Nerve stimulation promotes resolution of inflammation
Authors: April S. Caravaca Alessandro L. Gallina Laura Tarnawski Vladimir S. Shavva Romain A. Colas Jesmond Dalli Stephen G. Malin Henrik Hult Hildur Arnardottir Peder S. Olofsson
Inflammation is essential for an effective antimicrobial defense and response to tissue injury, but it must be well regulated and ultimately resolved to avoid excessive tissue damage and development of inflammatory disease. An important mechanism to restore homeostasis is the active resolution of inflammation. Defective resolution may result in nonresolving inflammation, which underlies many chronic inflammatory diseases (1, 2). Key processes in the resolution of inflammation include the cessation of neutrophil infiltration and the promotion of macrophage reparative functions, including clearance of apoptotic cells from affected tissues by efferocytosis (3, 4). However, the mechanisms that regulate resolution of inflammation are not yet fully understood, and specific therapeutic options to actively promote inflammation resolution are lacking (5, 6).It is becoming increasingly clear that the onset and intensity of inflammation are regulated by homeostatic neural reflexes, such as the “inflammatory reflex,” in which the vagus nerve plays a key role (7, 8).
In this study, we found that VNS accelerated resolution of inflammation. Activation of the vagus nerve promoted efferocytosis and shifted the LM balance toward a more proresolving profile in wild-type mice with peritonitis. VNS-mediated acceleration of resolution was impaired in mice genetically deficient in either Alox15, a key enzyme in LM biosynthesis, or the ACh receptor subunit α7nAChR, which is commonly found on immune cells. Together, these observations demonstrate that the vagus nerve controls resolution of inflammation by cholinergic regulation of the inflammation microenvironment and SPM biosynthesis.Inflammation resolution is characterized by cessation of neutrophil infiltration and increased efferocytosis (36). Studies pioneered by the laboratory of Serhan and coworkers (22, 29) showed that mice that were vagotomized 7 d before initiation of peritonitis had higher peak neutrophil numbers and delayed resolution of inflammation compared with vagus nerve–intact mice, supporting a role for the vagus nerve in the regulation of peritoneal inflammation. In the present study, we observed that activation of the vagus nerve by electrical stimulation shortened resolution time compared with in sham-treated mice in zymosan-induced peritonitis. Importantly, this study shows that activation of the vagus nerve accelerated the resolution phase of inflammation and exerted cholinergic control of SPM biosynthesis and efferocytosis.VNS enhanced the production of specific DHA- and n-3 DPA–derived SPMs in peritoneal exudates during inflammation resolution in vivo. In particular, VNS enhanced biosynthesis of the RvD, PD, and MaR families, as demonstrated through identification of RvD2,17R-PD1, and MaR2 as well as their pathway markers 10S,17S-diHDHA (PDX) and 7S,14S-diHDHA, respectively, that are biosynthesized via double lipoxygenation (37). While many mechanistic details of SPM biology, such as their precise mode of action in different inflammatory contexts and the cognate receptors for some of the compounds, at present are incompletely defined, the activity of several compounds (for example, MaR1, RvD2, RvD3, RvD5, RvE1, PD1, and LxA4) was evidenced by the effect on resolution when administered in experimental inflammatory conditions (28, 30, 31, 38). The VNS-dependent regulation of Alox15-dependent LMs reported here indicates that activation of signals in the vagus nerve can shift the balance of LMs during inflammation toward a more proresolving state.
