OUP user menu

Increased Expression of BDNF and Proliferation of Dentate Granule Cells After Bacterial Meningitis

Simone C. Tauber MD, Christine Stadelmann MD, Annette Spreer MD, Wolfgang Brück MD, Roland Nau MD, Joachim Gerber MD
DOI: http://dx.doi.org/10.1097/01.jnen.0000178853.21799.88 806-815 First published online: 1 September 2005

Abstract

Proliferation and differentiation of neural progenitor cells is increased after bacterial meningitis. To identify endogenous factors involved in neurogenesis, expression of brain-derived neurotrophic factor (BDNF), TrkB, nerve growth factor (NGF), and glial cell line-derived neurotrophic factor (GDNF) was investigated. C57BL/6 mice were infected by intracerebral injection of Streptococcus pneumoniae. Mice were killed 30 hours later or treated with ceftriaxone and killed 4 days after infection. Hippocampal BDNF mRNA levels were increased 2.4-fold 4 days after infection (p = 0.026). Similarly, BDNF protein levels in the hippocampal formation were higher in infected mice than in control animals (p = 0.0003). This was accompanied by an elevated proliferation of dentate granule cells (p = 0.0002). BDNF protein was located predominantly in the hippocampal CA3/4 area and the hilus of the dentate gyrus. The density of dentate granule cells expressing the BDNF receptor TrkB as well as mRNA levels of TrkB in the hippocampal formation were increased 4 days after infection (p = 0.027 and 0.0048, respectively). Conversely, NGF mRNA levels at 30 hours after infection were reduced by approximately 50% (p = 0.004). No significant changes in GDNF expression were observed. In conclusion, increased synthesis of BDNF and TrkB suggests a contribution of this neurotrophic factor to neurogenesis after bacterial meningitis.

Key Words
  • Bacterial meningitis
  • Brain-derived neurotrophic factor (BDNF)
  • Hippocampus
  • Neurogenesis
  • Neurotrophin
  • Streptococcus pneumoniae
  • TrkB

Introduction

Survivors of bacterial meningitis frequently experience long-term sequelae, including learning and memory deficits (1-4). Neuronal injury is caused by migrating leukocytes releasing cytokines, chemokines, free radicals and proteolytic enzymes, stimulation of resident immune cells, and direct toxicity of bacterial compounds (5). Morphologically, in various models of experimental meningitis and in human autopsy cases, neuronal injury occurs most frequently in the hippocampal formation (6-9). Neurogenesis in the hippocampal dentate gyrus is increased in response to several modes of brain damage, e.g. ischemia (10-12), trauma (13, 14), and epileptic seizures (15). Similarly, proliferation and differentiation of neural progenitor cells in the dentate gyrus is increased after experimental meningitis, probably reflecting an endogenous response to brain injury and cell loss (16).

Neurotrophic factors may act as regulators of neurogenesis. Important members of the neurotrophin family are brain-derived neurotrophic factor (BDNF) and nerve growth factor (NGF). These structurally related proteins promote differentiation and survival of neurons (17, 18). Administration of NGF stimulated proliferation of brain precursor cells (19). BDNF plays a role during early differentiation of neurons and is expressed in the adult brain (20). In the mature central nervous system, BDNF is most abundant in the hippocampal formation and hypothalamus (21, 22). Intracerebroventricular administration of BDNF increased neurogenesis in rats (23). The expression of BDNF and NGF is influenced by neuronal activity such as afferent stimulation (24), stress (25), light exposure (26), and epileptic seizures (27). Hippocampal BDNF expression is increased after training in spatial memory tasks (28, 29) and is associated with phosphorylation of the BDNF receptor tyrosine kinase receptor B (TrkB) (30), which is expressed at high levels in the hippocampal formation (31). BDNF and TrkB mRNA were elevated in hippocampal cells after chronic administration of antidepressant drugs (32, 33). Mice lacking BDNF have learning deficits (34) and impaired long-term potentiation (35).

Glial cell line-derived neurotrophic factor (GDNF), which is not structurally related to the neurotrophin family, belongs to the transforming growth factor β (TGF-β) superfamily and signals through the GDNF family receptors (GFRα) (36). Neuroprotective effects of GDNF have been observed in injured dopaminergic midbrain neurons (37). Moreover, GDNF has been shown to diminish damage caused by kainic acid-induced tonic-clonic convulsions (38), to promote axonal growth after spinal cord transections (39), and to attenuate ischemic brain damage (40).

Neurotrophic factors are essential for cellular growth, differentiation, and survival. To identify endogenous factors regulating neurogenesis after experimental meningitis, expression of BDNF and its receptor TrkB, NGF, and GDNF was investigated.

