- Email: [email protected]
S100B in the cerebrospinal fluid—A marker for glial damage in the rabbit model of pneumococcal meningitis
Neuroscience Letters 475 (2010) 104–107
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
S100B in the cerebrospinal fluid—A marker for glial damage in the rabbit model of pneumococcal meningitis H. Schmidt ∗ , J. Gerber, K. Stuertz, M. Djukic, S. Bunkowski, F.R. Fischer, M. Otto, R. Nau University of Goettingen, Department of Neurology/Department of Medical Psychology and Medical Sociology, Robert-Koch-Str. 40, D-37099 Goettingen, Germany
a r t i c l e
i n f o
Article history: Received 5 November 2009 Received in revised form 8 March 2010 Accepted 21 March 2010 Keywords: Bacterial meningitis CSF S100B Rabbit Experimental meningitis
a b s t r a c t The rabbit model provides an important experimental setting for the evaluation of antibiotic agents against pneumococcal meningitis. One of the primary targets of this model is the study of neuronal and glial cell damage in bacterial meningitis. The aim of this investigation was to evaluate whether a significant increase of S100B in the cerebrospinal fluid (CSF) as an indicator of white matter damage could be observed in this meningitis model. Seven rabbits were infected intracisternally with S. pneumoniae, and CSF S100B concentrations were examined serially before infection, at 12 h, 14 h, 17 h, 20 h, and at 24 h after infection. The course of CSF S100B increase and its relation to other parameters of brain tissue destruction and CSF inflammation were measured. Axonal damage was visualized by amyloid precursor protein (APP) immunostaining and demyelination by Luxol Fast Blue/Periodic Acid Schiff (LFB-PAS) stain. In each animal, we observed a distinct rise in S100B concentration in the CSF due to pneumococcal meningitis. We conclude that the CSF concentration of the glial S100B protein can be used as an additional parameter for future interventional studies focusing on glial cell damage in the rabbit model of bacterial meningitis. © 2010 Elsevier Ireland Ltd. All rights reserved.
Human bacterial meningitis can cause neurological [1] and neuropsychological sequelae [2,3] resulting from the brain swelling [4] and toxic bacterial products [5,6]. Once the brain is affected by the inflammation, both neuronal and glial cells can die and release specific intracellular proteins into the cerebrospinal fluid (CSF). Patients with bacterial meningitis and meningoencephalitis show increased CSF concentrations of neuron-specific enolase (NSE) as a neuronal marker of destruction [7], and of S100B as a marker of glial activation [8] and glial damage [9–11]. Morphologically, white matter lesions are seen at autopsy [12] in patients with bacterial meningitis, and sometimes also intra vitam by cerebral magnetic resonance tomography [13–15]. The clinical importance of these brain lesions in humans is reflected by their association with persisting cognitive sequelae [3]. Thus, studies on meningitis in animal models should aim at reducing both neuronal damage and white matter lesions. The increase of S100B as a result of bacterial meningitis has been reported in several studies in humans [10,11,16–18] but has not yet been studied in the rabbit model of pneumococcal meningitis. After intramuscular induction of anesthesia with ketamine (25 mg/kg of body weight) and xylazine (5 mg/kg), seven New Zealand White rabbits (weight approximately 2.5 kg) were anaes-
∗ Corresponding author. Tel.: +49 551 398484; fax: +49 551 398405. E-mail address: [email protected] (H. Schmidt). 0304-3940/$ – see front matter © 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2010.03.059
thetized with intravenous (i.v.) urethane for 24 h and placed in a stereotaxic frame by means of a dental acrylic helmet fixed at the scull with four screws as originally described by Dacey and Sande [19]. A spinal needle (22 × 3.5 in., Braun, Melsungen, Germany) was suboccipitally inserted into the cisterna magna of the rabbits. Meningitis was induced by injection of 106 colony forming units (CFU) of S. pneumoniae type 3. Blood and CSF were taken before, at 12 h, 14 h, 17 h, 20 h, and 24 h after infection. At 12 h after infection, antimicrobial therapy was initiated with a 20 mg/kg i.v. bolus and a maintenance dose of 10 mg of ceftriaxone per kg/h i.v. until the rabbits were sacrificed 24 h after infection with an i.v. bolus of thiopental. Seven anaesthetized uninfected rabbits served as control animals. After 24 h their CSF S100B concentrations were measured to rule out any hypoxic effect of the narcosis. A single rabbit was inoculated with heat killed uncapsulated R6 pneumococci instead of viable SP3 pneumococci to evaluate whether S100B release into the CSF can be induced by killed pneumococci. The rabbit brains were fixed in formalin, embedded in paraffin, and immunostained for amyloid precursor protein (APP) performed with the primary monoclonal antibody mAb348 (Chemicon, Temecula, CA) and the secondary antibody Z0259 (Dako, Danmark) which was exposed afterwards to immunocomplexes of alkaline phosphatase/anti-alkaline phosphatase (D0651 Dako, Danmark). The reaction was visualized by neufuchsin, and the sections were counterstained with Mayer’s hemalum. This solution detects axons with pathological axonal
H. Schmidt et al. / Neuroscience Letters 475 (2010) 104–107
105
Table 1 Descriptive statistics of CSF parameters and neuronal/axonal damage during experimental pneumococcal meningitis. Time after infection (h)
CSF S100B (ng/ml) SP3 pneumococci‡ CSF leukocyte count (/ml) CSF bacterial density (log CFU/ml) CSF TNF activity (U/l) CSF LTA concentration (ng/ml) CSF lactate (mmol/l) CSF protein (mg/l) CSF NSE (pg/ml) APP positivity rate [area % of white matter] Apoptotic neurons (n/mm2 )
0
12
14
3.35 ± 3.87
88.88 ± 145.56
151.27 ± 202.2§
3 ± 2*
9711 ± 23386 7.78 ± 0.72*§
8051 ± 8698* 7.43 ± 1.02*§
4.63 501.5 4.6 928.5
± ± ± ±
2.36§ 547.6*§ 2.4* 1337.2
17
7.8 ± 1.55*§ 1164.7 ± 1210.8*§
287.3 ± 280.31*§ 14732 ± 6957* 5.82 ± 1.08*§ 13.64 ± 4.22*§ 1350.0 ± 1413.4*§
20
24
310.74 ± 379.00*§
252.57 ± 188.58*§
8507 ± 4381* 5.08 ± 1.43*§ 12.38 ± 1.89*§ 1031.8 ± 1071.6*§
17886 ± 8261* 3.00 ± 1.62* 12.53 1073.9 10.8 4435.3 58.3 1.88
± ± ± ± ± ±
7.97*§ 924.2*§ 1.60* 2149.1* 69.6 1.11
145 ± 58
#
Baseline values in rabbits derived from historical controls. Normally distributed parameters were marked by an asterisk (*). A significant increase of the respective mean as compared to the mean initial value at 0 h is marked by a paragraph sign (§ ). ‡ CSF S100B values for a rabbit inoculated with heat-killed R6 pneumococci: 12 h: 21.1 ng/ml; 14 h 80.6 ng/ml; 17 h 20.5 ng/ml, 20 h 27.0 ng/ml; 24 h 8.0 ng/ml. Baseline CSF values without infection: TNF 0.65 + 0.65 U/l (n = 11), lactate 2.15 + 1.33 mmol/l (n = 11), protein 537 + 284 mg/l (n = 25), NSE 2.48 + 2.47 pg/ml (n = 25).
