positive cells was detected by an immunofluorescence assay as previously described [30]. The rate of increase in the γ-H2AX positive-cell numbers in the APEC XM group significantly increased at 0 and 72 hpi, compared to that in the control group. However, the expression of γ-H2AX in bEnd.3 cells infected with APEC ΔclbH was lower than that in the APEC XM group at 0 and 72 hpi (p < 0.01, Figure 2 A-D). Furthermore, there were no differences between the APEC ΔclbH and control groups at 72 hpi (p > 0.05; Figure 2 C, D). Additionally, the expression of γ-H2AX in the APEC ΔclbH/pclbH group was higher (p < 0.01; Figure 2 A-D) than that in the control group. Compared with that in the APEC XM group, the rate of γ-H2AX positive-cell numbers in the APEC ΔclbH/pclbH group significantly decreased (p < 0.01; Figure 2 A-D) at both time points indicating partial restoration of genotoxicity.
Meanwhile, APEC XM induced megalocytosis in bEnd.3 cells, which was characterized by a progressive enlargement of the cell body and nucleus, and is quantified using methylene blue staining. APEC ΔclbH resulted in fewer giant cells and higher absorbance of staining (p < 0.01; Figure 2 E, F) than the APEC XM group. Furthermore, APEC ΔclbH/pclbH caused higher absorbance of methylene blue staining than that in the APEC XM group (p < 0.01; Figure 2

Colibactin decreases serum resistance and does not contribute to the adhesion and invasion of bEnd.3 cells
E. coli requires a high degree of bacteremia, and binding to and invasion of brain microvascular endothelial cells before it can traverse the blood brain barrier (BBB) [31][32][33][34]. The serum resistance assay showed that the APEC ΔclbH strain displayed significantly lower serum resistance than the APEC XM strain inoculated into 100% mouse serum (p < 0.01; Figure 3 A). In addition, the adhesion and invasion assays showed that there were no significant differences in the binding and invasion of bEnd.3 cells among the APEC ΔclbH, APEC ΔclbH/pclbH, and APEC XM groups (p > 0.05; Figure 3 B).

Relative mRNA cytokine expression in in vitro infection
During the development of meningitis, E. coli induces an increase in the local production of inflammatory cytokines in endothelial cells [35]. After 4 h of infection, the relative expression of tumor necrosis factor alpha (TNF-α) (Figure 4 A), interleukin-1β (IL-1β) (Figure 4 B), and IL-6 (Figure 4 C) in the APEC ΔclbH infection group were significantly decreased, compared with those in the APEC XM group (p < 0.01), while the expressions of above cytokines in APEC ΔclbH/pclbH group were similar with those in APEC XM group.

mRNA and protein expression of tight junction proteins in infected bEnd.3 cells
The disruption of tight junction proteins from the BBB is an essential step in meningitis development [36]. The relative expressions of claudin-5 (p < 0.01; Figure 5

Colibactin plays an important role in the in vivo pathogenicity of APEC XM
The mice were monitored continuously, and their clinical symptoms were recorded after 8 h poi. Lethargy, Megalocytosis induced by colibactin was observed at 72 hpi by an inverted microscope (bar = 100 μm). (F) Quantification of megalocytosis was carried out by measuring the absorbance of methylene blue-stained cells at optical density (OD) of 600 nm. The data of the control group was set as 100%. (G) The cell cycles of bEnd.3 cells were assayed by flow cytometry (purple, G0/G1 phase; yellow, S phase; green, G2/ M phase). APEC XM and APEC XM ΔclbH/pclbH resulted in G2/M phase accumulation in infected cells at 48 hpi. (H) Quantification analyses of the cell cycle analysis of bEnd.3 cells in each group after 48 hpi. The results were analyzed with one-way ANOVA and presented as the mean ± standard errors of the mean for three independent experiments. (**, p < 0.01, *, 0.01 < p < 0.05, n.s., p > 0.05.)