Materials and Methods

Mouse Model of Experimental Meningitis

Male C57BL/6 mice (weight 20-26 g, age 2-3 months) were anesthetized with 100 mg/kg body weight ketamine and 10 mg/kg xylazine intraperitoneally before injection of 10 μL of 0.9% NaCl containing 104 colony-forming units (CFU) of a Streptococcus pneumoniae type 3 strain into the right forebrain. Control groups received the same treatment with injection of 10 μL of 0.9% NaCl without bacteria. Animal experiments were approved by the Animal Care Committee of the University Hospital of Göttingen and by the District Government of Braunschweig, Lower Saxony. Bacterial titers in this mouse model of meningitis were investigated in untreated and in ceftriaxone-treated animals: in untreated animals, bacterial counts in cerebellar homogenates rose from 4.1 ± 1.3 log CFU/mL at 12 hours after infection to 5.9 ± 0.3 (18 hours), 6.9 ± 0.5 (24 hours), and 8.9 ± 0.5 log CFU/mL at 36 hours after infection (41). Mice treated with ceftriaxone had cerebellar bacterial titer of 5.3 ± 2 and 4.3 ± 1.8 log CFU/mL 1 and 2 days after initiation of treatment, respectively. At later time points, no bacteria were detectable (detection limit of 102 CFU/mL) (42).

All infected mice developed clinical signs of disease within 24 hours. One set of animals (infected and saline-treated control group, n = 19 each) was killed by decapitation 30 hours after infection without any specific antibiotic treatment. A second set of mice (infected and saline-treated control group, n = 29 each) received antibiotic treatment twice daily with subcutaneous ceftriaxone (100 mg/kg body weight; Hoffmann-LaRoche, Grenzach-Wyhlen, Germany). Therapy was initiated 24 hours after infection and mice were killed on day 4 after infection. Five animals died between 30 hours and 52 hours after infection and were excluded from further analysis. Mice were weighed twice daily and the tightrope test (43) was performed every 12 hours.

Brains for mRNA analysis (n = 11 each respective group) and Western blot at day 4 after infection (n = 10 each group) were removed and the hippocampal formation was separated from the neocortex. The tissue was immediately transferred into liquid nitrogen and stored at -80°C. For immunohistochemical analysis, 8 animals of each respective group received 5 intraperitoneal injections of 50 mg/kg bromodeoxyuridine (BrdU) (Sigma-Aldrich, St. Louis, MO) every 3 hours from 27 to 15 hours before death. These mice were deeply anesthetized and perfused with 4% formalin.

To exclude an effect of ceftriaxone on cellular proliferation in the hippocampal formation, additional experiments were performed: after an injection of 10 μL of sterile saline into the right forebrain, mice were treated with ceftriaxone (100 mg/kg body weight twice daily) or received the same volume of sterile saline subcutaneously (n = 8 each group). BrdU was given as described previously (5 intraperitoneal injections of 50 mg/kg BrdU every 3 hours), and animals were killed at day 4 after the saline injection.

Motor Performance

The motor performance of mice was evaluated before and every 12 hours after infection by the tightrope test (43). Mice were placed in the middle of a tightrope and the time for reaching one end of the rope was measured. Mice reaching one end in less than 6 seconds were scored 0; one additional point was given for every additional 6 seconds needed. Mice hanging for more than 60 seconds on the rope without reaching one end were scored 10. Mice falling from the rope before 60 seconds received additional points and 11 points for every 6 seconds earlier than 60 seconds (range of the score: 0-20).

Neuronal Damage and Meningeal Inflammation

Formalin-fixed, paraffin-embedded, 1-μm-thick brain sections were stained with hematoxylin and eosin (H&E). Neuronal damage and meningeal inflammation were evaluated by a semiquantitative score in infected and uninfected control animals at 30 hours and 4 days after inoculation (41). Neuronal damage was investigated in 4 areas of the brain (neocortex, striatum, hippocampus, dentate gyrus), and the density of necrotic or apoptotic neurons was assessed as absent or minimal (< 10% of all neurons: 0), moderate (10-30% of all neurons: 1), or numerous (> 30% of all neurons: 2). Apoptosis was defined morphologically by cell shrinkage, homogenous chromatin condensation, nuclear shrinkage, and disintegration of the nucleus into apoptotic bodies. Necrosis was characterized by initial cell swelling, eosinophilic degeneration of the cytoplasm, and shrinkage of the nucleus with clumping of chromatin. For each animal, the scores of the individual areas were added (score range: 0-8).