This notion is evidenced by the up-regulation of MaR2, RvD2, 17R-PD1, and 10S,17S-diHDHA in the vagus nerve–stimulated mice. As part of their key characteristics, SPMs, including MaR2, RvD2, and PD1, are potent regulators of macrophage phagocytosis and neutrophil trafficking to sites of inflammation (39–41). PD1, also termed neuroprotection D1, is neuroprotective (37, 40), and MaR1, in addition to its potent proresolving actions with immune cells, is also neuroprotective and enhances recovery from spinal cord injury (42). Furthermore, RvD2 and 17R-RvD3, which accelerate resolution of inflammation (42), were also associated with the separation of the VNS and sham clusters on the PLS-DA plot (Fig. 2A). These observations also infer that VNS accelerates inflammation resolution and enhances efferocytosis by activating the Alox15 biosynthetic pathways, a key enzyme in SPM biosynthesis (37). Of note, cyclooxygenase-derived PG and TX were not significantly altered by VNS in wild-type mice, indicating that the vagus nerve may specifically regulate lipoxygenase but not cyclooxygenase biosynthetic pathways, which resulted in an overall proresolving LM profile in exudates from wild-type mice. An imbalance in SPM to proinflammatory eicosanoids (PG, TX, and LT) has been connected with chronic disease progression, including impaired efferocytosis associated with advanced atherosclerotic lesions (2).
Additionally, the ratio of SPM to proinflammatory LT or PG is being proposed as a potential biomarker or measure of nonresolving cardiovascular inflammation (43–45). Taken together, these results show that VNS can increase SPM production and accelerate resolution of inflammation in vivo, providing a potential means to activate resolution in chronic inflammatory conditions.VNS is known to regulate the release of proinflammatory cytokines in early stages of inflammation through an α7nAChR-dependent mechanism (9, 11, 13, 14), and activation of α7nAChR attenuates inflammation in experimental models of chronic disease (46–48). In the present study, the α7nAChR was also required for the VNS-mediated effect on resolution of inflammation in peritonitis (Fig. 4) (13, 15). As expected from studies of acute inflammation, levels of IL-1β and IL-6 were decreased in exudates of VNS-treated wild-type mice compared with sham 12 h after zymosan-induced peritonitis. We did not observe any significant effect by VNS treatment on TNF levels at 12 h after zymosan administration, likely because TNF levels are known to peak early in this model and are considerably reduced toward baseline at 12 h (49, 50).
Furthermore, there were no significant differences in the levels of measured cytokines 24 h after zymosan injection. The lack of significant difference between VNS- and sham-treated wild-type mice on levels of the neutrophil chemoattractant, CXCL1, suggests that neutrophil egress and efferocytosis play a role in the reduction of resolution time (Ri) and the increase in inflammation decay (Id) by VNS. These observations together support the notion that SPMs that rely on biosynthesis by Alox15 are important mediators of resolution of inflammation regulated by cholinergic signals elicited by the vagus nerve in this model.ACh is a key neurotransmitter in the inflammatory reflex, and VNS increases ACh in the spleen (7, 13, 14). ACh also up-regulates the Alox15 biosynthetic pathway in both human and mouse cells (29). Congruently, ACh up-regulates the RvD and PD biosynthetic pathway marker 17-HDHA in innate lymphoid cells group 3 (29). Interestingly, the peritoneum in rodents is innervated by the vagus nerve (51), and peritoneal ACh levels are reduced in peritonitis after vagotomy (22).
It is possible that peritoneal ACh levels regulated by the vagus nerve control resolution of peritoneal inflammation through activation of α7nAChR on innate immune cells (11, 13, 15). Our findings here (that the VNS treatment on α7nAChR-deficient mice failed to reduce neutrophil numbers and levels of proinflammatory cytokines during the resolution phase of zymosan-induced peritonitis) further support that cholinergic signaling is important for the effective resolution of inflammation and depends, at least partly, on the α7nAChR. Of note, injection of a set of synthetic SPMs into α7nAChR-deficient mice reduced neutrophil numbers in peritoneal exudates after zymosan injection, demonstrating that while the α7nAChR is essential for neural regulation of SPM biosynthesis, it is not required for the effect of SPMs.The use of mathematical modeling of the dynamic changes in local neutrophil numbers here facilitated analysis of the inflammation resolution phase. Our analysis of data from previous studies of neutrophil counts over the course of an episode of peritonitis revealed that the change in neutrophil counts over time showed exponential decay (SI Appendix, Fig. S3). In addition to conventional determination of resolution indices (SI Appendix, Fig. S2B), the model of exponential decay permits calculation of key indices of inflammation resolution without precise knowledge of the peak time or peak numbers of neutrophil infiltration, points that are challenging to experimentally determine and compare. Accordingly, we add to the indices of resolution the rate of “inflammation decay,” Id, as a measure of the speed with which inflammation resolves.