Figure 3. Serum resistance assay and adhesion and invasion assays.
(A) For the serum resistance assay, the APEC XM, APEC ΔclbH, and APEC ΔclbH/pclbH strains were treated with 100% or 50% serum for 30 min, and bacterial numbers were calculated using LB plates. Data of the APEC XM were set as 100%. bEnd.3 cells were infected with a multiplicity of infection (MOI) of 100 bacteria/cell for 4 h in the APEC XM, APEC ΔclbH, or APEC ΔclbH/pclbH groups. For the control group, cells were not infected with any bacteria for 4 h. (B) For the adhesion and invasion assay, after 4 h of treatment and three washes with PBS, the bacteria were collected and analyzed using LB plates. The data of APEC XM was set as 100%. The data were analyzed with one-way ANOVA and expressed as the mean ± standard error of the mean for three independent experiments. (**, p < 0.01, *, 0.01 < p < 0.05, n.s., p > 0.05.) After 4 h of treatment and three washes with PBS, the relative abundance of tumor necrosis factor (TNF)-α (A), interleukin (IL)-1β (B), and IL-6 (C) mRNA transcripts in bEnd.3 cells were measured using qRT-PCR. The data were analyzed with one-way ANOVA and expressed as the mean ± standard error of the mean for three independent experiments. (**, p < 0.01, *, 0.01 < p < 0.05, n.s., p > 0.05.) unresponsiveness, lackluster coat, eyelid closure with thick red eye discharge, diarrhea, and neurological symptoms were observed in APEC XM-infected mice ( Figure 6A). The mice in the APEC ΔclbH/pclbH group showed clinical symptoms similar to those described above. In the APEC ΔclbH group, the mice presented with mild or no clinical signs ( Figure 6A).
The CSF staining confirmed that all mice in the APEC XM and APEC ΔclbH/pclbH groups had meningitis (Figure 6 B, C). No mice had meningitis in the APEC ΔclbH group (Figure 6 B, C). The bacterial loads in CSF samples from the APEC ΔclbH/pclbH group were similar to those from the APEC XM group (Figure 6 D). The complete blood count analysis showed that the absolute white blood cell and lymphocyte counts in the APEC XM and APEC ΔclbH/pclbH groups were lower than those in the control group (p < 0.01; Figure 6 D-F). There were no significant differences in absolute neutrophil counts among the APEC ΔclbH, APEC ΔclbH/pclbH, and control groups (p > 0.05; Figure 6 E-G).
At 12 h poi, bacterial loads in the lungs and blood significantly decreased in the APEC ΔclbH and APEC ΔclbH/pclbH groups, compared with those in the APEC XM group (p < 0.01; Figure 6 H, I). Importantly, no bacteria were isolated from brain tissue samples from the APEC ΔclbH or control groups (Figure 6 J). The APEC ΔclbH/pclbH bacterial load in the brain of mice was significantly reduced compared to that in the APEC XM group (p < 0.01; Figure 6 J).

Relative cytokine mRNA expression, pathological features, and MRI findings in mouse brains
The relative expression levels of IL-1β, IL-6, and TNF-α mRNA in the brain tissue samples were measured by qRT-PCR. In contrast to the control group, the relative expression of all detected cytokines increased significantly (p < 0.01; Figure 7 A-C) in the APEC XM group. Compared with that in the APEC XM group, there was a significant decrease (p < 0.01; Figure 7 A-C) in the mRNA expression of IL-1β, IL-6, and TNF-α in the APEC ΔclbH group, which was similar to that observed in the control group (p > 0.05; Figure 7 A-C). The relative mRNA expression of IL-1β, IL-6, and TNF-α increased   Histopathological analysis showed severe thickening of the pia mater and an infiltration of leukocytes in the pia mater, cerebral cortex, and medulla of the mouse brain induced by APEC XM (Figure 7 D). In addition, there were hemorrhages in the pia mater and an infiltration of leukocytes into the pia mater, cerebral cortex, and medulla in the APEC ΔclbH/pclbH group (Figure 7 D). No similar changes were detected in the mouse brains of the APEC ΔclbH or control groups (Figure 7 D).
MRI examination was performed to assess brain damage at 12 h poi. The enhanced T1W1 image demonstrated an abnormal contrast continuous linear enhancement of the pia mater (Figure 7 E, blue arrowhead) and a diffusion enhancing of the cerebral parenchyma in the APEC XM group. These abnormal MRI features partly decreased in the ΔclbH/pclbH group compared to those in the APEC XM group. In addition, these typical MRI findings for meningitis were not detected in the APEC ΔclbH group (Figure 7 E).

