Background
Death in the acute phase of the disease and neurological as well as neuropsychological sequelae are frequent complications of bacterial central nervous system (CNS) infections.
Escherichia coli is a Gram-negative bacillus causing local infections in the urinary tract, abdomen, and lungs. Systemic spread of these infections is frequent, leading to sepsis and meningoencephalitis, and is associated with high rates of mortality and morbidity in newborn infants, and in immunocompromised or elderly persons despite advances in antimicrobial chemotherapy [
1]. The presence of the capsule K1 confers invasiveness to the strains and enables them to penetrate the blood–brain barrier
in vivo [
2,
3]. Several studies in human beings and in the newborn rat model of hematogenous
E. coli meningoencephalitis suggest that a high degree of bacteremia is required for meningeal invasion [
4,
5]. The ability of bacteria to achieve high bacterial concentrations in blood, increasing the probability of invasion of the CNS, is greater in immunocompromised individuals (for example, neonates) than in immunocompetent adults, thus explaining the differences in the occurrence of
E. coli K1 meningoencephalitis [
6-
8]. Invasion of brain microvascular endothelial cells appears to be a prerequisite for
E. coli K1 to induce meningoencephalitis [
8]. Some
E. coli K1 structures, such as outer membrane protein A (OmpA), Ibe proteins, and cytotoxic necrotizing factor 1, are necessary for successful bacterial traversal across the blood–brain barrier [
8]. In recent years, a significant increase in multidrug-resistant
E. coli strains has been observed [
9]. In Europe, antimicrobial resistance in Gram-negative bacteria is spreading, particularly in
E. coli, which constitutes a large portion of invasive Gram-negative isolates in European countries [
10]. Antibiotics are essential for the control and treatment of
E. coli infections in human beings and animals. However, it is generally accepted that antimicrobial resistance is associated with the quantity of antibiotic consumption [
11]. Vaccination may be an important primary prevention strategy in human beings against most
E. coli strains. To date, no effective vaccine is available for the prevention of these infections [
10]. Therefore, the development of novel therapeutic strategies for these infectious diseases is of high priority. Vitamin D has long been known to play a role in building the skeletal system and in calcium homeostasis; vitamin D deficiency is known to be a cause of rickets and osteomalacia and aggravates osteoporosis [
12]. In addition to this well-known role in mineral and skeletal homeostasis, 1,25-dihydroxyvitamin D3 (1,25(OH)
2D3) affects both innate and adaptive immune responses [
13,
14]. Cells of the immune system possess vitamin D receptors and are capable of metabolizing the active form of vitamin D (1,25(OH)
2D3) [
15], suggesting vitamin D as an important factor in the immune response to infection [
16]. Studies of the innate immune response to pathogens such as
Mycobacterium tuberculosis have shown that pathogen-recognition receptor (PRR)-mediated activation of localized vitamin D metabolism and signaling is a key event associated with resistance to infection [
17]. Epidemiological studies have established that vitamin D deficiency plays an important role in susceptibility to tuberculosis [
18]. Vitamin D supplementation showed a beneficial modulating effect on sepsis [
19] and on endotoxin shock in mice [
20]. We have previously shown in microglia cultures that vitamin D3 deficiency may impair the resistance of the brain against bacterial infections [
21]. Taken together, these data indicate an important role of vitamin D3 in the clearance of infections and containment of inflammation by the body’s immune cells. Here we extended our analysis to the action of vitamin D3
in vivo. The aims of this study were (i) to investigate the immunomodulatory capacity of vitamin D and (ii) to examine the impact of vitamin D3 supplementation as a preventive or adjuvant therapeutic intervention on the course and mortality of experimental