Three genetically different mouse models were used in this study. A limitation is that the genetic modifications in question have important effects on the immune system. Alox15-deficient mice are known to show lower intensity of the immune response, particularly neutrophil infiltration (52, 53), precluding direct comparisons of the absolute values of neutrophil numbers with the other two models. The α7nAChR-deficient mice have impaired regulation of inflammation with significantly increased cytokine release in proinflammatory events (13), and in the long-term, α7nAChR deficiency may increase the propensity to develop inflammatory diseases. For example, ablation of α7nAChR in hematopoietic cells promotes vascular inflammation and atherosclerosis (54). Therefore, it must be considered that the baseline status of the immune systems in the three models is dissimilar and that direct comparisons between the genotypes are challenging.
While there was no significant difference in the inflammation decay between VNS- and sham-treated Alox15-deficient mice, we did note a difference in the mean numbers of peritoneal exudate neutrophils, with lower numbers and higher mean Id in the VNS-treated group than in the sham, but the observations were not statistically significant. Of note, VNS treatment did have a significant effect on levels of several non–Alox15-dependent LMs in both Alox15-deficient and α7nAChR-deficient mice, although the relative effect on LMs was opposite between the strains, the mechanism of which is yet not known. It is conceivable that the effects by VNS on resolution of inflammation are not restricted to regulation of Alox15-dependent mediators. In support of this notion, the relative differences in the mean numbers of peritoneal exudate neutrophils were much smaller between VNS- and sham-treated animals in the α7nAChR-deficient mice, and there was less separation between VNS- and sham-treated animals by PLS-DA in α7nAChR-deficient mice compared with both Alox15-deficient and wild-type mice. The observations in the three different mouse strains together suggest that the α7nAChR is essential for VNS-mediated regulation of the inflammation decay in this model, and while likely not alone responsible for the entire effect, Alox15 is a component of this regulation.Together, these observations demonstrate that VNS promotes resolution of inflammation in vivo through a mechanism that involves SPM biosynthesis and requires cholinergic signaling. The findings encourage further exploration of neural regulation in resolution of inflammation.
Experimental studies over the past decades have demonstrated that vagus nerve stimulation (VNS) reduces the release of proinflammatory cytokines in acute inflammation (9–11). Rodents subjected to minutes-long VNS treatment show reduced proinflammatory cytokine release for ≥24 h (10–12); activation of this cholinergic anti-inflammatory pathway attenuates inflammation in experimental models of inflammatory diseases (7, 13–15), and VNS reduced levels of proinflammatory cytokines in clinical settings (16, 17). Consistent with these findings, 3 d after disruption of vagus nerve signaling by vagotomy, release of proinflammatory cytokines is increased in inflammation (18). Hence, animals subjected to prolonged disruption of vagus nerve signals by vagotomy respond more strongly to proinflammatory stimuli and show increased propensity for inflammation and higher inflammation-associated mortality (19, 20). This may potentially be explained by the loss of the homeostatic inhibition of inflammation that is normally conferred by the intact vagus nerve (7, 21). Interestingly, 1 wk after disruption of vagus nerve signaling by unilateral surgical vagotomy, peak numbers of neutrophils in zymosan-induced peritonitis are higher and the duration of neutrophil clearance is longer than in sham-treated mice (22). Consistent with these observations, levels of specialized proresolving mediators (SPMs), a class of potent bioactive lipid mediators (LMs) that actively regulate inflammation resolution, are significantly lower in vagotomized mice compared with sham (22). While it has been recently reported that stimulation of human and mouse vagus nerve ex vivo can enhance SPM production from the vagus nerve tissue itself (23), whether activation of the vagus nerve promotes resolution of inflammation and regulates proresolution mechanisms in inflammation in vivo remains to be determined. Here, we postulated that electrical VNS promotes resolution of inflammation in vivo.