|Home | About | Journals | Submit | Contact Us | Français|
Elevated levels of pterins and nitric oxide (NO) are observed in patients with septic shock and bacterial meningitis. We demonstrate that Escherichia coli K1 infection of human brain microvascular endothelial cells (HBMECs) induces the expression of guanosine triphosphate cyclohydrolase (GCH1), the rate-limiting enzyme in pterin synthesis, thereby elevating levels of biopterin. DAHP (2,4-diamino hydroxyl pyrimidine), a specific inhibitor of GCH1, prevented biopterin and NO production and invasion of E. coli K1 in HBMECs. GCH1 interaction with Ecgp96, the receptor for outer membrane protein A of E. coli K1, also increases on infection, and suppression of Ecgp96 expression prevents GCH1 activation and biopterin synthesis. Pretreatment of newborn mice with DAHP prevented the production of biopterin and the development of meningitis. These results suggest a novel role for biopterin synthesis in the pathogenesis of E. coli K1 meningitis.
Escherichia coli (E. coli) K1 is the bacterium that most commonly causes neonatal meningitis in industrialized countries . The morbidity and mortality rates associated with neonatal meningitis have not changed for last 2 decades despite the use of effective antibiotics and supportive care . This poor outcome is attributed to an incomplete understanding of the pathogenesis and pathophysiology of the disease, which inhibits the development of novel therapeutic strategies for prevention. Besides antimicrobial therapy, prevention of E. coli K1 traversal to the brain has become another area of focus for preventing meningitis. E. coli K1 translocation of the blood-brain barrier and invasion of human brain microvascular endothelial cells (HBMECs), a single-cell lining of the blood-brain barrier, requires unique interactions between the bacterial determinants and their cognate host receptors. Previous studies have demonstrated that Ecgp96, a 96-kDa glycoprotein, serves as a receptor in HBMEC to outer membrane protein A (OmpA) of E. coli K1 for binding to and invasion of the bacterium .
Additionally, E. coli K1 induces tight junction disruption and blood-brain barrier leakage by triggering iNOS activation and consequently nitric oxide (NO) production . Tetrahydrobiopterin (BH4), a pterin analogue and an obligate cofactor for all the isoforms of NOS, primarily controls NO production . Pterin (biopterin and neopterin) production occurs mainly by de novo synthesis, using guanosine triphosphate (GTP) as a source and GTP cyclohydrolase (GCH1; EC 188.8.131.52) is the first and rate-limiting enzyme in this reaction . Human GCH1 exists in its native form (approximately 28 kDa) and assembles to a homo-decamer to catalyze GTP to pterins . The vital role of GCH1 and pterins in cell differentiation, pain modulation and mRNA stability is well documented [8–10]. Furthermore, elevated levels of pterin are used as a diagnostic marker in infections caused by intracellular pathogens and malignant tumors . Lipopolysaccharide (LPS), interferon γ (IFN-γ), and tumor necrosis factor α (TNF-α) stimulate de novo synthesis of pterins in various immune and endothelial cells .
During sepsis, iNOS-dependent NO production and subsequent vasodilation is a major factor responsible for persistent hypotension . Of note, elevated pterin levels in cerebrospinal fluid (CSF) of pediatric patients with bacterial meningitis were also observed . Therefore, we hypothesize that E. coli K1 triggers NO production not only by activating iNOS expression but also by modulating pterin synthesis. Since endothelial cells have been reported to produce low levels of neopterin , we have focused only on biopterin production in this study. Here, our studies demonstrate that E. coli K1 invasion of HBMECs depends on the expression of GCH1 and biopterin synthesis. Importantly, we found that pretreatment of newborn mice with the GCH1-specific inhibitor 2,4-diamino hydroxyl pyrimidine (DAHP) protects the animals from E. coli K1 meningitis.
E. coli (OmpA+ E. coli), a spontaneous rifampin-resistant mutant of strain RS 218 (serotype O18:K1:H7), was isolated from the CSF of a neonate with meningitis . OmpA− E. coli is a mutant of RS218 that does not express OmpA or invade HBMECs. Antibodies to GCH1, iNOS, and β-actin were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-Ecgp96 antibody was generated as previously described . Fluorescent-tagged secondary antibodies were purchased from Invitrogen (Carlsbad, CA). DAHP and Lipofectamine were obtained from Sigma (St. Louis, MO). Griess reagent was purchased from Promega (Madison, WI). We purchased 6-methylpterin (internal standard), d-neopterin, and l-biopterin from Shricks Laboratories (Jona, Switzerland). Ascorbic acid was obtained from Calbiochem (La Jolla, CA). Lugol's iodine was purchased from Electron Microscopy Sciences (Hatfield, PA). Antibodies to GFAP and MPO were obtained from Leica Microsystems (Buffalo Grove, IL).