Meningeal inflammation was estimated by the invasion of granulocytes into the frontal interhemispheric region, the hippocampal fissure (both sides), 3 superficial meningeal regions over the convexities, and the third ventricle. One high-power field (diameter: 250 μm) was scored in each region: no granulocytes: 0, < 10 granulocytes: 1, 10 to 50 granulocytes: 2, and > 50 granulocytes: 3. The scores of the individual regions were added (range of the score: 0-21).

RNA Extraction and cDNA Synthesis

Total amount of RNA was extracted using Trizol reagent (Life Technologies, Gaithersburg, MD) according to the manufacturer's instructions. Integrity was confirmed by agarose gel electrophoresis. Two micrograms of total RNA from each sample was used for cDNA synthesis. Reverse transcription was performed using Omniscript RT Kit (Qiagen, Hilden, Germany). The reaction was carried out for 60 minutes at 37°C and stopped by heating (93°C for 5 minutes). Twenty-microliter reaction samples were diluted to a total volume of 100 μL, and 5 μL of diluted cDNA was used for real-time polymerase chain reaction polymerase chain reaction (PCR).

Real-Time Polymerase Chain Reaction Analysis

Expression levels of all genes evaluated were determined by quantitative real-time PCR in the BioRad iCycler iQ™ system (BioRad, Hercules, CA) using Qiagen Quantitect™ SYBR Green (Qiagen) and FITC (BioRad). cDNA was amplified using the following gene-specific primers. BDNF: for 5′ GTC-TGA-CGA-CGA-CAT-CAC-TGG 3′, rev 5′ AGA-GGA-GGC-TCC-AAA-GGC-ACT 3′; TrkB: for 5′ CGC-CCT-GTG-AGC-TGA-ACT-CTG 3′, rev 5′ CTG-CTT-CTC-AGC-TGC-CTG-ACC 3′; GDNF: for 5′ GGG-CCT-GAG-GTC-TAT-TAC-ATC 3′, rev 5′ GTT-TCT-GAG-GGC-ACG-AAG-GAG 3′; NGF (44): for 5′ CAG-ACC-CGG-AAC-ATC-ACT-GTA 3′, rev 5′ CCA-TGG-GCC-TGG-AAG-TCT-AG 3′; GAPDH: for 5′ CTG-GTC-ACC-AGG-GCT-GCC-ATT-TGC 3′, rev 5′ CCC-ATT-CTC-GGC-CTT-GAC-TGT-GCC 3′. By conventional reverse transcriptase-PCR, it was confirmed that all primer pairs produced only one amplification product ranging in size from 139 to 160 bp. For quantification, gene-specific standard serial dilutions were coamplified. Quantitative real-time PCR was performed under the following conditions: 94°C for 15 minutes, followed by 60 cycles of 94°C for 30 seconds and 60°C (BDNF, GDNF, and GAPDH) or 63°C (TrkB) or 67°C (NGF) for 30 seconds, and then 72°C for 45 seconds. At the end of each cycle, fluorescence was measured at 55°C. Specificity of individual real-time PCR products was assessed by melting curve analysis. All sets of reactions were measured as duplicates and each measurement included a nontemplate control. All gene-specific mRNA expression values were normalized against the internal housekeeping gene GAPDH.

Western Blot

For protein extraction, lysis buffer (50 nM Tris pH 8.0, 150 nM NaCl, 1% Triton X-100) containing proteinase inhibitors EDTA (0.2 mM), PMSF (2 mM), pepstatin (1 μM), leupeptin (10 μM), and aprotinin (2 μg/mL) was added to the isolated hippocampi. After incubation of tissue homogenates on ice for 20 minutes, the samples were centrifuged for 20 minutes at 13,000 × g at 4°C. The total amount of protein in the supernatant was determined by BCA-Assay (Pierce, Rockford, IL) according to the manufacturer's instructions. Equal amounts of hippocampal protein (40 μg per lane) were separated in SDS-PAGE according to Schagger and von Jagow (45) and transferred to nitrocellulose membrane according to standard methods. Immunodetection was performed using rabbit polyclonal anti-BDNF (1:400; Santa Cruz Biotechnology, Santa Cruz, CA) and peroxidase conjugated mouse antirabbit IgG (1:10000; Jackson Immunoresearch, West Grove, PA). The reaction was developed with ECL+ enhanced chemiluminescence Western blotting detection reagents, and the emitted light was detected and quantified using the chemiluminescence camera Fluor S MAX Multiimager (BioRad) and the software Quantity One 4.2.1 (BioRad). For each sample, emitted light counts per mm2 were detected.

