Systemic P. gingivalis infection induced increased IL-1β in leptomeninges and decreased synaptic marker in leptomeninges proximity cortex, accompanied memory decline in the middle-aged mice.
To test our hypothesis that leptomeningeal cell-induced neuroinflammation might be involved in synaptic failure by releasing soluble factors during P. gingivalis infection, we firstly examined the expression of IL-1β and SYP in the proximity of leptomeninges using the middle-aged mice (fifteen-month-old, female) which were used in our previous studies [14, 34]. The time course of systemic infection with P. gingivalis was shown as an illustration (Fig. 1A). Compared with the control mice, the mice significantly reduced the latency in middle-aged mice after systemic infection with P. gingivalis for 3 weeks (Fig. 1B). However, no different of body weight was found between control mice and the P. gingivalis-infected ones (Fig. 1C), suggesting that the systemic P. gingivalis infection-induced learning and memory deficits in the middle-aged mice are not associated with sickness behaviors. We next focused on expression of IL-1β in the leptomeninges, because IL-1β is the master mediator for inducting neuroinflammation [13, 21]. Compared with the control mice, the IL-1β immunoreactivity was significantly increased in the fibronectin positive leptomeningeal cells of the P. gingivalis infected mice (7.5-folds), demonstrating that IL-1β was induced in leptomeningeal cells in the middle-aged mice after systemic P. gingivalis infection (Fig. 1D, E). Compared with the control mice, the SYP immunoreactivity was significantly decreased (2.18-folds) in the cortex proximity of the leptomeninges in the middle-aged mice after systemic P. gingivalis infection (Fig. 1F, G). The expression of IL-1β was negative correlated with that of SYP in the systemic P. gingivalis- infected mice (Fig. 1H). These observations evidenced that upregulated IL-1β in leptomeningeal cells, downregulated of synaptic marker and memory decline are induced in the middle-aged mice after systemic P. gingivalis infection.
NLRP3 inflammasome involved in IL-1β secretion by primary leptomeningeal cells after P. gingivalis infection
Next, the mechanisms underlying the involvement of P. gingivalis in IL-1β production were studied using primary leptomeningeal cells. To determine the appropriate concentration of living P. gingivalis on primary leptomeningeal cells, we first examined the cell viability of primary leptomeningeal cells infected by P. gingivalis at a multiplicity of infection (MOI) of 1, 5 or 10. The time course of in vitro experiments was set at up to 12 h after P. gingivalis infection, as P. gingivalis can be kept alive for up to 12 h in our cellular models [14, 34].
Compared with the start culture time (0 h), the cell viability of the primary leptomeningeal cells was not decreased up to 12 h at MOIs of 1, 5 or 10 following P. gingivalis infection (Fig. 2A). Therefore, a MOI of 10 was used in subsequent experiments. We focused on IL-1β production by leptomeningeal cells after P. gingivalis infection, as in addition to being the master neuroinflammation regulator, IL-1β exerts direct effects on neurons [16, 35]. Compared with the cells at 0 h, the secretion of IL-1β by leptomeningeal cells was significantly increased from 1 h (10.7-fold), peaked at 3 h (21.8-fold) and was reduced from 6 to 12 h after P. gingivalis infection (Fig. 2B).
We next examined the involvement of NLRP3 inflammasome in IL-1β production by leptomeningeal cells after P. gingivalis infection, as NLRP3 inflammasome activation can be detected during Gram-negative bacteria infection [36]. Compared with 0 h, the mRNA expression of NLRP3 and Caspase-1 was significantly increased in the leptomeningeal cells from 1 h (4.3-fold, 1.6-fold), continued at 3 h (6.5-fold, 2.9-fold) and 6 h (6.9-fold, 3.8-fold) and lasted until 12 h (6.3-fold, 2.7-fold) after P. gingivalis infection (Fig. 2C, D).
To confirm the involvement of NLRP3 inflammasome in IL-1β production in leptomeningeal cells after P. gingivalis infection, we used siRNA to interfere with the NLRP3 mRNA expression. Compared with the control cells, the mRNA expression of NLRP3 was reduced by pretreatment with NLRP3 siRNA in both uninfected cells (49% reduced) and P. gingivalis-infected cells (68% reduced) (Fig. 2E). The protein expression of caspase-1 and IL-1β was also examined. Compared with the control cells, the protein expression of pro-caspase-1 and maturate caspase-1 was significantly increased in the leptomeningeal cells at 3 h after P. gingivalis infection (1.3-fold and 1.6-fold) and significantly decreased by NLRP3 siRNA (by 29% and 30%), respectively (Fig. 2F-H). Compared with the control cells, the protein expression of pro-IL-1β and mature IL-1β was significantly increased in the leptomeningeal cells at 3 h after P. gingivalis infection (2.6-fold, 5-fold) and significantly decreased by NLRP3 siRNA (by 26% and 18%), respectively (Fig. 2I-K). The P. gingivalis infection-induced increase in the secretion of IL-1β was significantly reduced by NLRP3 siRNA (28% reduced) (Fig. 2L). These observations showed that NLRP3 inflammasome involved in IL-1β secretion by leptomeningeal cells after P. gingivalis infection.