transport. An impaired axonal transport can be a late consequence of demyelination. There is no antibody against damaged myelin basic protein available for rabbit which could reveal early demyelination (dMBP-Ab) [20]. To detect the larger areas of demyelination, we performed a LFB-PAS staining known to detect myelin damage [21]. Computerized quantification of the demyelination detected by LFB-PAS was not possible due to low contrast differences but APP-positive areas could be segmented electronically (ImageJ software, Version 1.421; NIH, New Bethesda, USA). The ratio of the APP-positive area to the total white matter area was calculated. The density of apoptotic neurons in the dentate gyrus was quantified by in situ tailing, planimetry of the dentate gyrus and consecutive microscopical counting [6]. The CSF leukocyte density was counted with a Fuchs-Rosenthal haemocytometer. Pneumococcal titers in CSF were calculated by plating 10 l of serially 10-fold diluted samples on sheep blood agar plates, incubated overnight at 37 ◦ C with 5% CO2 . The bacterial densities at 12 h, 14 h, 17 h, 20 h, and 24 h were used for log-linear regression analysis in order to calculate the bacterial killing rate. The remaining CSF was centrifuged at 3000 × g for 5 min, and the supernatants were immediately frozen at −80 ◦ C. Leukocyte count, bacterial density, TNF activity, lipoteichoic acid (LTA), and S100B concentration were determined in each CSF specimen at each time. CSF protein and lactate concentrations were measured at 12 h and 24 h, and neuron-specific enolase (NSE) at the end of the experiment. S100B was measured using a commercially available immunoluminescence assay (Diasorin, Dietzenbach, Germany) designed for humans. CSF lactate concentrations were determined photometrically using a commercially available lactate oxidase reaction kit (Rolf Greiner Biochemica, Flacht, Germany), and protein concentration with a nephelometer (DosaScat nephelometer, Dosatech, Gilching, Germany). A luminescence immunoassay was used for the measurement of the NSE concentration (LIAmat, Diasorin, Germany). CSF lipoteichoic/teichoic acids (LTA) were quantified with an ELISA previously described in detail [22]. The CSF tumor necrosis factor (TNF) activity was measured using a cytolytic assay with a L929 fibroblast cell line [23]. Statistical calculations were performed with GraphPad Prism software (GraphPad Software, La Jolla, USA). Data were expressed as means ± standard deviations. The distribution for normality was tested by the Kolmogorov–Smirnov test. A significant increase of a parameter as compared to its initial value at 0 h was calculated by Wilcoxon’s signed rank test.
For parameters with multiple serial determinations, the maxima and the areas under the curve were correlated either with Pearson’s or with Spearman’s correlation (with normal or non-Gaussian distribution, respectively). The decrease of the bacterial density after the initiation of antibiotic treatment was calculated by logarithmiclinear regression. In all animals, inoculation of Streptococcus pneumoniae type 3 (SP3) resulted in an increase of CSF leukocytes after 12 h. TNF activity reached its maximum at 17 h after infection. All animals reliably displayed increasing CSF S100B concentrations after 12 h of meningitis. CSF S100B reached its peak concentration 20 h after infection. The S100B concentrations after 24 h of urethane anaesthesia measured in seven uninfected rabbits were mean ± SD 5.2 ± 9.1 ng/ml (median [25th /75th perc.] 0 [0/7.65]) which was not different from the S100B values before infection in the group with meningitis (U-test p = 0.2). In the rabbit inoculated with unencapsulated heat-killed R6 pneumococci a release of S100B into the CSF could also be observed, indicating that the cell wall constituents of pneumococci were sufficient to induce an increase of CSF S100B (Table 1). Small areas of axonal damage were detectable in every animal (Table 1). This axonal damage was distributed diffusely, and only in two rabbits were larger areas of APP-positive axons observed. Paler regions in the LFB-PAS stain were seen in each rabbit brain but a distinctive demyelination was visible in only one rabbit (see Fig. 1). The pattern of demyelination was less prominent in extent and severity than in other demyelinating diseases such as multiple sclerosis. After 12 h of infection TNF activity (r = 0.8, p = 0.03) was correlated with the area under the curve of CSF S100B, while NSE or other parameters of inflammation in CSF were not. The NSE concentration in CSF was, however, positively correlated with CSF lactate (r = 0.8, p = 0.02). Data for all measured parameters are provided in Table 1. In all animals APP-positive axons (Fig. 1) were detected but the area of APP-positive axons was correlated only with the bacterial concentration at 24 h after infection (r = 0.85, p = 0.02) although not with the concentration of NSE. Likewise, the density of apoptotic neurons in the dentate gyrus was neither associated with the percentage of APP-positive axons, nor with the CSF NSE or the S100B concentrations. CSF S100B protein is a glial protein well known as a prognostic tool for patients after cerebral hypoxia [24–28]. Although S100B was found to be increased in the CSF in human bacterial meningitis, the release of S100B into the CSF and its correlation with other parameters of inflammation or brain tissue damage in the rab-
106
H. Schmidt et al. / Neuroscience Letters 475 (2010) 104–107
protein is derived from rabbits and therefore unsuitable for this purpose. Despite the lack of a significant correlation of S100B concentrations with quantified axonal damage, lower concentrations of S100B correlated with fewer neurological complications in children [32] and a significantly (r > 0.5) better clinical outcome in adults after bacterial meningitis (Schmidt et al., unpublished data), again underlining the importance of white matter pathology for the formation of persisting functional damage in pneumococcal meningitis. Hence, white matter lesions are pathophysiological elements that should be studied in the future in experimental meningitis settings. In summary, S100B is reliably increased in the CSF of rabbits with pneumococcal meningitis and thus should be considered an indicator for white matter damage in this animal model. Fig. 1. (a) Light microscopy (10×) of LFB-PAS stain of a demyelinated region (marked by arrows) with (intersection b)) APP-positive axons (Light microscopy (10×)) in a rabbit after 24 h of pneumococcal meningitis.
