Streptococcal meningitis reveals the presence of residual streptococci and down-regulated aquaporin 4 in the brain

The pathology of streptococcal meningitis is poorly understood, even though streptococcal infection induces meningitis. The aim of this study was to clarify the relationship between streptococcal meningitis and aquaporin 4 (AQP4) in the mouse brain. After Streptococcus suis infection, the streptococcal number was calculated, and AQP4 mRNA expression in the brain was quantified at 2 and 7 days after infection. At 7-day post-infection, mice with neurological symptoms showed significantly higher S. suis levels in the brain than mice without neurological symptoms. AQP4 expression was significantly decreased in mice with neurological symptoms than in mice without neurological symptoms. Image analysis demonstrated that S. suis progressed to invade the white matter. Pathological analysis revealed that infected mouse brains had higher inflammation and neurological damage scores than uninfected mouse brains. Therefore, mice with neurological symptoms caused by streptococcal meningitis had high S. suis levels in the brain and reduced AQP4 expression.

Streptococcus suis is a gram-positive, facultative anaerobic bacterium. Approximately 35 S. suis serotypes have been reported (De-Greeff et al. 2002). Serotype 2 is the most virulent serotype and is frequently isolated in swine and humans. Several virulence-associated genes have been reported, such as muramidase-released protein (Smith et al. 1992), extracellular protein factors (Smith et al. 1993), and suilysin (Lun et al. 2003). The primary infection routes of S. suis are infection of a wound (Gottschalk et al. 2010) or the gut from consuming raw pork (Nakayama et al. 2013). Several descriptions of human clinical manifestations of S. suis infection have been published (Werheim et al. 2009). According to our epidemiological study, approximately 20% of patients had diarrhoea and altered consciousness, and hearing loss is a unique characteristic of this infection (Kerdsin et al. 2011). Although mouse models have been used as infectious experimental models to examine the responses of mice to cytokines and chemokines produced during S. suis infection (Dominguez-Punaro et al. 2008), the relationship between streptococcal meningitis and brain pathology has not been adequately studied. Almost all studies have only mentioned meningitis after detecting S. suis in the brain. Aquaporins (AQPs) are membrane proteins involved in water transport within the body (Verkman et al. 2000). AQP4 has been identified in the brain and participates in water homeostasis (Iacovetta et al. 2012). The astrocyte plasma membrane domains that ensheath the cerebral microvessels are enriched in AQP4 water channels, which are strongly implicated in brain Verkman 2007, Tang et al. 2013). However, the relationship between AQP4 expression and streptococcal meningitis remains unknown. Our aim was to clarify the relationship between streptococcal meningitis and AQP4 in the mouse brain.
The S. suis 31533 strain was provided by the National Institute of Animal Health in Japan. Todd-Hewitt broth (Difco Laboratories, Detroit, MI, USA) was used for streptococcal cultures. 22 female specific pathogen-free A/J mice (7-9 weeks) (SLC, Shizuoka, Japan) were acclimated under standard laboratory conditions and provided free access to rodent chow and water. Animal studies were performed in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of Osaka University (Osaka, Japan), and all animal experiments were approved by the Animal Welfare Assurance of Compliance H-21-09-0. 500 µL of streptococcal suspension (5.0 × 10 7 colony forming units (CFU) of S. suis) or the control solution (sterile PBS) were administered to mice via intraperitoneal injection. Animals in a septic state lose the ability to maintain their body temperature, and a decrease in body temperature beyond a certain point (a decrease of 6 °C from normal body temperature) has been correlated with death in several infectious disease models (Olfert et al. 2000). Therefore, mouse body temperature was measured to determine the clinical endpoints.