E. coli meningoencephalitis.
Methods
Vitamin D3 concentration in the blood
To produce different vitamin D3 concentrations in the circulation, mice were fed with a diet containing either low (L-VitD; vitamin D3 concentration below the detection level; normal calcium and phosphate concentrations), standard (S-VitD; vitamin D3 concentration 1,500 IU/kg food; normal calcium and phosphate concentrations), or high (H-VitD; vitamin D3 concentration 75,000 IU/kg food; normal calcium and phosphate concentrations) vitamin D3 concentrations (all from ssniff Spezialdiäten GmbH, Soest, Germany). After 6 weeks, the 25-hydroxyvitamin D3 serum concentrations were measured in the mice using liquid chromatography-tandem mass spectrometry. Serum samples were obtained by puncture of the retroorbital plexus and measured using a MassChrom® 25-OH-Vitamin D3/D2 LC-MS/MS kit (Chromsystems, Munich, Germany). Measurements were made using an AB Sciex API 4000 LC/MS/MS system (AB Sciex, Darmstadt, Germany). The HPLC component was from Shimadzu (Duisburg, Germany). The assay was adapted to a 50 μl sample volume.
Animals
The animal experiments were approved by the Animal Care Committee of the University Hospital of Göttingen and by the Niedersächsische Landesamt für Verbraucherschutz und Lebensmittelsicherheit, Braunschweig, Lower Saxony, Germany. C57Bl/6 wild-type mice (2 to 3 months old, weight 20 to 30 g, Charles River Laboratory) were used in all experiments [
10]. Water and food were available
ad libitum.
Bacteria
The E. coli strain K1 (serotype O18:K1:H7), originally isolated from the cerebrospinal fluid of a child with neonatal meningoencephalitis (and the gift of Dr. Gregor Zysk, Institute of Medical Microbiology, Düsseldorf, Germany) was used in all experimental infections. Bacteria were grown overnight on blood agar plates, harvested in 0.9% saline and stored at −80 °C. Frozen aliquots were thawed immediately before the experiments and diluted with saline to the required bacterial concentration.
Induction of meningoencephalitis
For survival experiments, meningoencephalitis was induced by the slow injection of 4,000 colony-forming units (CFUs) of
E. coli K1 in 10 μl sterile saline (0.9% NaCl) into the right frontal lobe of the cerebral cortex using a 27-gauge disposable needle [
22] under intraperitoneal anesthesia with ketamine (100 mg/kg of body weight) and xylazine (10 mg/kg of body weight). All animals resumed their normal behavior after awaking from anesthesia. During the acute disease phase, animals were weighed and scored every 12 h (0, no apparent behavioral abnormality; 1, moderate lethargy; 2, severe lethargy; 3, unable to walk; 4, dead) [
19]. Mice with a clinical score of 3 were killed for ethical reasons. In survival experiments, animals were monitored for 14 days after infection. In bacteriological studies, using a 27-gauge needle, 10 μl of a suspension containing 9,000 CFUs of
E. coli K1 or an equal amount of saline were slowly injected into the right frontal lobe of the cerebral cortex. Mice were killed 20 h after infection.
Sample processing
Mice that were killed 20 h after infection were anesthetized with ketamine (100 mg/kg of body weight) and xylazine (10 mg/kg of body weight). Blood was drawn by cardiac puncture and 10 μl was used for the determination of bacterial concentrations. The remaining blood was stored at 4°C for 30 min and then centrifuged at 3,000 g for 10 min at 4°C. Serum was then transferred to another tube and stored at −20°C. The whole brain and spleen were removed. The cerebellum was dissected from the brain stem. Half of the spleen and the whole cerebellum were homogenized in 0.9% saline. Bacterial titers in homogenates and blood were determined by plating serial 10-fold dilutions in 0.9% saline on sheep blood agar plates (detection limit: 100 CFU/ml, respectively). The whole cerebrum and the other half of the spleen were fixed in 4% paraformaldehyde and then embedded in paraffin. In all experiments, a control group of five mice per group was injected with 0.9% NaCl.
Histological analysis
Paraffin-embedded, 2 μm coronal brain sections from killed or dead mice from the survival experiments as well as from animals killed 20 h after infection in the bacteriological studies were analyzed. Chloroacetate esterase staining was performed to evaluate the degree of inflammation in three superficial meningeal regions and the hippocampal fissure. This stain is used to detect neutrophils but it can also stain some monocytes or macrophages [
23]. The numbers of chloroacetate-esterase-stained leukocytes were counted in one high-power field (×40 objective) per region by a blinded investigator. For each animal, the leukocyte numbers of the individual fields were added and then divided by the number of counted regions.