bit model of experimental pneumococcal meningitis has not been evaluated up to now. The data presented here show that CSF S100B concentrations rise reproducibly 12 h after infection, reaching their maximum 20 h after the induction of meningitis. The finding of increased S100B CSF levels throughout the experiment and the axonal and white matter lesions after 24 h of meningitis demonstrate the important effect of inflammation on glial activation and damage. Glial injury did not depend on the presence of viable bacterial by the increased S100B release into the CSF of a rabbit that was inoculated with heat-killed uncapsulated R6 pneumococci (Table 1). Persistent inflammation in meningitis is maintained by the increased levels of bacterial cell wall constituents in the CSF leading to brain damage long after the sterilization of the CSF. This has been demonstrated not only in the rabbit model [29] but also in other animal models of experimental meningitis [30] and in human beings [22]. Neither the concentrations of CSF S100B, nor the densities of APP-positive axons, nor the density of apoptotic neurons in the dentate gyrus correlated with the concentrations of neuron-specific enolase at the end of the experiment. However, due to the small number of animals in this study, statistical significance would only be possible if a very strong correlation of the investigated parameters had been found. In addition, in this experiment we could only study a small temporal segment of the overall time course of pneumococcal meningitis. Glial damage occurring later than in the first 24 h, would have escaped our notice. The total extent of glial and axonal damage and the correlation of both may therefore remain undetected with this experimental approach even though we found some APP-positive axons and pale regions in the LFBPAS stain as signs of axonal damage and demyelination in the white matter. Another reason for the discrepancy between microscopically visible cellular damage and CSF S100B release might be that glial activation [8,31] rather than the ensuing cellular damage was responsible for a fraction of the S100B concentration in the CSF. With this experiment we cannot determine to which proportion white matter damage and glial activation contribute to the total release of S100B. It is known that white matter is affected in pneumococcal meningitis of the rabbit; therefore this kind of damage can be studied in future experiments using this experimental model. Since we were unable to quantify the white matter damage in the LFB-PAS stain, we were not able to answer the question as to whether the extent of demyelination, correlated with the glial CSF marker S100B, or whether S100B release might be, due more to glial activation than to its degradation. The only available antibody against degraded myelin basic
Acknowledgment We thank Ms. C. Bunker for the thorough English proofreading.
References [1] V.A. Bohr, N. Rasmussen, Neurological sequelae and fatality as prognostic measures in 875 cases of bacterial meningitis, Dan Med. Bull. 35 (1988) 92–95. [2] M. Hoogman, D. van de Beek, M. Weisfelt, J. de Gans, B. Schmand, Cognitive outcome in adults after bacterial meningitis, J. Neurol. Neurosurg. Psychiatry 78 (2007) 1092–1096. [3] H. Schmidt, B. Heimann, M. Djukic, C. Mazurek, C. Fels, C.W. Wallesch, R. Nau, Neuropsychological sequelae of bacterial and viral meningitis, Brain 129 (2006) 333–345. [4] U.M. Niemoller, M.G. Tauber, Brain edema and increased intracranial pressure in the pathophysiology of bacterial meningitis, Eur. J. Clin. Microbiol. Infect. Dis. 8 (1989) 109–117. [5] Y.D. Bifrare, C. Gianinazzi, H. Imboden, S.L. Leib, M.G. Tauber, Bacterial meningitis causes two distinct forms of cellular damage in the hippocampal dentate gyrus in infant rats, Hippocampus 13 (2003) 481–488. [6] G. Zysk, W. Bruck, J. Gerber, Y. Bruck, H.W. Prange, R. Nau, Anti-inflammatory treatment influences neuronal apoptotic cell death in the dentate gyrus in experimental pneumococcal meningitis, J. Neuropathol. Exp. Neurol. 55 (1996) 722–728. [7] S. Inoue, H. Takahashi, K. Kaneko, The fluctuations of neuron-specific enolase (nse) levels of cerebrospinal fluid during bacterial meningitis: the relationship between the fluctuations of nse levels and neurological complications or outcome, Acta Paediatr. Jpn. 36 (1994) 485–488. [8] R. Bianchi, C. Adami, I. Giambanco, R. Donato, S100b binding to rage in microglia stimulates cox-2 expression, J. Leukoc. Biol. 81 (2007) 108–118. [9] R. Gerlach, G. Demel, H.G. König, U. Gross, J.H. Prehn, A. Raabe, V. Seifert, D. Kögel, Active secretion of s100b from astrocytes during metabolic stress, Neuroscience 141 (2006) 1697–1701. [10] D. Gazzolo, D. Grutzfeld, F. Michetti, A. Toesca, M. Lituania, M. Bruschettini, A. Dobrzanska, P. Bruschettini, Increased s100b in cerebrospinal fluid of infants with bacterial meningitis: relationship to brain damage and routine cerebrospinal fluid findings, Clin. Chem. 50 (2004) 941–944. [11] J. Unden, B. Christensson, J. Bellner, C. Alling, B. Romner, Serum s100b levels in patients with cerebral and extracerebral infectious disease, Scand J. Infect Dis. 36 (2004) 10–13. [12] J. Gerber, R.C. Seitz, S. Bunkowski, W. Bruck, R. Nau, Evidence for frequent focal and diffuse acute axonal injury in human bacterial meningitis, Clin. Neuropathol. 