HBMECs were isolated, cultured, and maintained as described elsewhere . The frozen stocks of HBMECs were revived, characterized for brain endothelial cell markers and used for invasion assays. Total cell association (represented as binding) and invasion assays were described previously . Cytotoxicity of DAHP was assessed using a Cytotox 96 assay kit (Promega) according to the manufacturer's protocol.
The total biopterin level was measured using the protocol of Fukushima and Nixon, with minor modifications . The whole sample preparation was performed in a dark environment, using brown centrifuge tubes. Total lysates of HBMECs and brain tissues were treated with 50 µL of 10% trichloroacetic acid on ice for 30 minutes. Samples were centrifuged at 10 000 × g at 4°C for 10 minutes, and the supernatants were subjected to iodine oxidation. A total of 125 µL of sample and the standard containing 200 nM neopterin, 100 nM biopterin, and 10 µL of internal standard (6-methylpterin) were mixed. Thirty microliters of 1 N HCl was added, and the oxidation was started by the addition of 10 µL of acidic iodine solution. Tubes were kept at room temperature for 60 minutes, and then 15 µL of ascorbic acid was added to reduce excess iodine. A total of 100 µL of the solution was subjected to high-performance liquid chromatography (HPLC; Thermo Fisher Scientific, Hercules, CA). Separation was performed on a C18 Spherisorb 5-µm precolumn (10 × 4.6 mm) and a ODS-1 Spherisorb analytical column (250 × 4.6 mM; Waters, Milford, MA), using 1.25 mmol/L potassium hydrogen phosphate buffer, with 6% (v/v) methanol at a flow rate 1.3 mL/minutes.
Plasma membrane and cytosolic fractions from HBMECs were isolated as described earlier, using the BioVision kit . For immunoprecipitation, cytosolic fractions (300 µg) or plasma membrane fractions (80–100 µg) were incubated with appropriate antibodies overnight at 4°C, washed, and incubated for 2 hours with protein A agarose beads. The beads with bound immune complexes were washed, boiled for 10 minutes in sodium dodecyl sulfate (SDS) sample buffer, and separated by 10% SDS–polyacrylamide gel electrophoresis (SDS-PAGE). For identification of iNOS dimerization, samples were dissolved in sample buffer (devoid of β-mercaptoethanol) without boiling, and SDS-PAGE was performed at 4°C . Western blotting was done as described previously . Protein expression levels were quantitated using ImageJ software (http://www.rsbweb.nih.gov/ij/).
To detect the expression of GCH1 and Ecgp96, HBMECs were infected with OmpA+ E. coli and OmpA− E. coli, with or without DAHP pretreatment for various periods, and the cells were subjected to flow cytometry as described previously .
HBMECs at 60% confluence were transfected with 40 pmol of Ecgp96 siRNA (catalog no. HSS110955; Invitrogen) by use of Lipofectamine, according to the manufacturer's instructions. The efficiency of silencing was verified by Western blotting of total HBMEC lysates.
Total nitrite (NO2–) content was used as an index of NO production. Nitrite levels were determined in supernatants of cell cultures by spectrophotometry, using Griess reagent (Promega, Madison, WI) and using sodium nitrite as a standard according to the manufacturer's protocol.
Animal studies were approved by the Institutional Animal Care and Use Committee (IACUC) of Children's Hospital of Los Angeles and followed guidelines for the performance of animal experiments, as mandated by the office Laboratory Animal Welfare at the National Institutes of Health. C57BL/6J timed-pregnant mice were purchased from Jackson Laboratories (Bar Harbor, ME). One litter of 3-day-old mouse pups was infected intranasally with 103 colony-forming unites (CFU) of bacteria as previously described . Control mice received pyrogen-free saline through the same route. To evaluate the effect of DAHP on OmpA+ E. coli meningitis, DAHP (100 mg/kg body weight) was injected intraperitoneally 6 hours before, at the time of, and 6 hours after infection. On the basis of a previous report  and dose optimization studies, a dose of 100 mg/kg was used for this study. CSF samples were collected aseptically under anesthesia by cisternal puncture and were directly inoculated into rifampicin-containing LB. Growth of E. coli in broth from CSF samples was considered positive for meningitis. Whole brain was aseptically removed. One half of the brain was homogenized in sterile phosphate-buffered saline to determine the bacterial count, and the other half was stored in formalin for histopathologic analysis. Bacterial counts in brain were determined by plating 10-fold serial dilutions on rifampicin LB agar plates.