Deletion of clbH reduced in vivo BBB disruption
Evans blue staining revealed that the permeability of the BBB increased significantly in the APEC XM group (p < 0.01; Figure 8 A, B) compared to that in the control group. The permeability of the BBB decreased significantly in the APEC ΔclbH group compared to that in the APEC ΔclbH/pclbH or APEC XM groups (p < 0.01; Figure 8 A, B), which proved that the deletion of clbH reduced the disruption of BBB permeability by APEC XM.
Based on the results above, the changes in tight junctional proteins in vivo were also measured by Western blot and immunohistochemical staining. Western blot analysis showed that the protein expression of ZO-1, occludin, and claudin-5 decreased significantly in the brains of mice infected with APEC XM (p < 0.01) or APEC ΔclbH/pclbH (p < 0.01) compared with that in the control (Figure 8 C-F). Compared to the APEC XM group, no obvious in vivo disruption of ZO-1, occludin, and claudin-5 proteins was found in the APEC ΔclbH group (p < 0.01; Figure 8 C-F). There were no significant differences between the APEC ΔclbH and control groups (p > 0.05; Figure 8C-F).
Additionally, immunohistochemical staining confirmed that the expressions of ZO-1, occludin, and claudin-5 in the pia mater, cerebral cortex, and hippocampus of the brain were consistent with those obtained from Western blot (Figure 8 G, K, O). APEC XM and APEC ΔclbH/pclbH infection resulted in a significant decrease in ZO-1, occludin, and claudin-5 protein expression compared to that in the control group (p < 0.01; Figure 8 G, K, O). However, the deletion of clbH reduced the ability of APEC XM to disrupt ZO-1, occludin, and claudin-5 expressions. There were no significant differences in these proteins between the APEC ΔclbH and control groups (p > 0.01; Figure 8 G, K, O).