28 (2009) 33–39. [13] S. Appenzeller, A.V. Faria, V.A. Zanardi, S.R. Fernandes, L.T. Costallat, F. Cendes, Vascular involvement of the central nervous system and systemic diseases: etiologies and mri findings, Rheumatol. Int. 28 (2008) 1229–1237. [14] P. Beleza, M. Ribeiro, J. Pereira, C. Ferreira, M.J. Jordao, F. Almeida, Probable acute disseminated encephalomyelitis due to haemophilus influenzae meningitis, Dev. Med. Child Neurol. 50 (2008) 388–391. [15] P.G. Jorens, P.M. Parizel, M. Wojciechowski, A. Laridon, A. De Weerdt, G. Mertens, B. Ceulemans, Streptococcus pneumoniae meningoencephalitis with unusual and widespread white matter lesions, Eur. J. Paediatr. Neurol. 12 (2008) 127–132. ˜ [16] J.R. Infante, A. Martínez, J. Ochoa, F. Canadillas, M. Torres-Avisbal, J.A. Vallejo, F.M. González, C. Pacheco, J.M. Latre, Cerebrospinal fluid s-100 protein levels in neurological pathologies, J. Physiol. Biochem. 59 (2003) 255–261. [17] H. Lins, C.W. Wallesch, M.T. Wunderlich, Sequential analyses of neurobiochemical markers of cerebral damage in cerebrospinal fluid and serum in cns infections, Acta Neurol. Scand. 112 (2005) 303–308. [18] P.C. Spinella, A. Donoghue, A. Rajendra, H.R. Drott, T.E. Dominguez, M. Helfaer, Cerebrospinal fluid levels of s-100beta in children and its elevation in pediatric meningitis, Pediatr. Crit. Care Med. 5 (2004) 53–57.
H. Schmidt et al. / Neuroscience Letters 475 (2010) 104–107 [19] R.G. Dacey, M.A. Sande, Effect of probenecid on cerebrospinal fluid concentrations of penicillin and cephalosporin derivatives, Antimicrob Agents Chemother. 6 (1974) 437–441. [20] J.K. Wasserman, L.C. Schlichter, White matter injury in young and aged rats after intracerebral hemorrhage, Exp. Neurol. 214 (2008) 266– 275. [21] W.L. Lin, C. Zehr, J. Lewis, M. Hutton, S.H. Yen, D.W. Dickson, Progressive white matter pathology in the spinal cord of transgenic mice expressing mutant (p301l) human tau, J. Neurocytol. 34 (2005) 397–410. [22] K. Stuertz, I. Merx, H. Eiffert, E. Schmutzhard, M. Mader, R. Nau, Enzyme immunoassay detecting teichoic and lipoteichoic acids versus cerebrospinal fluid culture and latex agglutination for diagnosis of streptococcus pneumoniae meningitis, J. Clin. Microbiol. 36 (1998) 2346–2348. [23] D.A. Flick, G.E. Gifford, Comparison of in vitro cell cytotoxic assays for tumor necrosis factor, J. Immunol. Methods 68 (1984) 167–175. [24] B.W. Böttiger, S. Möbes, R. Glätzer, H. Bauer, A. Gries, P. Bärtsch, J. Motsch, E. Martin, Astroglial protein s-100 is an early and sensitive marker of hypoxic brain damage and outcome after cardiac arrest in humans, Circulation 103 (2001) 2694–2698. [25] M.C. Geraghty, M.T. Torbey, Neuroimaging and serologic markers of neurologic injury after cardiac arrest, Neurol. Clin. 24 (2006), 107–121, vii.
107
[26] S. Hachimi-Idrissi, M. Van der Auwera, J. Schiettecatte, G. Ebinger, Y. Michotte, L. Huyghens, S-100 protein as early predictor of regaining consciousness after out of hospital cardiac arrest, Resuscitation 53 (2002) 251–257. [27] D.R. Li, B.L. Zhu, T. Ishikawa, D. Zhao, T. Michiue, H. Maeda, Postmortem serum protein s100b levels with regard to the cause of death involving brain damage in medicolegal autopsy cases, Leg. Med. (Tokyo) 8 (2006) 71–77. [28] K. Thorngren-Jerneck, C. Alling, A. Herbst, I. Amer-Wahlin, K. Marsal, S100 protein in serum as a prognostic marker for cerebral injury in term newborn infants with hypoxic ischemic encephalopathy, Pediatr. Res. 55 (2004) 406–412. [29] K. Stuertz, H. Schmidt, F. Trostdorf, H. Eiffert, M. Mader, R. Nau, Lower lipoteichoic and teichoic acid csf concentrations during treatment of pneumococcal meningitis with non-bacteriolytic antibiotics than with ceftriaxone, Scand. J. Infect. Dis. 31 (1999) 367–370. [30] R. Nau, H. Eiffert, Modulation of release of proinflammatory bacterial compounds by antibacterials: potential impact on course of inflammation and outcome in sepsis and meningitis, Clin. Microbiol. Rev. 15 (2002) 95–110. [31] R. Donato, G. Sorci, F. Riuzzi, C. Arcuri, R. Bianchi, F. Brozzi, C. Tubaro, I. Giambanco, S100b’s double life: Intracellular regulator and extracellular signal, Biochim. Biophys. Acta 1793 (2009) 1008–1022. [32] S.A. Hamed, E.A. Hamed, M.M. Zakary, Oxidative stress and s-100b protein in children with bacterial meningitis, BMC Neurol. 9 (2009) 51.