All data were derived from at least 3 independent experiments. Statistical analyses were conducted using SigmaPlot software (version 11.0). Significant differences (P < .05) between the groups were determined using the unpaired Student t test. For animal studies, statistical significance was determined using analysis of variance.
Our previous findings showed that the expression of OmpA in E. coli K1 is important for the invasion of HBMECs . Therefore, we examined the effect of OmpA+ E. coli and OmpA− E. coli K1 infection on GCH1 expression and biopterin production. OmpA+ E. coli K1 infection significantly induced GCH1 expression in a time-dependent manner, which peaked 60 minutes after infection and started declining by 90 minutes (Figure 1A). OmpA− E. coli K1 infection showed only a moderate effect on GCH1 expression. The increase in GCH1 expression was 2.5-fold greater during OmpA+ E. coli K1 infection than in control cells, after normalizing to β-actin levels, as measured by densitometry of the protein bands (Figure 1B). Flow cytometry of HBMECs infected with the bacteria after staining with anti-GCH1 antibodies revealed that OmpA+ E. coli induced greater levels of GCH1 expression 30 and 60 minutes after infection as compared to control or OmpA− E. coli–infected cells (Figure 1C and and11D). Next, the total biopterin level was measured by HPLC 60 minutes after infecting HBMECs with OmpA+ E. coli or OmpA− E. coli K1. Cells infected with OmpA+ E. coli K1 showed a significant increase in the quantities of total biopterin as compared to uninfected control HBMECs (P < .05). However, no significant increase in biopterin levels was observed in HBMECs infected with OmpA− E. coli K1 (Figure 1E). These results suggest that E. coli K1 infection stimulates biopterin synthesis by inducing GCH1 expression and that the presence of OmpA in E. coli is critical for this event to occur.
To analyze the role of GCH1 and biopterin in binding to and invasion of HBMECs by E. coli K1, the cells were pretreated with DAHP for various times and at different concentrations. A time- and dose-dependent inhibition in both binding and invasion of E. coli K1 was observed (Figure 2A and B). HBMECs treated with 5 mM DAHP for 3 hours reduced the binding and invasion of E. coli K1 by 45.6% and 56.2%, respectively, compared with control cells (mean ± SD CFU/well, 1.93 × 103 ± 0.4 × 103; P < .05). Interestingly, DAHP cotreatment (0-hour treatment) along with E. coli K1 infection also showed significant inhibition of binding and invasion (16.5% and 46.8%, respectively). DAHP treatment of HBMECs for 6 hours showed minimal cytotoxicity, even at a concentration of 7.5 mM (Figure 2C) and no antimicrobial activity (data not shown). Pretreatment of HBMECs with 5 mM of DAHP for 3 hours was found to be the minimal condition for 50% inhibition of invasion by E. coli K1. Hence, the same time and dose of DAHP treatment were used for further experiments.
We have previously reported that E. coli K1 triggers NO production via iNOS for invading HBMECs . Of note, dimer formation is critical for sustained enzymatic activity of iNOS and mainly depends on a cofactor like BH4 . Since E. coli K1 infection stimulates total biopterin production, we analyzed the monomer/dimer status of iNOS in HBMECs. A time-dependent increase in iNOS dimer formation was observed in HBMECs infected with OmpA+ E. coli K1 (Figure 2D). Similarly, the total iNOS expression also increased, beginning 30 minutes after infection. Pretreatment of HBMECs with DAHP (5 mM) for 3 hours attenuates both iNOS monomer and dimer expression. In parallel, production of NO (measured as the total nitrate level) induced by E. coli K1 was significantly diminished by DAHP pretreatment (Figure 2E). These results suggest that biopterin synthesis is necessary for increased production of NO during OmpA+ E. coli infection of HBMECs.
It was demonstrated that only the GCH1 decamer exhibits enzyme activity for catalysis of GTP to pterin . Chaperones play a vital role in protein maturation, folding, and multimerization , and Ecgp96 (also known as “HSP90β,” “gp96,” and “GRP94”) is a very abundant molecular chaperone in endoplasmic reticulum . Hence, we hypothesized that the association of Ecgp96 and GCH1 is essential for the enzymatic activity of GCH1. To explore this, cytosolic and membrane fractions were prepared from HBMECs infected with OmpA+ E. coli or OmpA− E. coli and were immunoprecipitated with anti-GCH1 antibodies, followed by immunoblotting with anti-Ecgp96 antibodies. In cytosolic fractions, a significant association of GCH1 and Ecgp96 was observed in OmpA+ E. coli–infected cells, with maximal interaction at 60 and 90 minutes, as determined by densitometric analysis (Figure 3A). Since OmpA− E. coli showed minimal stimulation of GCH1 expression and Ecgp96, it was not surprising that a slight increase in association of GCH1 and Ecgp96 was also observed at 30 minutes. To reconfirm this observation, reverse immunoprecipitation was performed using anti-Ecgp96 antibodies, followed by immunoblotting with anti-GCH1 antibodies. Maximum association of Ecgp96 with GCH1 was observed in HBMECs 60 and 90 minutes after infection with OmpA+ E. coli.