Flow cytometry
Anesthetized animals were perfused transcardially with PBS 20 h after infection, and the whole brain was removed and processed. Brain was digested and homogenized with collagenase D (2.5 mg/ml, Roche Diagnostics GmbH, Mannheim, Germany) and DNase I (2 mg/ml, Roche Diagnostics GmbH) using a gentleMACS dissociator (Miltenyi Biotec, Germany). The resultant homogenates were mechanically dissociated and passed through a 70-μm nylon cell strainer (BD Biosciences, Franklin Lakes, NJ, USA). Leukocytes were separated in a 37/70% Percoll gradient (GE Healthcare, Chalfont St Giles, Buckinghamshire, UK). Single cells were stained with the following antibodies: CD45 (30-F11), CD4 (RM4-5), CD27 (LG.3A10), CD11b (M1/70), and Ly6C (HK1.4) purchased from BioLegend (San Diego, CA, USA), CD3 (145-2C11), CD25 (PC61.5), CD19 (eBio1D3) NK1.1 (PK136), and FoxP3 (FJK-16 s) provided by eBioscience (San Diego, CA, USA), and Ly6G (1A8, BD Pharmigen, Franklin Lakes, NJ, USA) and CCR2 (FAB5538A, R&D Systems, Minneapolis, MN, USA). At least 50,000 events were acquired on a FACSCanto II cell analyzer (BD Biosciences) and analyzed using FlowJo software (version 8.8; Tree Star).
Cytokine and chemokine measurement
Cytokines and chemokines were measured in cerebellum homogenates of mice killed 20 h after infection and in the cerebellum and spleen of dead mice from survival experiments as well as in the cerebellum and spleen homogenates of five mice per group injected with 0.9% sterile saline. Concentrations of IL-6, IL-10, KC (CXCL1), IFN-γ, and MIP-2 (CXCL2) were determined using DuoSet ELISA development kits (R&D Systems, Wiesbaden, Germany). Procedures were performed according to the manufacturer’s instructions. The sensitivity of the assays for these cytokines and chemokines was 7.5 pg/ml.
Statistics
Vitamin D3 serum concentrations, the weight of animals, bacterial loads in cerebellum and spleen, and the mean number of leukocytes per area are shown as mean and standard deviation (SD) and were analyzed by one-way-ANOVA and corrected for repeated testing with the Bonferroni multiple comparisons test. The cytokine and chemokine concentrations are reported as medians and corresponding interquartile ranges (Q25 and Q75) and were analyzed by the Kruskal-Wallis test and corrected for repeated testing using the Bonferroni method. For survival analysis, the log-rank test based on a Kaplan-Meier plot was used. The clinical scores were reported as box and whiskers (min to max) and were analyzed using the Kruskal-Wallis test. P values were corrected for repeated testing with the Bonferroni method. For all analyses, GraphPad Prism version 5 (GraphPad Software, San Diego, CA) was used; P ≤ 0.05 was considered statistically significant.
Discussion
Here we provide experimental evidence that in bacterial CNS infection an adequate supply of vitamin D3 decreases susceptibility of the brain to
E. coli infection and reduces mortality. Moreover, vitamin D has anti-inflammatory properties, illustrated in this study by increased anti-inflammatory IL-10 and decreased proinflammatory IL-6 concentrations in brain tissue. Clinical and experimental observations support the hypothesis that vitamin D has an effect on the immune response to infection and the ability to eliminate pathogens after entry into the host. Vitamin D deficiency may therefore be an underlying cause of infectious diseases and immune disorders [
14,
16,
33]. A central feature of many of these non-classical actions of vitamin D is related to the synthesis of active 1,25(OH)
2D in a cell-specific manner: enzyme 25OHD-1α-hydroxylase (encoded by the gene
CYP27B1) is expressed by many extrarenal tissues, including the immune system [
14,
34]. For example, activated human T- and B-cells can convert the inactive intermediate of vitamin D (also known as 25-hydroxyvitamin D or 25(OH)D) to active 1,25(OH)
2D
in vitro [
35], and this locally produced 1,25(OH)
2D acts on immune cells in an autocrine or paracrine fashion [
36]. The concentration of the biologically active form of vitamin D, 1,25(OH)
2D, is dependent on the serum vitamin D (25(OH)D) concentration, the substrate for CYP27B1 [
13,
37].