Three and five mice with neurological symptoms were sacrificed at 2 and 7 days after infection, respectively, and three non-infected mice were sacrificed as a control using 100 µL of Ketalar solution, composed of 10 mL Ketalar (Daiichi-Sankyo, Tokyo, Japan) mixed with 2.2 mL of 2% Selactar (Bayer Health Care, Leverkusen, Germany). The brains were aseptically excised and transferred to 500 µL of sterile PBS in a cell strainer (BD Biosciences, San Jose, CA, USA) and homogenised using the rubber tip of a syringe bar (Terumo, Tokyo, Japan). The homogenised solution was used for RNA extraction. Serial dilutions of 10 µL of the homogenate in PBS were spread onto sheep blood agar plates and incubated at 37 °C for 24 h. Streptococcal colonies were counted and expressed as CFU g -1 for brain samples.
RNA was extracted as previously described (Nakayama et al. 2010). Briefly, PBS-Trizol (Invitrogen, Carlsbad, CA, USA) and chloroform (Wako, Osaka, Japan) were added after washing. After centrifugation, the supernatant was treated with 2-propanol (Wako, Osaka, Japan) and 70% ethanol (Wako). Complementary DNA (cDNA) was obtained using a Roche cDNA kit (Roche, Basel, Switzerland). Real-time polymerase chain reaction (PCR) was performed (Applied Biosystems, Warrington, UK) under the following conditions: 40 cycles of 94 °C for 20 s, 53 °C for 20 s, and 72 °C for 30 s. Two overlapping primers for aquaporin 4 (forward, 5-CTG GAG CCA GCA TGA ATC CAG-3; reverse, 5-TTC TCT CTT CTC CAC GGT CA-3) and β-actin (forward, 5-GTC CCT CAC CCT CCC AAA AG-3; reverse, 5-GCT GCC TCA ACA CCT CAA CCC-3) were used in the present study. Standard DNA amounts corresponding to the target sequences were required to perform real-time PCR. This standardisation was achieved by purifying plasmid DNA containing the target sequences (Furrie et al. 2005). Briefly, cDNA from S. suis 31533 was amplified using a specific PCR primer pair. The product of the correct size and sequence was purified using a PCR purification kit (Qiagen, Valencia, CA, USA) and ligated into a vector using TOPO TA cloning (Invitrogen). DH5 alpha competent Escherichia coli cells were transformed with each ligated vector, and positive colonies were selected. The plasmid from each selected colony was purified using a miniprep system (Qiagen, Valencia, USA).
To clarify the localisation of S. suis, the brain tissue was stained and observed using a fluorescence microscope. The brain of one uninfected mouse was used as a negative control, and the brains of two mice that exhibited neurological symptoms at 7 days after infection were used for immunostaining assays. After excision, the brain tissues were preserved in 10% neutral-buffered formalin (Fujimi Pharmaceutical Company, Osaka, Japan). The brains were sectioned by the Research Foundation for Microbial Diseases at Osaka University, and the sections were embedded in paraffin. The slides were deparaffinised in xylene three times for 5 min each. The slides were immersed in 100% ethanol twice for 5 min each and then immersed in 90% and 80% ethanol. The slides were immersed in 1 mM EDTA (pH 8.0) at 121 °C for 5 min. For immunostaining, the specimens were immersed in 5% albumin in PBS for 60 min. Rabbit anti-S. suis polyclonal antibody (10,000 ×) (Statens Serum Institute, Copenhagen, Denmark) was used as the primary antibody and incubated at 37 °C for 60 min. Alexa Fluor 488-conjugated goat anti-rabbit IgG (Invitrogen) (200 ×) was used as a secondary antibody and incubated with the slides for 60 min. After washing, the slides were stained with haematoxylin and eosin (H&E). Finally, the aqueous mounting medium Permafluor (Thermo Fisher Scientific, Waltham, USA) was added to the specimens. All slides were examined under a fluorescence microscope. A histopathological analysis was conducted to clarify the inflammatory and neuronal damage scores at 7-day post-infection, according to a previously described method (Wellmer et al. 2001). The tissue was sectioned and stained with haematoxylin and eosin (H&E). The stained tissue was assessed using inflammatory and neuronal damage scores. The data are presented as the mean values and standard errors of the mean, and were generated using Student's t test in Fig. 1a, b and the Mann-Whitney test in Fig. 1c.
