DISCUSSION
Past clinical experience has shown that patients with sepsis benefit from de-escalation antibiotic therapy, but the mechanism has not been explored in depth. This study is the first time to elaborate the mechanism of DE therapy from animal experiments. Our results show that different β-Lactams have different effects on NETs formation bothex vivo and in vivo , and we demonstrated that de-escalation antibiotic therapy, including the early use of NET-promoting antibiotics and the late use of NET-inhibiting antibiotics, could fully exploit the immunomodulatory effects of NET formation during the different stages of sepsis. More importantly, these results underline the important role of NETs directed initial empirical antibiotic therapy in the infection of polymicrobial sepsis in vivo .
Similar with previous studies, our data emphasize the critical antibacterial role of NETs in the early stage and the deleterious role of NETs in the late stage of sepsis [17, 19]. Neutrophils serve as the first line of defense in the innate immune system against bacterial pathogens. The release of intracellular contents, which is called neutrophil extracellular traps, is critical for killing bacteria in infection and sepsis [3]. Current studies have shown that NETs and their elements can cause organ injury, coagulation disorders, autoimmune diseases, and thromboembolism [20]. Although NETs are a double-edged sword in sepsis, few studies have focused on their antibacterial function in the early stage of sepsis, during which NETs can trap pathogens and prevent bacterial dissemination [17, 18, 21]. Our data show clearing NETs with DNaseI will lead to the spread of infection and increased organ damage, these results prove that NETs formation playing an important role in organ protection during the early stage of sepsis.
We demonstrated, for the first time, that in addition to macrolide and quinolone drugs, β-lactam antibiotics can also modulate the formation of NETs in vivo and in vitro . Previous studies have proven that some antibiotics may exert immunomodulatory functions, and with the discovery of NETs, more antibiotics have been found to affect the immune system [22-24]. However, no studies have shown immunomodulatory functions of β-lactam antibiotics, and some studies have shown a lack of immunomodulatory function of β-lactam antibiotics. This study has proven that different subgroups of β-lactam antibiotics have distinct effects on NET formation and that imipenem and ceftriaxone regulate NETs in a ROS-dependent manner. ROS generation is a well-recognized step in NET formation, and many studies have shown that NET-associated diseases may be alleviated by the usage of ROS scavengers or inhibitors [25]. The β-lactam antibiotic-induced NET regulation observed in this study could be inhibited by NOX inhibitors or promoted by exogenous ROS, which confirmed the involvement of ROS.
De-escalation therapy has been shown to be associated with reduced mortality in patients with severe sepsis and septic shock and has been shown to be effective and safe; however, there is no adequate, direct evidence or high-quality RCT studies to prove these claims, and the necessary studies to elucidate the mechanism of de-escalation therapy have not been conducted [26]. Although there is no consensus when the ridge line of early and late stage of sepsis is, it is acknowledged that the pathogen confirmed in 24-72h after sepsis diagnose. Thus, in our study we defined that 3 days were the early stage in according to the unknown pathogen period and discover whether the initial empiric antibiotic act differently. We confirmed that although it cannot improve the 6-day mortality of CLP mice in our study, de-escalation therapy can reduce organ injury and inflammatory responses, which is consistent with observations in human patients. According to the guidelines of sepsis treatment, antibiotics were administered 1 h after the CLP procedure and de-escalated. Escalation therapy was used under the guidance of empirical antibiotic usage in our study. We showed better reductions in the AST, ALT, serum creatine, and cytokine levels (IL-6, IL-10, TNF-α, IFN-γ, MIP-2 and MCP-1) in both the early and late stages of sepsis.
Based on the recovery of the function of β-lactam antibiotics in the regulation of NETs, we then explored the mechanism of de-escalation therapy. In our study, we found that the combination of different antibiotics affects the serum levels of MPO-DNA, which is a biomarker of NETs. With the sequential treatment with NET-promoting and NET-inhibiting antibiotics, NET formation could be regulated, ultimately maximizing the function of NETs and reducing their damage. In the experiment with imipenem-ceftriaxone de-escalation therapy, NET formation was higher in the early stage and lower in the late stage, which was accompanied by milder organ injury and inflammatory responses. In addition, the administration of DNase I with de-escalation therapy reduced the formation of NETs, but organ injury and inflammatory responses were exacerbated. To confirm the role of NETs, we conducted histological and immunohistochemical staining, and with greater infiltration of citH3, which is a credible NET biomarker, increased injury to organs and enhanced apoptosis of tissue cells were observed. It seems, therefore, that in addition to the useful antibacterial function of antibiotics, the function of de-escalation therapy in the regulation of NETs is more important in sepsis; this regulatory role enhances the antibacterial function of innate immunity and prevents further NET-associated damage in sepsis. Although there was no significant difference in survival between the DE group and the ES group due to sample size. Meanwhile, several therapeutic strategies such as anti-infection, nutrition and organ function support are employed in clinical treatment of sepsis, and antibiotic treatment alone does not reflect the overall picture of sepsis treatment. However, organ damage and the levels of inflammatory cytokines in the DE group were significantly decreased compared to the ES group. These results show the benefits of de-escalation antibiotic therapy strategies in sepsis.
However, there are some limitations. Polymicrobial sepsis model was used and we did not examine the bacterial loads and differentiate bacterial infiltration. It is important to detect the antibacterial spectrum, which takes about 24-72h to confirm in clinical practice. The initial administration of different antibiotics may shift bacterial colonization with a certain resistance to NET-dependent killing and influence the real outcome. Besides, bacterial colonization changes may also affect homeostasis in intestine which contribute to the injury of intestine. Those speculative assumption need to be addressed in future studies. Lastly, we assessed the 6-day mortality in the DE group which was shorter compared with the real clinical setting. Extended observation time may be expected.