To assess the role of vitamin D in an
in-vivo experimental model of bacterial CNS infection, we analyzed the influence of the infection on the survival and immune response of mice fed a diet containing different amounts of vitamin D3 after injection of
E. coli directly into the right frontal lobe. The mice did not receive antibiotic treatment; that is, this model tests the ability of mice to combat pathogens after they have entered the brain and cerebrospinal fluid. The importance of vitamin D for the resistance to infections has long been appreciated but poorly understood. This has been especially true for tuberculosis. Indeed, prior to the development of specific drugs for the treatment of tuberculosis, getting out of the city into sunlight (and fresh air) was the treatment of choice [
16]. In a survey of patients with tuberculosis in London [
38], 56% had undetectable 25(OH)D levels, and an additional 20% had detectable levels below 9 ng/ml (22 nM). A randomized double-blind intervention study showed that supplementation with vitamin D may reduce disease burden in patients with frequent respiratory tract infections [
33].
In-vivo studies in mice and rats also showed beneficial effects of vitamin D in lipopolysaccharide-induced sepsis [
20,
39], and antenatal vitamin D therapy improved survival in newborn rat pups and enhanced their lung structure after exposure to endotoxin [
40]. The influence of vitamin D on the course of CNS infections is not known. We therefore asked whether vitamin D might modulate the course of and survival from infection after bacterial meningoencephalitis and whether vitamin D could potentially be used therapeutically to alleviate CNS pathology. We observed that vitamin D3-deficient mice died earlier and more frequently than mice fed a diet containing standard or high amounts of vitamin D3. The bacterial burdens in mice dying in the acute phase of infection and mice killed at 20 h were similar in all three groups, that is, the antibacterial action of vitamin D3 in our experimental design was less apparent. The high number of vitamin D3-deficient animals with persistent bacteria in the brain compared with animals with an adequate or high vitamin D3 supply, however, indicated that vitamin D3 also supported pathogen elimination in our model. In an early study, Rook and colleagues [
17] demonstrated that active vitamin D, (1,25(OH)
2D), can inhibit the growth of
Mycobacterium tuberculosis in vitro. At this time, the physiological significance of this finding was unclear. To clarify the activity of vitamin D against
M. tuberculosis, Liu and colleagues [
41] observed that activation of the toll-like receptor TLR1/2 by a lipoprotein extracted from
M. tuberculosis reduced the viability of intracellular
M. tuberculosis in human monocytes and macrophages, concomitant with an increased expression of the vitamin D receptors and of CYP27B1 (the enzyme that produces 1,25(OH)
2D) in these cells. Killing of
M. tuberculosis occurred only when the serum in which the cells were cultured contained adequate levels of 25(OH)D, the substrate for CYP27B1. Activated vitamin D (1,25(OH)
2D) bound to the monocyte vitamin D receptors and then was able to act as a transcription factor leading to the induction of cathelicidin, a potent antimicrobial protein [
42], and the promotion of phagocytosis and intracellular killing [
41,
43,
44]. We were also able to show that, in cultured microglia cells, vitamin D3 deficiency led to a decreased phagocytosis and intracellular killing rate of
E. coli [
21]. One reason for the mild antimicrobial properties of vitamin D3 in the present
in-vivo model might be that, unlike the human
cathelicidin gene, the mouse gene does not contain a vitamin D response element, and is not induced directly by 1,25(OH)
2D [
16]. Presumably, in addition to cathelicidin, other immune mediators are involved in the vitamin D-dependent immune pathways.