Immunohistochemistry

Brain sections from 8 mice of each group (infected and uninfected mice 30 hours and 4 days after inoculation) were examined for BrdU incorporation, BDNF, and TrkB protein. For visualization of colocalization of 2 markers in one cell, double immunofluorescence labeling was performed: coexpression of TrkB with the neuron-specific nuclear antigen (NeuN), neuronal calbindin, and with the astrocytic glial fibrillary acidic protein (GFAP) was evaluated. Moreover, double labeling of bromodeoxyuridine (BrdU) with the microglia/macrophage marker “ionized calcium-binding adaptor molecule 1” (Iba1), and B-cell (CD45R/B220) and T-cell markers (CD3) was performed.

Furthermore, paraformaldehyde-fixed coronal brain sections from uninfected mice that had received antibiotic treatment with ceftriaxone (n = 8) and control animals (n = 8) were investigated for BrdU incorporation. Briefly, after deparaffinization, sections were pretreated with microwaving for 3 × 5 minutes in citric acid buffer, 10 mmol/L, pH 6.0. After blocking with 10% FCS/PBS, primary antibodies were applied at the concentrations indicated and permitted to bind overnight at 4°C (BrdU, BDNF, TrkB, and Iba1) or for 90 minutes at room temperature (NeuN, calbindin, GFAP, CD45R/B220, and CD3).

BrdU incorporation was detected by binding of 1:15 diluted peroxidase-conjugated monoclonal mouse-anti-BrdU antibodies (Roche, Mannheim, Germany) and visualized by 3,4-diaminobenzidine (DAB, Roche, Mannheim, Germany). BDNF and TrkB were detected by binding of rabbit polyclonal anti-BDNF (1:300, Santa Cruz Biotechnology) and rabbit polyclonal anti-TrkB (1:300, Santa Cruz Biotechnology) antibodies. The anti-TrkB antibody recognizes full-length TrkB exclusively. For light microscopy, sections were incubated with appropriate biotinylated secondary antibodies (Amersham, Buckinghamshire, U.K.) followed by treatment with avidin-peroxidase (Sigma-Aldrich). DAB was used as the chromogenic substrate.

NeuN, calbindin, and GFAP were visualized using monoclonal mouse anti-NeuN (1:100, Chemicon, Temecula, CA), monoclonal mouse anticalbindin-D-28k (1:150, Sigma-Aldrich), and monoclonal mouse anti-GFAP antibodies (1:50, DakoCytomation, Carpinteria, CA). CD45R/B220 and CD3 were visualized by using rat antimouse CD45R/B220 (1:200; PharMingen, BD Biosciences, Heidelberg, Germany), rat antihuman CD3 (1:200; Serotec, Düsseldorf, Germany), and Iba1 (Serotec). For fluorescence double labeling, sections were incubated for 1 hour with 1:200 Cy-2-streptavidin (Jackson Immunoresearch), transferred to 10% FCS for 30 minutes, and then incubated overnight with rabbit polyclonal anti-TrkB (1:300; Santa Cruz Biotechnology) at 4°C or monoclonal mouse anti-BrdU (1:16; Roche, Mannheim, Germany) for 90 minutes at room temperature. Finally, slices were incubated with 1:400 (for TrkB) or 1:200 (for BrdU) goat-antirabbit Cy3 (Jackson Immunoresearch) for 1 hour. Control sections were incubated with isotype control antibodies, rabbit serum, or without primary antibody.

Quantification of TrkB- and BrdU-Labeled Cells

An Analysis Software Imaging System (BX51; Olympus, Hamburg, Germany; software AnalySIS® 3.2; Soft Imaging System GmbH, Münster, Germany) was used to count immunolabeled cells and to measure the area of the granular cell layer and hilus of the dentate gyrus. The density of immunolabeled cells was expressed as the number of marked cells per mm2 of the area measured.

Statistical Analysis

Values are expressed as medians (25/75 quartile). For statistical comparison, the 2-tailed nonparametric Mann-Whitney U test was used and p < 0.05 was considered statistically significant.

Results

Motor Performance in Experimental Meningitis

The tightrope test showed a distinct impairment of motor function in mice with bacterial meningitis. Comparison of scores from uninfected versus infected animals revealed a significantly higher score (i.e. a poorer motor performance) in infected animals between 24 and 60 hours with a maximum score at 36 hours after infection. Thereafter, the motor function recovered without reaching the level of uninfected animals at the end of the experiment 4 days after infection (Fig. 1).

FIGURE 1.

Time course of the tightrope test score in uninfected (n = 19 ▪) and infected (n = 19 •) mice revealing a reduced motor function of animals with bacterial meningitis with maximally impaired motor performance 36 hours after infection and recovery thereafter without reaching the levels of uninfected controls at 4 days after infection (median and 25175 quartile; **, p < 0.01; ***, p < 0.001).

