Abstract
Inflammatory resolution is a process that, when uncontrolled, impacts many organs and diseases. As an active, self-limited inflammatory process, resolution involves biosynthesis of specialized proresolving mediators (SPM) (e.g., lipoxins, resolvins [Rv], protectins, and maresins). Because vagal stimulation impacts inflammation, we examined human and mouse vagus ex vivo to determine if they produce lipid mediators. Using targeted lipid mediator metabololipidomics, we identified lipoxins, Rv, and protectins produced by both human and mouse vagus as well as PGs and leukotrienes. Human vagus produced SPM (e.g., RvE1, NPD1/PD1, MaR1, RvD5, and LXA4) on stimulation that differed from mouse (RvD3, RvD6, and RvE3), demonstrating species-selective SPM. Electrical vagus stimulation increased SPM in both human and mouse vagus as did incubations with Escherichia coli. Electrical vagus stimulation increased SPM and decreased PGs and leukotrienes. These results provide direct evidence for vagus SPM and eicosanoids. Moreover, they suggest that this vagus SPM circuit contributes to a new proresolving vagal reflex.
Introduction
The acute inflammatory response is critical in host defense and, when unresolved, can lead to chronic inflammation associated with many human diseases (1, 2). New therapeutic approaches are needed for diseases in which unresolved inflammation contributes to progressive loss of organ function. The vagus nerve–based inflammatory reflex uncovered by Tracey and colleagues (3) regulates immune function and inflammation. One mechanism of neural–immune control involves activation of macrophage α7 nicotinic acetylcholine receptors that inhibit proinflammatory cytokines. This macrophage α7 receptor inhibits NF-κB nuclear translocation and stimulates the JAK2/STAT3 pathway to reduce cytokines (4).
Mechanisms controlling the magnitude and duration of inflammatory responses have recently attracted considerable attention (1, 2). Self-limited acute inflammatory responses activate biosynthesis of novel specialized proresolving mediators (SPM) that stimulate resolution. SPM function by 1) limiting further neutrophil infiltration, 2) reducing collateral tissue damage, and 3) activating macrophages to engulf apoptotic cells and debris as well as 4) clearing microbial infections (2). The SPM include lipoxin (LX), resolvin (Rv), protectin (PD), and maresin (MaR) families biosynthesized from essential polyunsaturated fatty acids. Each SPM family member also counter-regulates cytokines, chemokines, and proinflammatory eicosanoids (e.g., PGF2α and leukotrienes [LT]) to reduce inflammation and activate IL-10 (2). Rv also block macrophage NLRP3 inflammasome, reducing IL-1β (5), and reduce pain (6, 7). Recently, new SPM structures containing peptide conjugates were elucidated that stimulate resolution and activate tissue regeneration (8).
We found that vagotomy delays resolution of inflammation (9). This delay involves shifting lipid mediators (LM) with reduced Rv to proinflammatory status, demonstrating a novel vagus-resolution circuit (9, 10). During bacterial infection, vagus also controls resolution via biosynthesis of specific SPM that function as immunoresolvents (e.g., PD conjugate in tissue regeneration [PCTR]1) upregulated by acetylcholine via ILC-3 control of macrophage SPM biosynthesis and phenotype (10).
In view of these findings, we investigated whether vagus can directly produce LM. In this article, we report that human vagus produces specific SPM, identified using liquid chromatography–tandem mass spectrometry (LC-MS/MS)–based metabololipidomics, that differed from those produced by mouse vagus. Escherichia coli increased LM-SPM, and electrical vagus stimulation (EVS) ex vivo increased SPM and reduced both PGs and LT.
Materials and Methods
Human and mouse tissues
Fresh human vagi (deidentified) purchased from Tissue for Research (Ellingham, Bungay, Suffolk, U.K.) were analyzed under protocol no. 1999P0001279 approved by the Partners Human Research Committee. Each postmortem, full-length human vagus was thawed on arrival, measured, dissected, and incubated in PBS (with calcium and magnesium) for 20 min at 37°C with 5% CO2 in parallel with direct EVS with 2.5 mA 18 V direct current (DC) for 20 min in PBS at 37°C (ApeX Type A stimulator; ApeX Electronics, Schenectady, NY), or coincubated with E. coli (109 CFU for 3 h at 37°C). Deuterium-labeled standards for SPM and eicosanoid extraction recoveries were from Cayman Chemical (Ann Arbor, MI). For abbreviations and stereochemical assignments with the full name for each of the SPM, see (11, 12). Animal experimental procedures were approved by the Institutional Animal Care and Use Committee of Brigham and Women’s Hospital (protocol no. 2016N000145) and complied with institutional and U.S. National Institutes of Health guidelines. Six- to eight-week-old FVB male mice (Charles River Laboratories, Wilmington, MA) were fed ad libitum Laboratory Rodent Diet 20-5058 (Purina Mills, Great Summit, MO).
