Continuous injection of LPS for 7 days improved systemic inflammation in septic mice
We aimed to construct a mice model of sepsis by administering continuous LPS injections. To evaluate systemic inflammation in mice that received LPS injection that last for a period of time, we assessed scored the overall condition of the mice by various aspects such as smoothness of hair, locomotor status, and post-stunning response. This was used as the basis for the subsequent experimental grouping. We also measure the blood neutrophil-to-lymphocyte ratio (NLR), inflammatory cytokines and their body weight variations in three groups of mice respectively. Overall, the group of mice that received LPS injection had lower body weight compared to mice that received PBS injection, while mice that received LPS injection for 7 days regained body weight compared with the control group (Fig. 1a). The total scores of the 4-day group also showed a deterioration in the status of the mice, while in 7-day group the mice are getting better (Fig. 1b). In addition, the ratio of neutrophil to lymphocyte (NLR) in mouse blood, which is used to reflect systemic inflammatory status, found significantly increased inflammatory response in 4 days of mice. While the NLR ratio decreased in mice that received 7 days of LPS treatment (Fig. 1c). We also verified the inflammatory factors in the blood of the mice, in order to further illustrate the systemic inflammatory response of the mice. ELISA showed that the pro-inflammatory factor, such as TNF-α, IL-6, elevated in 4-day mice, while the pro-inflammatory factors in the 7-day mice decreased to approximately the same level as those in the vehicle-treated group (Fig. 1d). Whereas the anti-inflammatory factors IL-10 were raised in the 7-day mice, but did not change significantly in the 4-day mice (Fig. 1e), indicating that pro-inflammatory factors were increased in mice receiving LPS injections for 4 days. But anti-inflammatory factors were increased in mice receiving LPS injections for 7 days. Weight, MSS and blood indicators from mice showed that mice injected with LPS for 4 days were in worse condition and showed more severe systemic inflammatory responses compared to mice injected for 7 days.
Continuous injection of LPS improves behavioral cognitive dysfunction in mice
To verify whether intraperitoneal injection of LPS could cause inflammation in CNS, we conducted novel object recognition (NOR) and Y-maze experiments to explore whether learning and spatial memory abilities were impaired in mice that received injections over a long period of time. We performed behavioral tests on mice in both the group of control and in each of groups injected with LPS on days one to seven, and we found a similar trend of post-impairment re-recovery of cognitive dysfunction in mice (Fig. S1a, b). In the absence of statistical differences in the total time spent exploring objects by the three groups of mice (Fig. 2b), the novel object discrimination index (NODI) was significantly lower in the group of 4-day mice compared to the vehicle and 7-day mice group of mice (Fig. 2a). Similarly in the Y-maze experiment, our data showed that 7-day mice had more spontaneous alternation than 4-day mice, which was not significantly different from the vehicle mice (Fig. 2c). Thus, mice injected with LPS for 4 consecutive days had impaired cognitive function, while mice injected with LPS for 7 consecutive days had restored cognitive function. By HE staining of hippocampal tissue, the nerve cells in the hippocampus of 4-day mice showed nucleus sequestration, deep staining and morphological changes compared with the vehicle group (Fig. 2d). Furthermore, the mitochondria of 4-day mice group by electron microscopy revealed that mitochondrial membrane was broken, mitochondrial cristae disappeared and vacuole formation (Fig. 2e), indicating that the mitochondrial structure destruction of mouse hippocampus in 4 days would further cause the dysfunction of mitochondria.
Serial injection of LPS for 7 days attenuated neuroinflammation in septic mice
Other than that, ELISA about inflammatory cytokines in the hippocampus indicated increased expression of TNF-α, IL-6 in 4-day mice (Fig. 3a, b and increased expression of IL-10 and IFN-β in hippocampal tissues of 7-day mice compared with the vehicle group (Fig. 3c, d). Similarly, we validated this result by qPCR (Fig. 3e, f). The above results indicate that not only the systemic inflammatory response was increased in the 4-day group of mice, but also affected the inflammatory response in the hippocampal tissue of the same group of mice. In the 7-day group, not only the systemic inflammatory response was reduced, but also the inflammation of the CNS was restored.