Neuronal Damage and Meningeal Inflammation

The density of apoptotic and necrotic cells was higher in mice with bacterial meningitis compared with uninfected control animals at 30 hours after inoculation (neuronal damage score 2 [1-6] vs. 0 [0-1], p = 0.0043), whereas no significant difference was detectable 4 days after infection (0.5 [0-2] vs. 0 [0-1], p = 0.19).

Meningeal inflammation was determined by a semiquantitative score: a strong granulocytic invasion of the meninges was observed in all infected mice 30 hours after inoculation, whereas infiltration of granulocytes was absent in control animals (12 [11-15] vs. 0.0, p < 0.01). Four days after infection and 3 days after initiation of treatment, the number of granulocytes in the meninges of infected mice had distinctly decreased (2 [0-4] vs. 0.0, p < 0.01).

Microglia/Macrophages in Meninges and the Hippocampal Formation

In the acute phase of bacterial meningitis, immunofluorescence staining of Iba1 (a calcium-binding protein specifically expressed in brain microglia and macrophages) showed the presence of microglia/macrophages in the meninges as well as in the hippocampal formation. In uninfected control animals, no Iba1-positive cells were detected (Fig. 2). The density of microglia/macrophages had decreased 4 days after infection. Immunofluorescence studies for CD45R/B220 as well as CD3 revealed no B- or T-cells in the hippocampal formation and the meninges during the acute phase of infection. However, 4 days after infection, single B-cells were detected in the meninges, whereas in the hippocampal formation, no B- or T-cells were observed (data not shown).

FIGURE 2.

Double-label fluorescent immunohistochemistry of Iba1 (green), a marker for microglia/macrophages, showing a strong infiltration of meninges 30 hours after intracerebral infection with Streptococcus pneumoniae(A). In contrast, no reactivity for Iba1 was observed in uninfected control animals (B). Iba1-positive cells in the hilus of the hippocampus of infected (C) and uninfected (D) mice 30 hours after inoculation demonstrating activated microglia/macrophages in the hippocampal formation of infected mice. Scale bar = 50 μm.

Hippocampal and Neocortical Gene Expression of BDNF, NGF, and GDNF in Bacterial Meningitis

In median, BDNF mRNA levels in the hippocampal formation of infected mice were increased 1.7-fold 30 hours after infection compared with control animals (no significant difference: p = 0.4). Four days after infection, BDNF gene expression was significantly upregulated (in median 2.4-fold; p = 0.026) (Fig. 3A). In contrast, expression of hippocampal NGF mRNA 30 hours after pneumococcal inoculation was decreased by over 50% (p = 0.004) and showed no significant difference from control animals 4 days after infection (Fig. 3B). No major variation of GDNF mRNA expression was observed (Fig. 3C). In the neocortex, mRNA expression of BDNF, NGF, and GDNF showed no significant differences between mice with meningitis and control animals both 30 hours and 4 days after infection (data not shown).

FIGURE 3.

BDNF (A), NGF (B), GDNF (C), and TrkB (D) mRNA expression in the hippocampal formation of uninfected control mice (C) and animals with pneumococcal meningitis (M) 30 hours and 4 days after infection (n = 11). Values are expressed as arbitrary units normalized to GAPDH (median, 25/75 quartile and minimum/maximum; *, p < 0.05; **, p < 0.01).

In infected animals, gene expression of BDNF was higher in the hippocampal formation than in the neocortex (p = 0.027): in median (25/75 quartile), mRNA levels 4 days after infection were 0.197 (0.133/0.368) versus 0.106 (0.083/0.157). Gene expression of NGF in median was higher in the neocortex than in the hippocampal formation, yet the difference failed to reach statistical significance (0.0004 [0.0003/0.0005] vs. 0.001 [0.0004/0.003]; p = 0.06). The mRNA expression of GDNF 4 days after infection, however, was higher in the neocortex (0.0003 [0.0001/0.0004] vs. 0.001 [0.0008/0.004]; p = 0.0003).

Conversely, in control animals, mRNA gene expression of BDNF and NGF showed no differences between the hippocampus and neocortex (p = 0.514 and p = 0.097) and GDNF mRNA levels were higher in the neocortex (p = 0.001) 4 days after treatment with saline.

Hippocampal Gene Expression of TrkB in Bacterial Meningitis

Analysis of TrkB gene expression in the hippocampal formation 30 hours after infection showed similar levels of mRNA in both the infected and uninfected control group (0.009 [0.007/0.01] vs. 0.007 [0.006/0.02]). Four days after infection, however, TrkB mRNA levels were significantly upregulated in infected animals (Fig. 3D; p = 0.0048).

Detection of BDNF Protein in the Hippocampal Formation

Analysis of BDNF by Western blot revealed higher protein levels in the hippocampal formation of infected mice than in control animals 4 days after infection (Fig. 4A, B; p = 0.0003).