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
S100B in the cerebrospinal fluid—A marker for glial damage in the rabbit model of pneumococcal meningitis H. Schmidt ∗ , J. Gerber, K. Stuertz, M. Djukic, S. Bunkowski, F.R. Fischer, M. Otto, R. Nau University of Goettingen, Department of Neurology/Department of Medical Psychology and Medical Sociology, Robert-Koch-Str. 40, D-37099 Goettingen, Germany
a r t i c l e
i n f o
Article history: Received 5 November 2009 Received in revised form 8 March 2010 Accepted 21 March 2010 Keywords: Bacterial meningitis CSF S100B Rabbit Experimental meningitis
a b s t r a c t The rabbit model provides an important experimental setting for the evaluation of antibiotic agents against pneumococcal meningitis. One of the primary targets of this model is the study of neuronal and glial cell damage in bacterial meningitis. The aim of this investigation was to evaluate whether a significant increase of S100B in the cerebrospinal fluid (CSF) as an indicator of white matter damage could be observed in this meningitis model. Seven rabbits were infected intracisternally with S. pneumoniae, and CSF S100B concentrations were examined serially before infection, at 12 h, 14 h, 17 h, 20 h, and at 24 h after infection. The course of CSF S100B increase and its relation to other parameters of brain tissue destruction and CSF inflammation were measured. Axonal damage was visualized by amyloid precursor protein (APP) immunostaining and demyelination by Luxol Fast Blue/Periodic Acid Schiff (LFB-PAS) stain. In each animal, we observed a distinct rise in S100B concentration in the CSF due to pneumococcal meningitis. We conclude that the CSF concentration of the glial S100B protein can be used as an additional parameter for future interventional studies focusing on glial cell damage in the rabbit model of bacterial meningitis. © 2010 Elsevier Ireland Ltd. All rights reserved.
Human bacterial meningitis can cause neurological [1] and neuropsychological sequelae [2,3] resulting from the brain swelling [4] and toxic bacterial products [5,6]. Once the brain is affected by the inflammation, both neuronal and glial cells can die and release specific intracellular proteins into the cerebrospinal fluid (CSF). Patients with bacterial meningitis and meningoencephalitis show increased CSF concentrations of neuron-specific enolase (NSE) as a neuronal marker of destruction [7], and of S100B as a marker of glial activation [8] and glial damage [9–11]. Morphologically, white matter lesions are seen at autopsy [12] in patients with bacterial meningitis, and sometimes also intra vitam by cerebral magnetic resonance tomography [13–15]. The clinical importance of these brain lesions in humans is reflected by their association with persisting cognitive sequelae [3]. Thus, studies on meningitis in animal models should aim at reducing both neuronal damage and white matter lesions. The increase of S100B as a result of bacterial meningitis has been reported in several studies in humans [10,11,16–18] but has not yet been studied in the rabbit model of pneumococcal meningitis. After intramuscular induction of anesthesia with ketamine (25 mg/kg of body weight) and xylazine (5 mg/kg), seven New Zealand White rabbits (weight approximately 2.5 kg) were anaes-
∗ Corresponding author. Tel.: +49 551 398484; fax: +49 551 398405. E-mail address: [email protected] (H. Schmidt). 0304-3940/$ – see front matter © 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2010.03.059
thetized with intravenous (i.v.) urethane for 24 h and placed in a stereotaxic frame by means of a dental acrylic helmet fixed at the scull with four screws as originally described by Dacey and Sande [19]. A spinal needle (22 × 3.5 in., Braun, Melsungen, Germany) was suboccipitally inserted into the cisterna magna of the rabbits. Meningitis was induced by injection of 106 colony forming units (CFU) of S. pneumoniae type 3. Blood and CSF were taken before, at 12 h, 14 h, 17 h, 20 h, and 24 h after infection. At 12 h after infection, antimicrobial therapy was initiated with a 20 mg/kg i.v. bolus and a maintenance dose of 10 mg of ceftriaxone per kg/h i.v. until the rabbits were sacrificed 24 h after infection with an i.v. bolus of thiopental. Seven anaesthetized uninfected rabbits served as control animals. After 24 h their CSF S100B concentrations were measured to rule out any hypoxic effect of the narcosis. A single rabbit was inoculated with heat killed uncapsulated R6 pneumococci instead of viable SP3 pneumococci to evaluate whether S100B release into the CSF can be induced by killed pneumococci. The rabbit brains were fixed in formalin, embedded in paraffin, and immunostained for amyloid precursor protein (APP) performed with the primary monoclonal antibody mAb348 (Chemicon, Temecula, CA) and the secondary antibody Z0259 (Dako, Danmark) which was exposed afterwards to immunocomplexes of alkaline phosphatase/anti-alkaline phosphatase (D0651 Dako, Danmark). The reaction was visualized by neufuchsin, and the sections were counterstained with Mayer’s hemalum. This solution detects axons with pathological axonal
H. Schmidt et al. / Neuroscience Letters 475 (2010) 104–107
105
Table 1 Descriptive statistics of CSF parameters and neuronal/axonal damage during experimental pneumococcal meningitis. Time after infection (h)
CSF S100B (ng/ml) SP3 pneumococci‡ CSF leukocyte count (/ml) CSF bacterial density (log CFU/ml) CSF TNF activity (U/l) CSF LTA concentration (ng/ml) CSF lactate (mmol/l) CSF protein (mg/l) CSF NSE (pg/ml) APP positivity rate [area % of white matter] Apoptotic neurons (n/mm2 )
0
12
14
3.35 ± 3.87
88.88 ± 145.56
151.27 ± 202.2§
3 ± 2*
9711 ± 23386 7.78 ± 0.72*§
8051 ± 8698* 7.43 ± 1.02*§
4.63 501.5 4.6 928.5
± ± ± ±
2.36§ 547.6*§ 2.4* 1337.2
17
7.8 ± 1.55*§ 1164.7 ± 1210.8*§
287.3 ± 280.31*§ 14732 ± 6957* 5.82 ± 1.08*§ 13.64 ± 4.22*§ 1350.0 ± 1413.4*§
20
24
310.74 ± 379.00*§
252.57 ± 188.58*§
8507 ± 4381* 5.08 ± 1.43*§ 12.38 ± 1.89*§ 1031.8 ± 1071.6*§
17886 ± 8261* 3.00 ± 1.62* 12.53 1073.9 10.8 4435.3 58.3 1.88
± ± ± ± ± ±
7.97*§ 924.2*§ 1.60* 2149.1* 69.6 1.11
145 ± 58
#
Baseline values in rabbits derived from historical controls. Normally distributed parameters were marked by an asterisk (*). A significant increase of the respective mean as compared to the mean initial value at 0 h is marked by a paragraph sign (§ ). ‡ CSF S100B values for a rabbit inoculated with heat-killed R6 pneumococci: 12 h: 21.1 ng/ml; 14 h 80.6 ng/ml; 17 h 20.5 ng/ml, 20 h 27.0 ng/ml; 24 h 8.0 ng/ml. Baseline CSF values without infection: TNF 0.65 + 0.65 U/l (n = 11), lactate 2.15 + 1.33 mmol/l (n = 11), protein 537 + 284 mg/l (n = 25), NSE 2.48 + 2.47 pg/ml (n = 25).