LPS reduces the expression of TREM2
The ability of TREM2 to produce neuroprotective effects in central inflammation has been well documented[14]. TREM2 exerts anti-inflammatory effects and promotes apoptotic neurons in many diseases including neurodegenerative diseases, ischemia/ reperfusion injury and bacterial infections phagocytosis[14–18]. To investigate whether the LPS affects expression of TREM2 in microglia, we explored whether TREM2 expression is altered in mouse microglia. It was found by qPCR and Western blot that the expression of TREM2 in LPS-treated mice decreased in the first period (before 4 days), but increased from day 5 onwards (Fig. 4a, b). In addition, the hippocampal sections from three groups were also examined by confocal microscopy to examine the co-localization of TREM2 and microglial marker myeloid cell marker ionized calcium-binding adapter molecule 1 (Iba-1). The results suggested that TREM2 was significantly upregulated and overlapped with activated microglia Iba-1 in 7-day mice (Fig. 4c). These results suggest LPS exacerbates central inflammatory conditions whereas it attenuates TREM2 expression. However, a significant rise in TREM2 expression in activated microglia in the 7-day mouse group alleviated the inflammatory response in CNS cells. Additionally, the polarization of microglia was assessed by PCR, and the expression of M1 markers (iNOS and CD86) was increased in the hippocampus of 4-day mice compared to vehicle mice. Yet the expression of M2 markers (IL-10, CD206) was increased in the hippocampus of 7-day mice (Fig. 4d). From these results we speculated that TREM2 participates in microglial M2 polarization thereby attenuates central inflammatory response.
To further confirm the results of in vivo experiments, in vitro assays were performed on BV-2 from mice. The results showed that BV2 cells treated with 1 µg/ml, 5ug/ml or 10 µg/ml LPS for 24 h had much lower expression of TREM2 than that in the control group (Fig. 4e-g). Among them, 1 ug/ml LPS treatment downregulated TREM2 levels to half of the control group. As is well understood, LPS treatment promotes microglia polarization toward M1 type, at the same time M1 type microglia increase with increasing LPS concentration, reducing the M2 anti-inflammatory phenotype exacerbating the inflammatory response (Fig. 4h) The above results indicate that LPS treatment can reduce TREM2 expression in vitro and in vivo and alternate the microglial polarization status. To further verify whether LPS changed the polarization state of microglia via TREM2 and thus, we overexpressed TREM2 in BV-2. TREM2 was found to be significantly elevated in BV-2 cells by exogenous treatment (Fig. S2a), and the same treatment of LPS on BV-2 cells with overexpression TREM2 revealed a decrease in the M1-type polarization of microglia (Fig. S2b).
IFN-β enhances TREM2 expression
Since the level of IFN-β and IL-10 in the hippocampal tissue was validated by ELISA and qPCR, we investigate whether there is a connection between IFN-β or IL-10 and TREM2. We handled BV-2 cells with different concentrations of IFN-β or IL-10 for 24 hours in vitro. We found TREM2 expression improving in the IFN-β treated compared to control (Fig. 5a, Bb. Nevertheless, the expression of TREM2 was not significantly changed in cells treated with IL-10 (Fig. 5c, d). In particular, IFN-β treatment induced a concentration-dependent upregulation of TREM2 (Fig. 5a, b, e). IFN-I has different roles in multiple diseases. Nevertheless, in many neurological diseases, it has been reported IFN-I is able to block pro-inflammatory mediators and induce the effects of anti-inflammatory factors[19, 20].
Recent studies have shown that TREM2 is able to affect Arg-1 expression, which is induced by anti-inflammatory factors, after affecting the expression of microglial transcription factor STAT6[11]. This indicates that TREM2 plays an instrumental role in microglial polarization to M2 phenotype and in promoting central nervous anti-inflammatory effects. To mimic the in vivo environment, we co-incubated 1 ug/ml of LPS with BV-2 for 1h and then treated BV-2 cells with different concentrations of IFN-β for 24h. In this regard, we speculate that a certain concentration range of IFN-β has a promoting effect on the conversion of microglia to the M2 phenotype. The levels of IL-10 and Arg-1 were examined to reveal the activation of M2 phenotype microglia. We discovered the expression of IL-10 and Arg-1 were significantly upregulated, and the production of CD 86 and CD16 decreased (Fig. 5f). Therefore, the above results demonstrated that a certain concentration range of IFN-β promoted microglia polarization to M2 phenotype by increasing TREM2 expression, which was able to mitigate neuroinflammation.
TREM2 shRNA in BV-2 cell reduces the conversion of the microglial M1 phenotype to M2 phenotype
Next, we have investigated the effect of Trem2 shRNA on the alternative anti-inflammatory activation of microglia in response to IFN-β treatment. We verified the expression of TREM2 after TREM2 shRNA and showed a significant decrease in TREM2 expression in sh-TREM2 compared to the NC group (Fig. 6a, b). Since the greatest increase in TREM2 expression was observed in the IFN-β group at 1ug/ml, thus this concentration was chosen to complete the subsequent experiments. It was evident that protein levels of TREM2 in sh-TREM2 group were reduced after IFN-β treatment compared to NC group (Fig. 6c). To examine the activation of the M1 and M2 phenotypes in microglia, we detected decreased mRNA levels of IL-10 and ARG-1 of the M2 phenotype (Fig. 6d). This result indicated that shRNA of TREM2 reduces the anti-inflammatory capacity of microglia, and exogenous IFN-β is able to recover few portions of the TREM2 expression, but can’t promote microglia to M2 phenotype conversion totally, which played a pivotal role in abating central nervous inflammation.