FIGURE 4.

Western blot analysis of BDNF in the hippocampal formation of uninfected control mice and mice with pneumococcal meningitis 4 days after infection. (A) Increased BDNF protein levels in animals 4 days after infection with Streptococcus pneumoniae. (B) Densitometric values are expressed as arbitrary units (n = 10) (median, 25/75 quartile and minimum/maximum; ***, p < 0.001).

Immunohistochemical analysis showed BDNF expression in the hippocampal formation and occasionally in single cells of the neocortex. Within the hippocampal formation, BDNF staining was seen almost exclusively in the hilus of the dentate gyrus and the CA3/4 region of the hippocampus (Fig. 5A, B).

FIGURE 5.

BDNF in experimental pneumococcal meningitis 4 days after infection (A) and in an uninfected control animal (B) (scale bar = 200 μm). ([B], insert) Negative control (staining without primary antibody). BDNF was seen predominantly in the CA3/4 region of the hippocampus and the hilus of the dentate gyrus. Cells expressing tyrosine kinase receptor B (TrkB) in the subgranular layer and the hilus of the dentate gyrus 4 days after infection (C) and in uninfected control (D) (scale bar = 50 μm). ([D], insert) Negative control (staining without primary antibody). Incorporation of BrdU into cells of the subgranular and granule cell layer of the dentate gyrus 4 days after infection. Single immunoreactive cells were also seen outside of the granule layer (E). BrdU immunohistochemistry in uninfected control animals (F) (scale bar = 50 μm). ([F], insert) Negative control (staining without primary antibody).

Detection of TrkB Protein in the Hippocampal Formation

TrkB protein was detected by immunohistochemistry in the hippocampal formation (Fig. 5C, D) and to a lesser extent in the neocortex. The density of cells expressing TrkB was not significantly different between mice with bacterial meningitis and saline-treated control animals 30 hours after initiation of the experiment (data not shown). At day 4, the density of cells expressing TrkB in the dentate gyrus (located predominately in the subgranular layer) was higher in infected than in uninfected control animals (38.2 [12.8/68.7]/mm2 vs. 7.8 [0/13.3]/mm2; p = 0.027). Furthermore, TrkB-positive cells outside of the subgranular layer were detected in the CA3/4 region of the hippocampus and the hilus; however, in this area, the difference between both treatment groups failed to reach statistical significance (62.3 [30.6/84.3]/mm2 vs. 10.5 [5.2/101.2]/mm2; p = 0.13).

Increased Proliferation of Dentate Granule Cells After Bacterial Meningitis

The density of BrdU-labeled cells in the dentate gyrus 30 hours after infection was not significantly different in infected animals and controls (30.1 [16.3/51.9]/mm2 vs. 54.6 [36.3/58.3]/mm2; p = 0.25). The number of proliferating dentate granule cells, however, was higher in infected mice than in the control group 4 days after infection: the density of BrdU-labeled cells was 81.8 (71.6/104.9)/mm2 in mice infected with S. pneumoniae and 28.1 (14.6/38.9)/mm2 in saline-injected control animals (p = 0.0002) (Figs. 5E, F, 6). Single BrdU-positive cells were also present outside of the granule layer.

FIGURE 6.

Density of BrdU-labeled cells in the dentate gyrus of mice 30 hours and 4 days after infection with Streptococcus pneumoniae. BrdU incorporation was increased 4 days after infection (median, 25/75 quartile and min/max; ***, p < 0.001).

Control experiments showed no influence of ceftriaxone on cell proliferation in uninfected mice. The density of BrdU-labeled cells in the dentate gyrus was similar in mice treated with ceftriaxone subcutaneously and in the saline-treated group (35.5 [26.5/55.5]/mm2 vs. 31.5[20.0/36.5]/mm2; p = 0.33).

Double Labeling of Iba1 and BrdU

Double immunofluorescence staining of Iba1, CD45R/B220, and CD3 with BrdU was performed to exclude that BrdU-positive cells represented nonneuronal immune cells. The majority of BrdU-incorporating cells in the dentate gyrus did not colocalize with Iba1 (Fig. 7A-C). Double-labeling of BrdU and Iba1 was observed rarely in single cells in the granule cell layer. Approximately 5% of the BrdU-positive cells counted in the dentate gyrus were positive for both markers (insert of Figure 7A-C). No CD45R/B220- or CD3-positive cells were detected in the hippocampal formation (data not shown).

FIGURE 7.