CatB is involved in NLRP3 inflammasome activation for IL-1β secretion by primary leptomeningeal cells after P. gingivalis infection
We next examined the involvement of CatB in NLRP3 inflammasome activation in primary leptomeningeal cells after P. gingivalis infection, as CatB is required for NLRP3 inflammasome activation, resulting in IL-1β secretion [20, 21]. Compared with the control cells, fluorescent-labeled P. gingivalis was detected in the lysosome-associated membrane protein 2(LAMP2)-positive lysosomes in leptomeningeal cells at 1 h after P. gingivalis infection, and the prevalence of fluorescent-labeled P. gingivalis in the lysosomes was reduced by pre-treatment with cytochalasin D (Cyto D), a specific phagocytosis inhibitor. These findings show that P. gingivalis can be phagocytosed into lysosomes in leptomeningeal cells after infection (Fig. 3A). Compared with the control cells, the punctate acridine orange aggregates had disappeared by 2 h after P. gingivalis infection, showing that lysosomal damage was induced in leptomeningeal cells after living P. gingivalis was phagocytized (Fig. 3B).
To further confirm the leakage of CatB from the damaged lysosomes, the protein expression of CatB in the cytosol was examined using the cytosol fraction of P. gingivalis-infected leptomeningeal cells. Compared with the control cells, the protein expression of CatB in the cytosol was significantly increased (1.5-fold) at 2 h after P. gingivalis infection (Fig. 3C, D). The involvement of cytosol leakage of CatB in NLRP3 inflammasome activation was then examined. Compared with the control cells, the protein expression of pro-caspase-1 and mature-caspase-1 was significantly increased at 3 h after P. gingivalis infection (1.3-fold, 1.8-fold), and pre-treatment with CA-074Me, the CatB specific inhibitor, significantly decreased the pro-caspase-1 and mature-caspase-1 (by 61% and 20%, respectively) (Fig. 3E-G) and the P. gingivalis-induced increase in IL-1β secretion from the leptomeningeal cells (by 50%) (Fig. 3H).
The effect of phagocytosed P. gingivalis on IL-1β secretion was further examined. Compared with the P. gingivalis infected cells, pre-treatment with Cyto D significantly inhibited the P. gingivalis-upregulated protein expression of mature IL-1β (by 53%) but not that of pro-IL-1β (Fig. 3I-K) and the P. gingivalis-upregulated secretion of IL-1β from the leptomeningeal cells (by 70%) (Fig. 3L). These observations showed that the cytosol leakage of CatB activated NLRP3 inflammasomes, which was required for the secretion of IL-1β by leptomeningeal cells after P. gingivalis infection.
CatB is involved in NF-κB activation for IL-1β production by primary leptomeningeal cells after P. gingivalis infection
CatB plays a critical role in nuclear factor kappa B (NF-κB) activation [13, 14, 37]. We therefore examined the involvement of CatB in NF-κB activation in leptomeningeal cells after P. gingivalis infection. Compared with the cells at 0 h, the mRNA expression of toll-like receptor 2 (TLR2) was significantly increased from 1 h (1.7-fold), lasting through 3, 6 and 12 h (5.2-fold, 5.2- fold, 3.3-fold, respectively) in the leptomeningeal cells after P. gingivalis infection (Fig. 4A). Compared with the cells at 0 h, the mRNA expression of CatB was significantly increased from 1 h (1.7-fold), lasting through to 3, 6 and 12 h (1.2-fold, 2.0-fold, 1.2-fold, respectively) in the leptomeningeal cells after P. gingivalis infection (Fig. 4B). To confirm the involvement of CatB in NF-κB activation in leptomeningeal cells, we next examined the protein expression of CatB from the early time before the peak of IL-1β secretion in leptomeningeal cells after P. gingivalis infection. Compared with the cells at 0 h, the protein expression of CatB was significantly increased from 0.5 h (1.31-fold), lasting through to 1, 2 and 3 h (1.12-fold, 1.12-fold, 1.33-fold, respectively) after P. gingivalis infection (Fig. 4C, D). In contrast, the protein expression of the nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor alpha (IκBα) was significantly decreased at 1 h after P. gingivalis infection (by 24%) compared with the cells at 0 h, and the decrease in IκBα was significantly inhibited by pre-treatment with CA-074Me (by 58%) (Fig. 4E, F).