After measuring the number of bacteria in the brain after infection, we detected 4.5 × 10 6 and 10 CFU/g in mice with neurological symptoms and 8.6 × 10 5 and 2.4 × 10 2 CFU/g in mice without neurological symptoms at 2 and 7 days after infection, respectively (Fig. 1a). There was no difference in the number of bacteria between mice with and without neurological symptoms 2 days after infection; however, the difference was apparent at 7 days after infection. We quantified AQP4 expression after infection and detected 59 and 29 copy numbers in mice with neurological symptoms and 3.7 × 10 3 and 1.2 × 10 5 copy numbers in mice without neurological symptoms at 2 and 7 days after infection, respectively (Fig. 1b). AQP4 expression was significantly decreased after infection and was significantly lower in mice with neurological symptoms than in mice without neurological symptoms at 2 and 7 days after infection.
Streptococcal localisation in the brain 7 days after infection was observed, and the bacteria were localised in the lateral ventricles in the brains of mice that exhibited neurological symptoms. No S. suis was detected in the brains of non-infected mice or mice without neurological symptoms (Fig. 2a). H&E staining results showed leucocyte aggregation in the lateral ventricles of mice with neurological symptoms 7 days after infection, which was consistent with the localisation of the bacteria in the lateral ventricles (Fig. 2b). The brains were analysed based on inflammatory Fig. 1 Influence of streptococcal number and AQP4 expression by streptococcal infection. Brain tissues were sampled at 2and 7-day post-infection. a The number of streptococcal CFUs was calculated, and b mRNA AQP4 expression was quantified by real-time PCR. All data were statistically analysed using Student's t test (a) and Mann-Whitney test (b) (*p < 0.05). and neurological damage scores 7 days after infection. Pathological analysis based on H&E staining showed that mouse brains with and without neurological symptoms had inflammation and neurological damage scores, whereas the brains of uninfected mice had no scores (Table 1).
Streptococcal infection can cause meningitis (brain oedema) and neurological symptoms (Dutkiewicz et al. 2018); however, the pathogenesis of neurological symptoms is not fully understood. When mice become infected and develop neurological symptoms, they never recover. Although almost all mice that exhibited neurological symptoms died of sepsis and bacteraemia, a few mice survived for more than 6 months following the onset of neurological symptoms.
Immunohistochemistry and histopathological analyses revealed that S. suis was localised to the lateral ventricles in the white matter of the brain, and lymphocyte aggregation in mice that exhibited neurological symptoms. Previous studies have reported that brain abnormalities are associated with the development of sensorineural hearing loss caused by ependymoma, which is derived from ependymal cells traversing the central nervous system (Morris et al. 2009) and may develop in response to cytomegalovirus infection (Matsuno et al. 2014). Cytomegalovirus infections in the developing brain may result in abnormalities, such as mental retardation, microcephaly, chorioretinitis, seizures, intracranial calcification, and neurological disorders, including hearing loss. These infections are commonly found in the periventricular white matter region (Moinuddin et al. 2003).