transport. An impaired axonal transport can be a late consequence of demyelination. There is no antibody against damaged myelin basic protein available for rabbit which could reveal early demyelination (dMBP-Ab) [20]. To detect the larger areas of demyelination, we performed a LFB-PAS staining known to detect myelin damage [21]. Computerized quantification of the demyelination detected by LFB-PAS was not possible due to low contrast differences but APP-positive areas could be segmented electronically (ImageJ software, Version 1.421; NIH, New Bethesda, USA). The ratio of the APP-positive area to the total white matter area was calculated. The density of apoptotic neurons in the dentate gyrus was quantified by in situ tailing, planimetry of the dentate gyrus and consecutive microscopical counting [6]. The CSF leukocyte density was counted with a Fuchs-Rosenthal haemocytometer. Pneumococcal titers in CSF were calculated by plating 10 l of serially 10-fold diluted samples on sheep blood agar plates, incubated overnight at 37 ◦ C with 5% CO2 . The bacterial densities at 12 h, 14 h, 17 h, 20 h, and 24 h were used for log-linear regression analysis in order to calculate the bacterial killing rate. The remaining CSF was centrifuged at 3000 × g for 5 min, and the supernatants were immediately frozen at −80 ◦ C. Leukocyte count, bacterial density, TNF activity, lipoteichoic acid (LTA), and S100B concentration were determined in each CSF specimen at each time. CSF protein and lactate concentrations were measured at 12 h and 24 h, and neuron-specific enolase (NSE) at the end of the experiment. S100B was measured using a commercially available immunoluminescence assay (Diasorin, Dietzenbach, Germany) designed for humans. CSF lactate concentrations were determined photometrically using a commercially available lactate oxidase reaction kit (Rolf Greiner Biochemica, Flacht, Germany), and protein concentration with a nephelometer (DosaScat nephelometer, Dosatech, Gilching, Germany). A luminescence immunoassay was used for the measurement of the NSE concentration (LIAmat, Diasorin, Germany). CSF lipoteichoic/teichoic acids (LTA) were quantified with an ELISA previously described in detail [22]. The CSF tumor necrosis factor (TNF) activity was measured using a cytolytic assay with a L929 fibroblast cell line [23]. Statistical calculations were performed with GraphPad Prism software (GraphPad Software, La Jolla, USA). Data were expressed as means ± standard deviations. The distribution for normality was tested by the Kolmogorov–Smirnov test. A significant increase of a parameter as compared to its initial value at 0 h was calculated by Wilcoxon’s signed rank test.
For parameters with multiple serial determinations, the maxima and the areas under the curve were correlated either with Pearson’s or with Spearman’s correlation (with normal or non-Gaussian distribution, respectively). The decrease of the bacterial density after the initiation of antibiotic treatment was calculated by logarithmiclinear regression. In all animals, inoculation of Streptococcus pneumoniae type 3 (SP3) resulted in an increase of CSF leukocytes after 12 h. TNF activity reached its maximum at 17 h after infection. All animals reliably displayed increasing CSF S100B concentrations after 12 h of meningitis. CSF S100B reached its peak concentration 20 h after infection. The S100B concentrations after 24 h of urethane anaesthesia measured in seven uninfected rabbits were mean ± SD 5.2 ± 9.1 ng/ml (median [25th /75th perc.] 0 [0/7.65]) which was not different from the S100B values before infection in the group with meningitis (U-test p = 0.2). In the rabbit inoculated with unencapsulated heat-killed R6 pneumococci a release of S100B into the CSF could also be observed, indicating that the cell wall constituents of pneumococci were sufficient to induce an increase of CSF S100B (Table 1). Small areas of axonal damage were detectable in every animal (Table 1). This axonal damage was distributed diffusely, and only in two rabbits were larger areas of APP-positive axons observed. Paler regions in the LFB-PAS stain were seen in each rabbit brain but a distinctive demyelination was visible in only one rabbit (see Fig. 1). The pattern of demyelination was less prominent in extent and severity than in other demyelinating diseases such as multiple sclerosis. After 12 h of infection TNF activity (r = 0.8, p = 0.03) was correlated with the area under the curve of CSF S100B, while NSE or other parameters of inflammation in CSF were not. The NSE concentration in CSF was, however, positively correlated with CSF lactate (r = 0.8, p = 0.02). Data for all measured parameters are provided in Table 1. In all animals APP-positive axons (Fig. 1) were detected but the area of APP-positive axons was correlated only with the bacterial concentration at 24 h after infection (r = 0.85, p = 0.02) although not with the concentration of NSE. Likewise, the density of apoptotic neurons in the dentate gyrus was neither associated with the percentage of APP-positive axons, nor with the CSF NSE or the S100B concentrations. CSF S100B protein is a glial protein well known as a prognostic tool for patients after cerebral hypoxia [24–28]. Although S100B was found to be increased in the CSF in human bacterial meningitis, the release of S100B into the CSF and its correlation with other parameters of inflammation or brain tissue damage in the rab-
106
H. Schmidt et al. / Neuroscience Letters 475 (2010) 104–107
protein is derived from rabbits and therefore unsuitable for this purpose. Despite the lack of a significant correlation of S100B concentrations with quantified axonal damage, lower concentrations of S100B correlated with fewer neurological complications in children [32] and a significantly (r > 0.5) better clinical outcome in adults after bacterial meningitis (Schmidt et al., unpublished data), again underlining the importance of white matter pathology for the formation of persisting functional damage in pneumococcal meningitis. Hence, white matter lesions are pathophysiological elements that should be studied in the future in experimental meningitis settings. In summary, S100B is reliably increased in the CSF of rabbits with pneumococcal meningitis and thus should be considered an indicator for white matter damage in this animal model. Fig. 1. (a) Light microscopy (10×) of LFB-PAS stain of a demyelinated region (marked by arrows) with (intersection b)) APP-positive axons (Light microscopy (10×)) in a rabbit after 24 h of pneumococcal meningitis.