We further examined the involvement of CatB in the expression of molecules downstream of NF-κB activation. The P. gingivalis infection-induced increase in the mRNA expression of NLRP3, Caspase-1 and IL-1β was significantly inhibited by pre-treatment with CA-074Me in the leptomeningeal cells (by 55%, 87% and 66%, respectively) (Fig. 4G-I). Moreover, the P. gingivalis-induced increase in protein expression of both pro-IL-1β and mature IL-1β was significantly inhibited by pre-treatment with CA-074Me in leptomeningeal cells (by 24% and 78%, respectively) (Fig. 4J-L). These observations showed that CatB was involved in NF-κB activation, which is required for the production of NLRP3 and IL-1β in the leptomeningeal cells after P. gingivalis infection.
P. gingivalis -infected leptomeningeal cells induced IL-1β-dependent synaptic molecule loss in neurons
As leptomeningeal cells anatomically cover the cortex of the brain, we developed a cellular model to investigate the direct influence of P. gingivalis-infected leptomeningeal cells on cortical neurons. Conditioned medium from P. gingivalis-infected leptomeningeal cells (3 h after P. gingivalis infection) was collected (Pg LCM) and applied to primary cultured cortical neurons. As we focused on the outcomes of P. gingivalis-infected leptomeningeal cells in neurons at the synaptic level, we first established the conditions of Pg LCM for the culture of primary cortical neurons. Compared with the control neurons, the cell viability of the primary cortical neurons was not reduced by applying 30% Pg LCM for 24 h (Fig. 5A), although the mean neurite length was significantly reduced by the Pg LCM (by 38%), showing that the morphological features of primary cortical neurons were changed by the Pg LCM without inducing neuron death (Fig. 5B, C). Therefore, 30% Pg LCM was used in the subsequent experiments.
We examined the pre- and post-synaptic protein levels because synaptic deficit is reversible neuronal damage and an early sign of AD. Compared with control neurons, the mRNA expression of pre-synaptic markers, synaptophysin (SYP), vesicular glutamate transporter1(VGLUT1) and synapsin1 (SYN1) as well as the post-synaptic marker postsynaptic density protein 95 (PSD95) was significantly decreased by the Pg LCM in primary cortical neurons (by 57%, 73%, 44% and 50%, respectively) (Fig. 5D), and the Pg LCM-induced reduction in SYP was significantly and positively correlated with that in VGLUT1 (r = 0.9683, p = 0.0015) (Fig. 5E). We next analyzed the relationship between the secreted IL-1β in the Pg LCM and synaptic markers in the Pg LCM-applied primary cortical neurons. The mRNA expression of SYP, VGLUT1, SYN1 and PSD95 in the Pg LCM-applied primary cortical neurons was significantly and negatively correlated with the concentration of IL-1β in Pg LCM (r=-0.8635, p = 0.0267; r=-0.8628, p = 0.0270; r=-0.9066, p = 0.0127 and r=-0.8997, p = 0.0146, respectively) (Fig. 5F). The protein expression of SYP was further examined. Compared with control neurons, the protein expression of SYP was dramatically decreased by the Pg LCM (by 65%) in the primary cortical neurons which was significantly inhibited by pre-treatment with NLRP3 siRNA or CA-074 Me in leptomeningeal cells (by 15%, 52%) (Fig. 5G, H). The Pg LCM-induced reduction in the protein expression of SYP in primary cortical neurons was dramatically inhibited by pre-treatment with IL-1 receptor antagonist (IL-1Ra) in neurons (by 69%) (Fig. 5I, J). These observations showed that leptomeningeal cells induced soluble IL-1β-dependent synaptic distribution after P. gingivalis infection.
P. gingivalis -infected leptomeningeal cells induced the IL-1β-dependent suppression of BDNF signaling in neurons
To further explore the effect of IL-1β secreted by P. gingivalis-infected leptomeningeal cells on neurons, we focused on the effects of Pg LCM on BDNF signaling in neurons using stable mouse N2a cells [38].
We first examined the BDNF-induced expression of activity-regulated cytoskeleton-associated protein (Arc), a critical immediate early gene that plays an essential role in synaptic plasticity of neurons [39]. Compared with the start culture time (0 min), the mRNA expression of Arc was induced from 10 min (4-fold), peaked at 30 min (5.3-fold), continued through 60 min and lasted until 120 min (3-fold, 2-fold, respectively) in N2a cells after pre-treatment with BDNF, and the BDNF-induced Arc expression at 30 min was dramatically reduced by pre-treatment with Pg LCM (by 76%) (Fig. 6A).