Discussion
Despite advances in clinical techniques and antibiotic therapies, E. coli meningitis remains a significant cause of mortality [37,38] and neurological disabilities in young infants [39]. Colibactin was first identified in an NMEC strain (IHE3034) [1]. Previous studies have primarily focused on gut homeostasis or colorectal cancer but are limited to neonatal meningitis. In our previous work, we found that the mRNA levels of pks island genes, which encode the non-ribosomal peptide synthetase-polyketide synthase (NRPS-PKS) complex for producing colibactin, significantly changed in APEC-XM during infection of bEnd.3 cells [40]. The APEC XM strain was isolated from the brain of a duck with neurological symptoms and septicemia, and the bacteria exhibited meningeal tropism in ICR mice [28,41]. Therefore, a mouse meningitis model was established to elucidate the possible effects of colibactin-induced E. coli meningitis.
The NRPS unit on the pks island is composed of clbH, clbJ, and clbN [1]. clbH has two N-terminal domains (C-A-T and A1-C-A2-T), and the noncanonical A1 domain activates L-serine to assemble the AM-ACP formation in an analogous manner to the zwittermicin biosynthetic enzymes [23]. Consequently, pks + E. coli mutants lacking any component of the AM biosynthetic machinery are not genotoxic, and AM-ACP formation is closely related to colibactin assembly [1]. In this study, clbH was deleted to construct a noncolibactin-producing strain. Deletion of clbH reduced colibactin production and genotoxicity in bEnd.3 cells, the pia mater (Black Arrowhead, bar = 100 μm). (E) Magnetic resonance imaging (MRI) analyses of mouse brains. Mouse brains were scanned using MRI T1W1 and enhancement T1W1. Abnormal contrast continuous linear enhancement of the pia mater, and diffusion enhancement of the cerebral parenchyma (Blue Arrowhead) and dilatation of the mesencephalic aqueduct (Blue Rectangular Frame) are observed. Figure 8. The disruption of blood-brain barrier in vivo.
In vivo disruption of blood-brain barrier (BBB) in 4-week-old ICR mice inoculated intraperitoneally with 10 7 colony forming units (CFU) of test bacteria for 12 h. The mouse of the control group injected intraperitoneally with an equal volume of sterile saline. (A) The representative bright-field photographs of Evans blue stained mouse brains were taken in dorsal view. (B) Evans blue extravasation from brain homogenates was quantitated at an optical density of 630 nm (OD 630 ). The data of the control group was set as 100%. Permeability of the BBB increased significantly in the APEC XM group compared to that in the control group. (C) Representative Western blots of ZO-1, occludin, and claudin-5 in vivo. Densitometric analysis of ZO-1 (D), claudin-5 (E), and occludin (F) proteins. Protein levels were determined and normalized to those of GAPDH. The data of the control group was set as 100%. ZO-1, occludin, and claudin-5 proteins decreased significantly in the APEC XM and APEC ΔclbH/pclbH groups, compared to that in the control group. The representative immunohistochemistry staining of ZO-1 (G), claudin-5 (K), and occludin (O) protein levels in the infected mouse brain sections (bar = 20 μm). The black arrowheads point to the pia mater, and the black arrows point to the microvasculature. Immunohistochemistry positive intensity data analysis of ZO-1 protein levels in pia mater (H), cerebral cortex (I) and hippocampus (J); claudin-5 protein levels in the pia mater (L), cerebral cortex (M), hippocampus(N); and occludin protein levels in the pia mater (P), cerebral cortex (Q), and hippocampus (R). The data of the control group was set as 100%. The results were analyzed with one-way ANOVA and presented as the means ± standard error of the mean for three independent experiments. (**, p < 0.01, *, 0.01 < p < 0.05, n.s., p > 0.05.) which was similar to results obtained from previous studies on clbA [42] or clbP [16]. As shown in this study, the deletion of clbH affected both the cytotoxicity to cells in vitro and the pathogenicity to bEnd.3 cells and newborn mice. Many studies have revealed that the successful crossing of the BBB by E. coli requires three key steps: a degree of bacteremia, E. coli binding to and invasion of brain microvascular endothelial cells, and traversal of the BBB [31][32][33][34]43,44]. In this study, we found that deletion of clbH did not affect the adhesion and invasion to bEnd.3 cells by APEC XM. Bacterial colonization in organs is concomitant with the capacity to cause bacteremia and systemic infection [5]. Unlike the effective colonization in the APEC XM group, the bacterial loads in the APEC ΔclbH group were significantly decreased in the lung, blood, CSF, and brain samples. The mouse serum resistance assay confirmed that APEC ΔclbH had lower serum resistance than the APEC XM strain. In line with experimental pks + septicemic mice [21], APEC XM and APEC ΔclbH/pclbH induced profound lymphopenia in meningitis-affected mice, which was alleviated in APEC ΔclbH-infected mice. Lymphopenia might reduce the survival rate of mice or humans with sepsis and meningitis induced by pks + E. coli [21]. Bacterial colonization of the brain is another important step in meningitis development [39]. Brain injuries are a hallmark of meningitis, including necrotic cortical injury and apoptotic hippocampal injury [45,46]. E. coli meningitis resembled other bacterial meningitis in the MRI scanning and histopathological characteristics, such as abnormal contrast continuous linear enhancement and severe thickening of the pia mater, leukocyte infiltration into brain tissue, and hemorrhage [47]. However, none of these histopathological and MRI findings were found in the APEC ΔclbH group. clbH is a necessary component for the synthesis of genotoxic colibactin, which is also strongly associated with meningitis induced by APEC XM.