Detection of Iba1 (green) (A) and BrdU (red) (B) by double-label fluorescent immunocytochemistry ([C]: merge). The majority of BrdU-immunoreactive nuclei did not colocalize with Iba1, indicating that the newly divided cells in the subgranular layer of the dentate gyrus are not microglia/macrophages. However, very few double-labeled cells were detected identifying newly formed microglia/macrophages ([A-C], inserts) (scale bar = 50 μm).

Colocalization of TrkB With NeuN and Calbindin

To further characterize cells expressing TrkB, double-labeling with neuronal and glial marker proteins was performed. Cells expressing TrkB were colocalized with the neuronal proteins NeuN and, to a lesser extent, with calbindin (Fig. 8A-F). Approximately 70% of TrkB-immunolabeled cells expressed NeuN; colocalization with calbindin was observed only in a minority of cells located mainly in the CA3 region of the hippocampus. No evidence for TrkB expression in astrocytic cells was observed: TrkB-immunolabeled cells did not colocalize with GFAP- or BrdU-immunoreactive cells (data not shown).

FIGURE 8.

Colocalization of TrkB with NeuN (A-C) and calbindin (D-F) in the hilus of mouse hippocampal formation. Double-label fluorescent immunocytochemistry was used to detect TrkB (red) (A, D) and NeuN (green) (B) or calbindin (green) (E). The merge of both markers appear yellow, identifying TrkB-positive cells as neurons (scale bar = 50 μm).

Discussion

Proliferation and differentiation of neural progenitor cells is increased in response to a variety of stimuli. In bacterial meningitis, proliferation of neural cells peaked between 2 and 6 days after infection and declined thereafter to basal levels (16). Similarly, in this study, proliferation of dentate granule cells was low 30 hours after infection and showed no difference to the uninfected control group. Four days after infection, the density of BrdU-labeled cells in the subgranular layer of the dentate gyrus was increased significantly in infected compared with uninfected animals. Because BrdU is an unspecific marker for cell proliferation, single immunopositive nuclei were found also outside of the granule layer, in accordance with other models of neural proliferation (46). In bacterial meningitis, invasion of leukocytes into the central nervous system and activation of resident immune cells have been observed. In our experiments, Iba1 immunohistochemistry identified activated microglia/macrophages in the hippocampal formation in infected but not in uninfected animals. In contrast, no B- or T-cells (CD45R/B220 and CD3 immunohistochemistry) were detected in the hippocampal formation during the course of bacterial meningitis. Double-labeling studies were performed to exclude the possibility that the BrdU-positive cells represent dividing immune cells. The vast majority of BrdU-positive cells in the subgranular layer in the dentate gyrus did not colocalize with Iba1-positive cells. As described in this model, approximately 60% of cells labeled by BrdU between 7 and 10 days after infection with S. pneumoniae express neuronal proteins 28 days later (16).

Unspecific stimuli like running (47) or enriched environment (48) are known to induce neurogenesis or to promote neuronal survival. In various models of brain injury, neurogenesis was increased in response to a noxious stimulus, reflecting an endogenous potential of cell replacement. Neurogenesis is also influenced by pharmacologic treatment. Antidepressant therapy by serotonin-reuptake inhibitors induces cell proliferation probably by a noradrenalin- and serotonin-dependent mechanism (33). Conversely, glucocorticoids as well as stress decrease neuronal cell proliferation (49).

In bacterial meningitis, both neurodegeneration and neural proliferation are present in the dentate gyrus. Neuronal injury is caused by migrating leukocytes releasing cytokines, chemokines, and free radicals by the stimulation of resident immune cells and by direct toxicity of bacterial compounds (5). To date, there is no evidence for a direct invasion of bacteria into the hippocampal formation. In experimental meningitis, a green fluorescent protein-transformed clinical isolate of S. pneumoniae was localized exclusively in the ventricles as well as in and around Gr-1+ granulocytes (50). Bacterial compounds, however, are present in the brain parenchyma of infected animals: pneumolysin was detectable both in the neocortex and the hippocampal formation in bacterial meningitis ([51], and unpublished observation). To what extent bacterial products are responsible for the induction of neurogenesis in bacterial meningitis is unknown. No pharmacologic effect of ceftriaxone on neural proliferation was seen in our experiments: the density of BrdU-labeled cells in the dentate gyrus of uninfected mice treated with ceftriaxone was similar to that of the vehicle-treated control group.