Several patients with S. suis exhibit a unique effect of hearing loss (Kerdsin et al. 2011). Thus, detectable S. suis levels in the lateral ventricle may play an important role in Fig. 2 Image analysis of mice brains infected with S. suis. The brain tissues were sampled at 7-day post-infection. a Sampled brains were subjected to immunohistochemistry, and the presence of S. suis was confirmed by fluorescence microscopy using an Alexa Fluor 488-conjugated antibody. b Histopathological analysis of the brain at 7-day post-infection. The white arrow shows that the lymphocyte aggregation in the lateral ventricle was confirmed in mice with neurological symptoms. Mice showing no sign of neurological symptoms had no detectable S. suis or lymphoid aggregation 7 days after infection the development of hearing loss. Mice that exhibited neurological symptoms harboured higher concentrations of S. suis in the whole brain than mice that did not show neurological signs. Therefore, high concentrations of residual S. suis in the brain white matter strongly correlate with the induction of neurological symptoms in the host. We also examined AQP4 expression. AQP4 is a membrane protein involved in water transport in many fluid-transporting tissues (Niu et al. 2012;Cruz et al. 2013). Although oedema is highly related to AQP4 expression (Tang et al. 2013), the relationship between neurological damage due to infection and AQP4 is poorly understood. AQP4 expression in the brain was decreased in mice that exhibited neurological symptoms at 2-day post-infection. Although few studies have reported the relationship between AQP4 and prevention of microbial infection, water channel AQP4 partially protects the host from cerebral malaria (Promeneur et al. 2013). Moreover, Shiga toxin released by E. coli decreases AQP4 levels throughout the cell, which compromises the integrity of the blood-brain barrier via the activation of astrocytes (Amran et al. 2013).
AQP4 expression in Streptococcus pneumoniae, a haemolysin-producing streptococci that can induce pneumococcal meningitis (Nakayama et al. 2014), was investigated using AQP4 null mice and rats (Papadopoulos and Verkman 2005;Du et al. 2015). The authors reported that pneumococcal infection upregulated AQP4 expression. In this study, the downregulation of AQP4 occurred as a result of infection and production of toxins from the S. suis strains. Therefore, the results here do not indicate that S. suis infection has an opposite effect on AQP4 than observed in S. pneumoniae. Further studies are needed to clarify these relationships. AQP4 upregulation is induced by brain oedema and bacterial meningitis (Huang et al. 2014). In this study, the significant decrease in brain AQP4 expression in mice infected with S. suis may be related to the oedema found around neurons of the cerebral cortex and periventricular regions. Oedema formation around vessels may occur as a result of the increase in blood-brain barrier permeability, allowing the passage of fluid into the extracellular space, and a decrease in AQP expression that prevents fluid elimination Lucero et al. 2012). In S. suis infection, AQP4 expression was decreased more in mice with neurological signs. The AQP4 downregulation detected in astrocytes of mice infected with S. suis could reduce the potassium ion clearance from neurons to adjacent astrocytes, resulting in alteration of neuron functionality . A previous study also showed that AQP4 levels were significantly higher in the cerebral cortex (grey matter) than in other parts of the brain (Han et al. 2004), and immunogold electron microscopy demonstrated that AQP4 is restricted to the glial membrane and ependymal cells (Balladh et al. 2004). Therefore, S. suis invaded the brain much better in mice that exhibited neurological symptoms than in symptom-free mice, and S. suis broke down the cells of the cerebral cortex, including AQP4. S. suis invaded the white matter, at which point the mice began to exhibit neurological symptoms. The present study demonstrated that S. suis was present in the white matter of mice that exhibited neurological symptoms. However, we could not clarify the region of white matter that was related to the induction of neurological symptoms due to S. suis infection, and this clarification is an important topic for future study.
The downregulation of AQP4 in the brain may also be due to functional mechanisms. Astrocytes are involved in inflammatory processes and are activated in response to brain damage. In this case, they may release inflammatory mediators, which may alter the integrity and permeability of the blood-brain barrier and neuronal survival (Abbott et al. 2000). In addition, astrocytes may release neurotrophic factors, which can be neurotoxic to neurons in the pursuit of brain damage (Pehar et al. 2004;Lucero et al. 2012). It is possible that astrocytes undergo astrogliosis when they come into contact with S. suis or extracellular proteins that contain toxins, such as haemolysin released by S. suis. Astrogliosis decreases the expression of AQP4 (Hassan-Olive 2019). Therefore, this may have been caused by S. suis infection in this study. Based on these findings, in this study, AQP4 might be downregulated by S. suis infection.
In conclusion, mice that exhibited neurological symptoms also harboured high S. suis levels and downregulated AQP4 levels in the brain. Image analysis demonstrated that S. suis progressed to invade the white matter in the brain of infected mice.