In human beings with bacterial meningoencephalitis, and in animal models of this disease, leukocytes, predominantly myelomonocytic cells such as monocytes, macrophages, and neutrophil granulocytes, quickly enter the subarachnoid space in response to local production of cytokines, chemokines, and other chemotactic stimuli [
22]. In terms of
S. pneumoniae and
E. coli, we previously showed,
in vivo, that granulocytes are predominantly involved in the restriction of the multiplication of extracellular bacteria [
24,
30]. Different vitamin D3 concentrations in the present study did not significantly influence the meningeal invasion with myeloid cells during
E. coli meningoencephalitis; the mean number of chloroacetate-esterase-stained leukocytes or areas and the recruitment of myeloid cells into the inflamed CNS 20 hours after
E. coli infection, as assessed by flow cytometry, were comparable in all three groups. A study on the immune response to an allergic stimulus in mice
in vivo suggested that vitamin D supplementation (100 ng 1,25(OH)
2D injection) given after the initial period of sensitization prevented high levels of eosinophils associated with a reduced local inflammatory response in bronchoalveolar lavage fluid and lung tissue. Constant vitamin D supplementation (100 ng 1,25(OH)
2D injection every other day during the whole study period), however, did not reduce the entry of eosinophils into the respiratory epithelia [
45]; this is in accordance with our observation.
One mechanism of how vitamin D3 might act beneficially in infections, could be an increase in pathogen phagocytosis with adequate levels of vitamin D [
21] and the ability of calcitriol to maintain antimicrobial peptide gene expression [
13,
14]. Other mechanisms could be the regulation of anti- and proinflammatory compounds by vitamin D. For this purpose, we investigated the pattern of cytokine and chemokine production, a central feature in the development of neuroinflammation, neurodegeneration, and demyelination in the CNS [
46,
47]. In bacterial meningoencephalitis, after the pathogen crossed the blood–brain barrier, microglia can respond directly to intact bacteria or to bacterial cell wall compounds and produce a wide array of inflammatory mediators, including TNF-α, IL-6, IL-12, keratinocyte-derived chemokine (CXCL1/KC), CCL2/MCP-1, CXCL2/MIP-2, and CCL5/RANTES [
48,
49]. Neuronal damage in bacterial meningoencephalitis is caused by the dual effects of an overwhelming inflammatory response and the direct action of bacterial toxins [
50,
51].
In this study,
E. coli led to the production of numerous cyto- and chemokines in the brain and spleen, including IL-6, IL-10, CXCL1 and CXCL2 shortly after bacterial exposure. IL-6 in the cerebellum, CXCL1 in the cerebellum and spleen and CXCL2 in the spleen were clearly down-regulated in high vitamin D3-fed animals, whereas the bacterial concentrations were not influenced. Our data imply that the protective role of vitamin D3 is related to the reduced release of proinflammatory IL-6 and other chemokines, such as CXCL1 and CXCL2. Moreover, anti-inflammatory IL-10 release [
52] was significantly increased in mice dying in the acute phase and fed a diet containing a high vitamin D3 concentration compared with mice dying in the acute phase and fed a diet containing low or standard vitamin D concentrations. This is in accordance with previously published data that vitamin D can increase IL-10 production [
16,
53]. In murine
S. pneumoniae meningitis, the absence of IL-10 was associated with higher proinflammatory cytokine and chemokine concentrations and more pronounced infiltrates [
54]. IL-10 reduced sepsis-associated hippocampal neuronal damage as a result of pneumococcal sepsis in mice overexpressing IL-10 [
55]. Also, intravenously administered recombinant IL-10 reduced the level of cerebrospinal fluid pleocytosis, cerebral edema, and intracranial pressure in a rat model of pneumococcal meningoencephalitis [
56]. In our study the IFN-γ levels of animals dying in the acute phase in all three groups were below the level of quantification. In early infection (20 h), low, but approximately equal IFN-γ levels were measured, suggesting that IFN-γ was not involved in the protective action of an adequate vitamin D supply.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
MD performed the experiments and wrote the manuscript. NS, MM, AM, and JT performed the experiments. SN analyzed the FACs data and discussed the manuscript. UKH and LCB reviewed and discussed the manuscript. TB, CS, and RN planned and designed the study. All authors read and approved the final version of the manuscript.