bit model of experimental pneumococcal meningitis has not been evaluated up to now. The data presented here show that CSF S100B concentrations rise reproducibly 12 h after infection, reaching their maximum 20 h after the induction of meningitis. The finding of increased S100B CSF levels throughout the experiment and the axonal and white matter lesions after 24 h of meningitis demonstrate the important effect of inflammation on glial activation and damage. Glial injury did not depend on the presence of viable bacterial by the increased S100B release into the CSF of a rabbit that was inoculated with heat-killed uncapsulated R6 pneumococci (Table 1). Persistent inflammation in meningitis is maintained by the increased levels of bacterial cell wall constituents in the CSF leading to brain damage long after the sterilization of the CSF. This has been demonstrated not only in the rabbit model [29] but also in other animal models of experimental meningitis [30] and in human beings [22]. Neither the concentrations of CSF S100B, nor the densities of APP-positive axons, nor the density of apoptotic neurons in the dentate gyrus correlated with the concentrations of neuron-specific enolase at the end of the experiment. However, due to the small number of animals in this study, statistical significance would only be possible if a very strong correlation of the investigated parameters had been found. In addition, in this experiment we could only study a small temporal segment of the overall time course of pneumococcal meningitis. Glial damage occurring later than in the first 24 h, would have escaped our notice. The total extent of glial and axonal damage and the correlation of both may therefore remain undetected with this experimental approach even though we found some APP-positive axons and pale regions in the LFBPAS stain as signs of axonal damage and demyelination in the white matter. Another reason for the discrepancy between microscopically visible cellular damage and CSF S100B release might be that glial activation [8,31] rather than the ensuing cellular damage was responsible for a fraction of the S100B concentration in the CSF. With this experiment we cannot determine to which proportion white matter damage and glial activation contribute to the total release of S100B. It is known that white matter is affected in pneumococcal meningitis of the rabbit; therefore this kind of damage can be studied in future experiments using this experimental model. Since we were unable to quantify the white matter damage in the LFB-PAS stain, we were not able to answer the question as to whether the extent of demyelination, correlated with the glial CSF marker S100B, or whether S100B release might be, due more to glial activation than to its degradation. The only available antibody against degraded myelin basic
Acknowledgment We thank Ms. C. Bunker for the thorough English proofreading.
References [1] V.A. Bohr, N. Rasmussen, Neurological sequelae and fatality as prognostic measures in 875 cases of bacterial meningitis, Dan Med. Bull. 35 (1988) 92–95. [2] M. Hoogman, D. van de Beek, M. Weisfelt, J. de Gans, B. Schmand, Cognitive outcome in adults after bacterial meningitis, J. Neurol. Neurosurg. Psychiatry 78 (2007) 1092–1096. [3] H. Schmidt, B. Heimann, M. Djukic, C. Mazurek, C. Fels, C.W. Wallesch, R. Nau, Neuropsychological sequelae of bacterial and viral meningitis, Brain 129 (2006) 333–345. [4] U.M. Niemoller, M.G. Tauber, Brain edema and increased intracranial pressure in the pathophysiology of bacterial meningitis, Eur. J. Clin. Microbiol. Infect. Dis. 8 (1989) 109–117. [5] Y.D. Bifrare, C. Gianinazzi, H. Imboden, S.L. Leib, M.G. Tauber, Bacterial meningitis causes two distinct forms of cellular damage in the hippocampal dentate gyrus in infant rats, Hippocampus 13 (2003) 481–488. [6] G. Zysk, W. Bruck, J. Gerber, Y. Bruck, H.W. Prange, R. Nau, Anti-inflammatory treatment influences neuronal apoptotic cell death in the dentate gyrus in experimental pneumococcal meningitis, J. Neuropathol. Exp. Neurol. 55 (1996) 722–728. [7] S. Inoue, H. Takahashi, K. Kaneko, The fluctuations of neuron-specific enolase (nse) levels of cerebrospinal fluid during bacterial meningitis: the relationship between the fluctuations of nse levels and neurological complications or outcome, Acta Paediatr. Jpn. 36 (1994) 485–488. [8] R. Bianchi, C. Adami, I. Giambanco, R. Donato, S100b binding to rage in microglia stimulates cox-2 expression, J. Leukoc. Biol. 81 (2007) 108–118. [9] R. Gerlach, G. Demel, H.G. König, U. Gross, J.H. Prehn, A. Raabe, V. Seifert, D. Kögel, Active secretion of s100b from astrocytes during metabolic stress, Neuroscience 141 (2006) 1697–1701. [10] D. Gazzolo, D. Grutzfeld, F. Michetti, A. Toesca, M. Lituania, M. Bruschettini, A. Dobrzanska, P. Bruschettini, Increased s100b in cerebrospinal fluid of infants with bacterial meningitis: relationship to brain damage and routine cerebrospinal fluid findings, Clin. Chem. 50 (2004) 941–944. [11] J. Unden, B. Christensson, J. Bellner, C. Alling, B. Romner, Serum s100b levels in patients with cerebral and extracerebral infectious disease, Scand J. Infect Dis. 36 (2004) 10–13. [12] J. Gerber, R.C. Seitz, S. Bunkowski, W. Bruck, R. Nau, Evidence for frequent focal and diffuse acute axonal injury in human bacterial meningitis, Clin. Neuropathol. 