Our previous studies showed surface translocation of Ecgp96 from cytosol during E. coli K1 infection in HBMECs . Thus, it was of interest to analyze whether these 2 proteins also associate at the membrane level. Immunoprecipitation of membrane fractions of infected cells revealed a significant association of GCH1 and Ecgp96. However, unlike the interaction observed in the cytosol, the maximum association began at 15 minutes and was maintained until 90 minutes after OmpA+ E. coli infection (Figure 3B). In contrast, OmpA− E. coli–infected HBMECs only showed a basal level of association. Collectively, these data demonstrate that increased expression and association of Ecgp96 and GCH1 occurs in the cytoplasm and that a portion of the complexes translocate to the plasma membrane of HBMECs during infection with OmpA+ E. coli.
OmpA+ E. coli infection induces total and cell surface expression of Ecgp96 in HBMECs, mediated by increased production of NO . Since pretreatment of DAHP attenuates NO production, we sought to analyze the effect of DAHP on the total and surface expression of Ecgp96. As shown in Figure 4A, OmpA+ E. coli activates total expression of Ecgp96 15–60 minutes after infection. Pretreatment of HBMECs with 5 mM of DAHP for 3 hours significantly reduced the expression of Ecgp96 despite infection with OmpA+ E. coli. Densitometric analysis of the protein bands revealed a 2.5-fold increase in the expression of Ecgp96 in HBMECs infected with OmpA+ E. coli, compared with control, which was inhibited by treating the cells with DAHP (Figure 4B). Flow cytometry of HBMECs revealed increased expression of Ecgp96 on the membranes 30–60 minutes after infection, while DAHP-treated and infected HBMECs exhibited significantly reduced expression of Ecgp96 by 60 minutes after infection (P < .05; Figure 4C). These results suggest that GCH1 regulates Ecgp96 expression and plasma membrane translocation via NO production.
We next sought to evaluate GCH1 expression and biopterin synthesis after silencing Ecgp96 expression, using siRNA. HBMECs transfected with 40 pmol of Ecgp96 siRNA suppressed the expression of Ecgp96 by approximately 75%, on the basis of densitometric analysis (Figure 5A). Untransfected HBMECs revealed increased expression of GCH1 with E. coli K1 infection, beginning 15 minutes after infection. However, such an increase of GCH1 expression was inhibited in Ecgp96-siRNA/HBMECs, despite infection with OmpA+ E. coli (Figure 5B). Densitometry analysis of GCH1 expression revealed a 2.5-fold decrease in expression, which was similar to that for control cells (Figure 5C). Furthermore, Ecgp96-siRNA–transfected cells with and cells without OmpA+ E. coli infection were analyzed for total biopterin production (Figure 5D). In control HBMECs, the total biopterin level was found to be 476.2 pmol/mg of protein, and the level decreased to 351.6 pmol/mg protein on silencing of Ecgp96 expression. Similarly, in HBMECs infected with OmpA+ E. coli, the total biopterin level declined from 676.8 pmol/mg protein to 365.6 pmol/mg protein on introduction of Ecgp96-siRNA into HBMECs (P < .05). Together, these results support our hypothesis that Ecgp96 plays vital role in cellular biopterin production by regulating the activity of GCH1.
We further examined the role of biopterin synthesis in our well-established newborn mouse model of meningitis. A significant elevation in total biopterin production was noticed in the brains of OmpA+ E. coli–infected pups, compared with controls (Figure 6A). Additionally, the mean bacterial load ( ± SD) in the brains of OmpA+ E. coli–infected newborn mice was found to be 5.22 ± 1.9 log CFU/g brain tissue at 72 hours (Figure 6B). In agreement with this finding, 100% of infected pups had CSF cultures that were positive for E. coli K1 (which indicates positivity for meningitis; Figure 6C). Of note, brains of newborn mice treated with DAHP had no positive CSF cultures and had decreased levels of biopterin induced by OmpA+ E. coli infection. Analysis of brain tissue sections of OmpA+ E. coli–infected pups showed significant brain damage by extensive infiltration of glial cells and neutrophils, compared with brain sections of uninfected pups (Figure 6D). DAHP pretreatment prevented the abnormalities induced by OmpA+ E. coli infection in the brains of newborn mice. These results suggest that GCH1 activity is important for the development of meningitis due to E. coli K1 in newborn mice.