Next, we examined the effect of IL-1β in Pg LCM on activation of protein kinase B/cAMP response element binding protein (Akt/CREB), the molecules of BDNF signaling. Compared with control cells, pre-treatment with BDNF for 2 h significantly increased the Akt phosphorylation at Ser473 (1.2-fold), and the BDNF-induced increase in Akt phosphorylation was significantly decreased by pre-treatment with Pg LCM (by 58%) (Fig. 5B, C). The Pg LCM-induced decrease in Akt phosphorylation was significantly reversed by pre-treatment with IL-1Ra (by 13%) (Fig. 6B, C).
We next examined the activation of CREB, a transcription factor downstream of Akt phosphorylation. Compared with control cells, pre-treatment with BDNF for 2 h significantly increased the CREB phosphorylation at Ser133 (1.2-fold), and the BDNF-induced increase in CREB phosphorylation was significantly decreased by pre-treatment with Pg LCM (by 15%). The Pg LCM-induced decrease in CREB phosphorylation was significantly reversed by pre-treatment with IL-1Ra (by 82%) (Fig. 6D, E).
In addition, the nuclear localization of phosphorylated CREB, which represents the activation of CREB, was also examined. Compared with control cells, the nuclear localization of phosphorylated CREB was significantly increased at 4 h in the BDNF pre-treated N2a cells (4.8-fold), and the BDNF-induced increase in CREB nuclear localization was significantly decreased by pre-treatment with Pg LCM (by 84%). The Pg LCM-induced decrease in CREB nuclear localization was significantly reversed by pre-treatment with IL-1Ra (by 40%) (Fig. 6F, G).
Taken together, these observations showed that P. gingivalis-infected leptomeningeal cells induced the IL-1β-dependent suppression of BDNF signaling in neurons.
Propolis modulated the IL-1β-related BDNF production by primary leptomeningeal cells after P. gingivalis infection
As structures covering the surface of the brain, leptomeningeal cells are known to protect neurons by producing neuroprotective factors. We therefore explored the involvement of IL-1β in the production BDNF, a critical neurotrophic factor, by leptomeningeal cells after P. gingivalis infection.
Compared with the cells at 0 h, the mRNA expression of BDNF and IL-1β was significantly increased from 1 h (8.7-fold, 8.5-fold) to 3 h (14-fold, 62-fold), respectively (Fig. 7A, B) in the leptomeningeal cells after P. gingivalis infection. We further explored the potential utility of natural materials for moderating the production of IL-1β and BDNF in P. gingivalis-infected leptomeningeal cells. To this end, we focused on propolis, which was shown to prevent cognitive decline in elderly subjects [33]. The cell viability was examined to determine the suitable condition of propolis for primary leptomeningeal cells. Compared with control cells, the cell viability was not significantly decreased until pre-treatment with propolis at 10 µg/ml (Fig. 7C), so 10 µg/ml of propolis was used in the following experiments. To our surprise, compared with the P. gingivalis-infected cells, the expression of BDNF was significantly increased (by 28%) while that of IL-1β was markedly decreased (by 78%) in the P. gingivalis-infected cells following pre-treatment with propolis (Fig. 7D, E). The mRNA expression of BDNF was significantly and negatively correlated with the IL-1β mRNA expression in leptomeningeal cells (r=-0.8968, p = 0.0154) (Fig. 7F). Compared with the P. gingivalis-infected cells, the mRNA expression of CatB in the P. gingivalis-infected leptomeningeal cells was significantly decreased by pre-treatment with propolis (by 40%) (Fig. 7G), which paralleled the findings of NLRP3 and Caspase-1 (20% and 24% decrease) (Fig. S1). The propolis-induced reduction in the mRNA expression of NLRP3 and Caspase-1 was significantly and positively correlated with that of CatB in the P. gingivalis-infected leptomeningeal cells (r = 0.9343, p = 0.0063; r = 0.9882, p = 0.0002, respectively) (Fig. S2). In contrast, the CatB mRNA expression was positively correlated with IL-1β but negatively correlated with BDNF mRNA (Fig. 7H, I).
Taken together, these observations showed that BDNF production was downregulated by the CatB-mediated IL-1β upregulation in leptomeningeal cells after P. gingivalis infection, and propolis upregulated BDNF by inhibiting IL-1β in leptomeningeal cells after P. gingivalis infection.