28 (2009) 33–39. [13] S. Appenzeller, A.V. Faria, V.A. Zanardi, S.R. Fernandes, L.T. Costallat, F. Cendes, Vascular involvement of the central nervous system and systemic diseases: etiologies and mri findings, Rheumatol. Int. 28 (2008) 1229–1237. [14] P. Beleza, M. Ribeiro, J. Pereira, C. Ferreira, M.J. Jordao, F. Almeida, Probable acute disseminated encephalomyelitis due to haemophilus influenzae meningitis, Dev. Med. Child Neurol. 50 (2008) 388–391. [15] P.G. Jorens, P.M. Parizel, M. Wojciechowski, A. Laridon, A. De Weerdt, G. Mertens, B. Ceulemans, Streptococcus pneumoniae meningoencephalitis with unusual and widespread white matter lesions, Eur. J. Paediatr. Neurol. 12 (2008) 127–132. ˜ [16] J.R. Infante, A. Martínez, J. Ochoa, F. Canadillas, M. Torres-Avisbal, J.A. Vallejo, F.M. González, C. Pacheco, J.M. Latre, Cerebrospinal fluid s-100 protein levels in neurological pathologies, J. Physiol. Biochem. 59 (2003) 255–261. [17] H. Lins, C.W. Wallesch, M.T. Wunderlich, Sequential analyses of neurobiochemical markers of cerebral damage in cerebrospinal fluid and serum in cns infections, Acta Neurol. Scand. 112 (2005) 303–308. [18] P.C. Spinella, A. Donoghue, A. Rajendra, H.R. Drott, T.E. Dominguez, M. Helfaer, Cerebrospinal fluid levels of s-100beta in children and its elevation in pediatric meningitis, Pediatr. Crit. Care Med. 5 (2004) 53–57.
H. Schmidt et al. / Neuroscience Letters 475 (2010) 104–107 [19] R.G. Dacey, M.A. Sande, Effect of probenecid on cerebrospinal fluid concentrations of penicillin and cephalosporin derivatives, Antimicrob Agents Chemother. 6 (1974) 437–441. [20] J.K. Wasserman, L.C. Schlichter, White matter injury in young and aged rats after intracerebral hemorrhage, Exp. Neurol. 214 (2008) 266– 275. [21] W.L. Lin, C. Zehr, J. Lewis, M. Hutton, S.H. Yen, D.W. Dickson, Progressive white matter pathology in the spinal cord of transgenic mice expressing mutant (p301l) human tau, J. Neurocytol. 34 (2005) 397–410. [22] K. Stuertz, I. Merx, H. Eiffert, E. Schmutzhard, M. Mader, R. Nau, Enzyme immunoassay detecting teichoic and lipoteichoic acids versus cerebrospinal fluid culture and latex agglutination for diagnosis of streptococcus pneumoniae meningitis, J. Clin. Microbiol. 36 (1998) 2346–2348. [23] D.A. Flick, G.E. Gifford, Comparison of in vitro cell cytotoxic assays for tumor necrosis factor, J. Immunol. Methods 68 (1984) 167–175. [24] B.W. Böttiger, S. Möbes, R. Glätzer, H. Bauer, A. Gries, P. Bärtsch, J. Motsch, E. Martin, Astroglial protein s-100 is an early and sensitive marker of hypoxic brain damage and outcome after cardiac arrest in humans, Circulation 103 (2001) 2694–2698. [25] M.C. Geraghty, M.T. Torbey, Neuroimaging and serologic markers of neurologic injury after cardiac arrest, Neurol. Clin. 24 (2006), 107–121, vii.
107
[26] S. Hachimi-Idrissi, M. Van der Auwera, J. Schiettecatte, G. Ebinger, Y. Michotte, L. Huyghens, S-100 protein as early predictor of regaining consciousness after out of hospital cardiac arrest, Resuscitation 53 (2002) 251–257. [27] D.R. Li, B.L. Zhu, T. Ishikawa, D. Zhao, T. Michiue, H. Maeda, Postmortem serum protein s100b levels with regard to the cause of death involving brain damage in medicolegal autopsy cases, Leg. Med. (Tokyo) 8 (2006) 71–77. [28] K. Thorngren-Jerneck, C. Alling, A. Herbst, I. Amer-Wahlin, K. Marsal, S100 protein in serum as a prognostic marker for cerebral injury in term newborn infants with hypoxic ischemic encephalopathy, Pediatr. Res. 55 (2004) 406–412. [29] K. Stuertz, H. Schmidt, F. Trostdorf, H. Eiffert, M. Mader, R. Nau, Lower lipoteichoic and teichoic acid csf concentrations during treatment of pneumococcal meningitis with non-bacteriolytic antibiotics than with ceftriaxone, Scand. J. Infect. Dis. 31 (1999) 367–370. [30] R. Nau, H. Eiffert, Modulation of release of proinflammatory bacterial compounds by antibacterials: potential impact on course of inflammation and outcome in sepsis and meningitis, Clin. Microbiol. Rev. 15 (2002) 95–110. [31] R. Donato, G. Sorci, F. Riuzzi, C. Arcuri, R. Bianchi, F. Brozzi, C. Tubaro, I. Giambanco, S100b’s double life: Intracellular regulator and extracellular signal, Biochim. Biophys. Acta 1793 (2009) 1008–1022. [32] S.A. Hamed, E.A. Hamed, M.M. Zakary, Oxidative stress and s-100b protein in children with bacterial meningitis, BMC Neurol. 9 (2009) 51.