Neurotrophic factors are important for cell differentiation and survival. During embryogenesis, BDNF is highly expressed. Its expression has also been found in pathologic conditions such as brain ischemia, hypoglycemic coma (52), or multiple sclerosis (53). Pharmacologic treatment with antidepressants can increase BDNF levels (54). Enhanced BDNF expression that coincided with increased neurogenesis was observed after dietary restriction in rats (55). Similarly, in our investigation, BDNF gene expression and protein synthesis in the hippocampal formation and proliferation of neural progenitor cells was elevated 4 days after induction of bacterial meningitis. BDNF protein was expressed predominantly in the area CA3/4 of the hippocampus and the hilus of the dentate gyrus. In other brain regions (e.g. the neocortex and the basal ganglia), BDNF expression was seen only in single cells at low levels. In rats, BDNF-immunoreactive cells have been observed at the Cornu ammonis of the hippocampus and the granule cell layer of the dentate gyrus. Immunoreactive fibers were most prominent within the areas CA2/3 and the dentate gyrus (56). Within the human hippocampal formation, the neuropil of the hilus and CA3/4 was intensely labeled and BDNF protein was localized in the cell bodies, dendrites, and axons of neurons (57). In our experiments, BDNF immunoreactivity was mainly located in the mossy fibers of the CA3/4 region.

Because effects of BDNF are mediated through TrkB receptors, immunohistochemistry for TrkB was performed to locate the possible areas of BDNF binding and to quantify cells expressing TrkB. TrkB expression was mainly found in the granule cell layer and hilus of the dentate gyrus. The density of TrkB-positive dentate granule cells was higher in mice 4 days after infection with S. pneumoniae than in uninfected controls. Similarly, TrkB mRNA levels were increased in the hippocampal formation. Double immunostaining of TrkB and NeuN identified approximately 70% of TrkB-labeled cells as neurons. No colocalization of TrkB- and BrdU-positive cells was observed.

Elevation of BDNF and TrkB in mature neurons (accompanied by increased BrdU incorporation in dentate granule cells) suggests that BDNF and surrounding Trk-B-expressing neurons contribute to proliferation of progenitor cells. Increased gene expression and protein synthesis of hippocampal BDNF and TrkB might therefore be part of an endogenous mechanism responsible for initiation of neuroregeneration.

NGF is also known as a neurotrophic factor involved in neurogenesis. NGF is expressed in hippocampal as well as neocortical neurons. Its expression is known to be upregulated by kainic acid-induced epileptic activity (58), as well as ischemia (52) and running (59). Therefore, we supposed that the regulation of NGF in bacterial meningitis would also be linked to neuroregeneration. Surprisingly, NGF mRNA levels were significantly decreased during the acute phase of the disease and rose back to basal levels on day 4 after infection. Decreased levels of NGF in acute bacterial meningitis might be the result of cellular stress and injury in neuronal cells. The neuronal damage score revealed a higher density of apoptotic and necrotic cells 30 hours after infection. Neuronal injury in this model of experimental meningitis involves the activation of caspases (60). The consequences of the reduced NGF levels 30 hours after infection, however, remain unclear. Because gene expression was not increased during neural progenitor cell proliferation, NGF does not appear to be a key regulator of progenitor cell activity in the hippocampal formation after bacterial meningitis.

The lack of GDNF mRNA concentration changes indicates that bacterial meningitis does not influence gene expression of this neurotrophic factor. Similarly, Tokunaga et al. found increased concentration of neurotrophin-4 but not of GDNF in the cerebrospinal fluid of patients with bacterial meningitis (61). Moreover, chronic antidepressant treatment did not change GDNF gene expression in the hippocampal formation (54). In contrast to these data, GDNF expression was found to be increased after transient ischemia in rat hippocampus (62, 63). The observations made in this study suggest that GDNF mRNA levels are not influenced by meningitis-induced nervous tissue damage and that this neurotrophic factor is not involved in progenitor cell proliferation in the hippocampal formation after bacterial meningitis.

In conclusion, expression of BDNF and its receptor TrkB were increased 4 days after intracerebral infection with S. pneumoniae. BDNF protein was localized predominantly in the hippocampal CA3/4 region and the hilus adjacent to the subgranular layer of the dentate gyrus. In this area, the density of cells expressing TrkB and the proliferation of neural progenitor cells was increased, suggesting an involvement of endogenous BDNF and TrkB signaling in neurogenesis after bacterial meningitis. Neither NGF nor GDNF expression was increased after infection with S. pneumoniae, suggesting no essential contribution of these growth factors to neurogenesis after bacterial meningitis.

Acknowledgments

The authors thank Stephanie Bunkowski, Doris Bode, and Jutta Laufenberg for excellent technical assistance.

Footnotes

  • This work was supported by the German Research Foundation (DFG) through the Center of Molecular Physiology of the Brain (CMPB) and by a grant to RN (Na 165/4-3). Christine Stadelmann is supported by the Gemeinnützige Hertie-Stiftung and the Medical Faculty of the University of Göttingen (Junior Research Group).

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
View Abstract