Our previous studies have demonstrated that E. coli K1 induces NO production via OmpA interaction with the receptor Ecgp96 in HBMECs . Here, we show that OmpA+ E. coli infection of HBMECs enhances the expression of GCH1, biopterin synthesis, and, concurrently, NO production. Although LPS has been shown to increase GCH1 expression, we observed only a moderate effect during OmpA− E. coli infection. Our unpublished data revealed that OmpA+ E. coli induces the expression of Toll-like receptor 2/Ecgp96 complexes on HBMEC plasma membranes, whereas OmpA− E. coli enhances mainly Toll-like receptor 4 (TLR4) expression. Therefore, it is possible that the moderate increase in GCH1 expression and biopterin production observed in OmpA− E. coli–infected HBMECs could be due to the interaction of LPS with TLR4. Since GTP is the substrate for cellular pterin synthesis, it should be noted that the GTP level is inversely proportional to the cellular pterin level, which in turn affects G-protein signaling and small-GTPase activation. This sequence of events could be the basis of our earlier observation that OmpA+ E. coli decreases the activity of the small GTPase Rac1 for cytoskeleton rearrangements .
Homo-decamerization of GCH1 is a critical factor in its activation, indicating that assembly of the GCH1 monomer into larger units is necessary for function. Hwu et al have reported the role of Hsp90, a heat shock protein, in regulating GCH1 expression and activity . Although Ecgp96 is a paralogue of the Hsp90β form, this study revealed that Ecgp96 also associates with GCH1 and that their levels increased significantly on OmpA+ E. coli infection of HBMECs. In addition, Ecgp96/GCH1 complexes also translocated to the cell surface during the invasion process. Coordinated expression of GCH1 and iNOS under inflammatory conditions has been reported in various models [29, 30]. Furthermore, the reduced forms of biopterin are critical for iNOS dimerization and NO production . Consistent with these reports, OmpA+ E. coli induced total expression, dimer formation of iNOS, and subsequent NO production 30–60 minutes after infection of HBMECs. Of note, NO-mediated Ecgp96 expression via iNOS was observed in HBMECs infected with OmpA+ E. coli. In DAHP-treated HBMECs, a significant decrease in total dimeric iNOS and in NO production was observed and, concurrently, reduced the total and surface expression of Ecgp96 during OmpA+ E. coli infection. A significant decrease in the total biopterin levels in HBMECs transfected with Ecgp96 siRNA and decreased expression of GCH1 induced by OmpA+ E. coli infection were also observed.
We translated our in vitro observations to a neonatal mouse model to further investigate the role of biopterin synthesis in OmpA+ E. coli infection. A beneficial role of DAHP in preventing septic shock induced by gram-positive and gram-negative pathogens was suggested by several studies [32–34]. However, this is the first report to show the effect of DAHP in preventing bacterial entry into brain. NO production by iNOS during infection has been reported to be 100–1000-fold higher than that of constitutive forms of NOS  and to cause severe brain injury and increased intracranial pressure . Additionally, elevated GCH1 expression and pterin metabolites were found to produce persistent neuropathic pain . OmpA+ E. coli–induced abnormalities in brain pathology completely resolved, and the total biopterin production was reduced in DAHP-treated newborn mice. In addition, the observation of a significant decrease in blood bacteremia among infected newborn mice (data not shown) indicates that the inhibition of biopterin synthesis reduced the bacterial load in the blood.
In sum, this study demonstrates that OmpA+ E. coli infection stimulates biopterin synthesis by manipulating the expression and activity of GCH1 for efficient brain translocation. Furthermore, the stimulation of biopterin synthesis also favors iNOS expression, dimerization, and subsequent NO production (Figure 7). Since iNOS-mediated NO production was triggered by many clinically important pathogens, it is of interest to examine the role of GCH1 and biopterin production in other infectious diseases.
Acknowledgments.We thank Barbara Driscoll, for a critical reading of the manuscript, and members of Dr Bellusci's laboratory, for assistance in immunohistochemistry. We are indebted to Larry Wang, Department of Pathology at CHLA, for evaluating the brain sections.
Financial support. This work was supported by the National Institutes of Health (grants AI40567 and NS73115 to N. V. P.).
Potential conflicts of interest. All authors: No reported conflicts.
All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.