During the development of meningitis, E. coli induces an increase in the local production of inflammatory cytokines in endothelial cells [35], microglia [48], and astrocytes [49]. The expression of TNF-α, IL-6, and IL-1β in the APEC XM group was in line with that previously reported in in vivo or in vitro E. coli meningitis studies [50][51][52]. Overexpression of proinflammatory cytokines could recruit leukocytes into the CNS to create a "cytokine storm" for exaggerated immune responses and CNS damage. In addition, inflammatory mediators play a role in BBB integrity [53,54]. IL-1β contributes to macrophage recruitment, Streptococcus pneumoniae clearance [55], and protects mice from lethal gramnegative infections [56]. At the peak of in vivo IL-1 expression, there is marked recruitment of neutrophils, breakdown of the BBB, and vasodilatation [57]. High levels of IL-1 in the CSF correlate with the development of neurological complications [58]. TNF-α is another important early response cytokine and is related to a fatal outcome in meningitis [58]. Both IL-1β and TNF-α are bone marrow stimulants that grow in a number of myeloid progenitors and promote the recruitment of neutrophils at the inflammation site [58]. TNF-α and IL-1β activate the p38/ ERK1/2 pathway and increase myosin light chain kinase [59]. TNF-α also activates the Hif-1α/ VEGF/ VEGFR-2/ ERK signaling pathway to decrease the expression of occludin in human cerebral microvascular endothelial cell lines [60]. IL-6 is a pleiotropic cytokine with both proinflammatory and antiinflammatory effects. It participates in inflammation, immune response, and hematopoiesis [61] and appears to be a good marker of severity during bacterial infection [62]. IL-6 has emerged as a pivotal player in neuroinflammation due to its influence on the three key branches of this process: astrogliosis [63], microgliosis [64], and BBB integrity. IL-6 increases endothelial permeability and produces ZO-1 mislocalization, actin structure remodeling, and increase in cell contraction [65]. As an antiinflammatory cytokine in neuroinflammatory conditions, IL-6 maintains BBB integrity by influencing endothelial cells and astrocytes [66]. Inflammatory stress, by using one or a combination of IL-17, IL-6, and/or TNF-α, could lead to the opening of the BBB in the bEnd.3 cell model, which is reflected by a significant increase in permeability and decrease in ZO-1 and claudin-5 [67]. In this study, APEC ΔclbH did not evoke TNF-α, IL-6, and IL-1β in the mouse brain or in bEnd.3 cells, which might have less ability to damage BBB integrity.
The BBB regulates the components in the CNS and minimizes the transfer of toxic compound pathogens to the CNS. BBB disruption is an essential step in meningitis development [36]. The state of brain capillaries and their polarized microvascular endothelial cells is responsible for the BBB structure and functional integrity by possessing tight junctions [68]. Occludin, claudin, junctional adhesion molecules (JAMs), and ZO-1 are the main elements of intercellular tight junction proteins and control the paracellular passage of substrates across the BBB [69]. In this study, we used ZO-1, claudin-5, and occludin to evaluate BBB breakdown. The in vitro disruption of ZO-1 and claudin-5 proteins in bEnd.3 cells infected with APEC XM was significantly alleviated in the APEC ΔclbH group. In addition, the relative mRNA expression of tight junction proteins in the three strain groups was similar, and decreased significantly compared with that in the control group. In the mouse model, vascular leakage was reduced and tight junction protein breakdown decreased in the APEC ΔclbH group. Further, all the tight junction proteins mentioned above decreased at the transcript and/or protein levels in brain endothelial cells in Group B Streptococcus [70][71][72], Neisseria meningitidis [73], Streptococcus suis [74], or E. coli [75] infection. In vivo, immunohistochemistry and Western blot assays also showed no significant disruption of tight junction proteins in the APEC ΔclbH group. Thus, colibactin might play an important role in disordering tight junction proteins in the BBB. ZO-1 is located on the cytoplasmic surface of endothelial cells. It serves as a recognition protein for tight junction placement and as a support structure for signal transduction [76]. Altering the structure or localization of ZO-1 protein leads to tight junction disconnection, opening of the intercellular gap, and increased BBB permeability. Claudin-5 is present in both human and mouse early fetal brain vessels and continues to increase during postnatal development and maturation of the BBB [77,78]. It is localized specifically to the endothelial cell layer in the brain and is also the most enriched tight junction protein at the BBB [79]. Dysfunction of claudin-5 protein is associated with either neurodegenerative [80], neuroinflammatory [79], or psychiatric disorders [81], and with CNS bacterial infections. Occludin is a central regulatory element in the assembly and function of tight junction proteins, and it is also required for cytokine signal transduction in cells such as for TNF-α and IFNγ [82]. Many studies have demonstrated that changes in redox conditions [83] and interactions with a wide range of kinases and phosphatases [84] can transform occludin domains to disrupt barrier functions.
As it is still difficult to extract pure mature colibactin from bacteria to date [22], it is difficult to investigate the direct role of colibactin on BBB integrity and induction of inflammation in the brain tissue by the tests used in this study. If colibactin can be purified in the future, it could be injected directly into the brain and its pathological effects on the brain tissue may be analyzed. In summary, colibactin is a key virulence factor for APEC XM to induce meningitis. It is responsible for increasing the inflammatory response and decreasing tight junction proteins expression in in vitro and in vivo infection, which is associated with blood survival mediated by colibactin.

Disclosure statement
No potential conflict of interest was reported by the authors.

Data Availability Statement
The datasets generated and/or analyzed during the current study are not publicly available due to the project is not finished yet but are available from the corresponding author on reasonable request.