LM metabololipidomics
Cold methanol containing deuterium-labeled (12) internal standards (500 pg/sample) were added to all samples. Following solid-phase extraction, LM-SPM were identified and quantified as in Ref. 13 and, for cysteinyl LT (CysLT), using LC-MS/MS and published criteria (i.e., six ions) (12, 13). Linear calibration curves were obtained using d5-LTC4, d5-LTD4, d2-PCTR3, and others (12), giving r2 values of 0.98–0.99.
Statistics
Results are mean ± SEM. Significance was calculated using one-tailed paired t test and GraphPad Prism software (La Jolla, CA). The p values were *p < 0.05 and **p < 0.01.
Results and Discussion
Human vagus produces endogenous SPM and eicosanoids
To determine if human vagus directly produces LM that could impact inflammation via the neural reflex (3), we assessed LM profiles with fresh human vagus. To this end, using LC-MS/MS–based LM metabololipidomics together with spectral libraries of MS/MS (11–13), we identified in human vagus specific mediators from each major bioactive LM-SPM metabolome. (Fig. 1A, Supplemental Table I). These included Rv, PD, and MaR from DHA, E-series Rv from EPA, arachidonic acid–derived LX, LT, thromboxane, and PGs, as well as CysLT (LTC4, LTD4). For each, LC-MS/MS results gave at least six diagnostic ions for identification (Fig. 1B).
Human vagus produced several Rv, including RvE1 and specifically RvD3, RvD4, and RvD5 (Fig. 1B). The Rv of human vagus did not include RvD1, RvD2, RvE2, or RvE3, which are produced by human leukocytes, lymph nodes, spleen (11), and emotional tears (13). These results indicate that, although some tissues produce all of the known d-series Rv (RvD1–RvD6), human vagus produces those biosynthesized via the 4(5)-epoxy-Rv intermediate rather than those from 7(8)-epoxy-Rv intermediate (i.e., RvD1 and RvD2) (compare Refs. 11, 13). D-series Rv control inflammation resolution, infection, and pain reduction (2, 8).
Human vagus also produced both PD and MaR pathways. This was concluded through identification of neuroprotectin D1 (NPD1/PD1) and its pathway marker (Fig. 1), biosynthesized via double lipoxygenation, namely 10S,17S-diHDHA (PDX) (14). Also, 17R-NPD1/PD1 was identified in human vagus (Fig. 1A). NPD1/PD1 stimulates resolution and is neuroprotective (15). This 17R epimer of NPD1/PD1 is longer acting and is produced via acetylated COX-2 following aspirin or by p450, which can produce the precursor 17R-hydroxydocosahexaenoic acid (16). Hence, 17R-NPD1/PD1 may have resulted from aspirin use by the organ donors. Alternately, aspirin-triggered Rv (17R epimer) and LX (15R epimer) are also produced by a new pathway in neural tissues that uses sphingosine kinase 1 to acetylate COX-2 as a mechanism to biosynthesize aspirin-triggered epimers of SPM (17). These longer-acting endogenous epimers of SPM are potent proresolving agonists (2). Human vagus also produced MaR1 and its pathway marker, 7S,14S-dihydroxy-DHA (Fig. 1). In addition to MaR1’s potent proresolving actions with human leukocytes (2, 14) and platelets (18), MaR1 is neuroprotective and activates recovery from spinal cord injury (19).
In human vagus, SPM from arachidonic acid (i.e., LXA4 and LXB4) were also identified (Fig. 1). Along with their ability to activate resolution (2), LXA4 reduces neuroinflammation and neuropathic pain following hemisection of spinal cord via reducing microglial activation (20), and both LXA4 and LXB4 are neuroprotective (21). Thus, their production by human vagus, as well as other SPM documented in this article from their physical properties, is of interest as potential mediators from vagal stimulation.
Electrical stimulation of human vagus increased RvD4 and MaR1, with trends for increases in other vagus SPM (Supplemental Table I). RvD4 is found in human bone marrow and controls bacterial clearance (22). Vagus expresses Toll receptors (3), and incubations with live E. coli increased both RvD4 and RvD6, as well as increased 15-epi-LXA4 and MaR1, which may together stimulate clearance of infections. RvD6 was not present in vagus alone or with electrical stimulation (Supplemental Table I). Fig. 1C shows the vagus LM network, depicting quantification, biosynthetic relationships between precursors, bioactive LM, and pathway marker products of each bioactive metabolome.
Human vagus produces a distinct and unique profile of SPM and eicosanoids
Because specific SPM were present in human vagus, we investigated LM of mouse vagus. For this, fresh mouse vagi were incubated, which demonstrated that LM profiles in mice differed from profiles in humans (Supplemental Fig. 1, Supplemental Tables I and II). Three mouse strains produced the same SPM (Supplemental Table III). Mouse vagus produced RvD4, RvE1, RvE3, LXB4, and 15-epi-LXA4. Mouse vagus with E. coli increased biosynthesis of only PDX, suggesting that this SPM may play a role in vagus control of infection, whereas human vagus increased several SPM (e.g., RvD4, NPD1, MaR1, 18-HEPE, and 15-epi-LXA4) that are each potent proresolving mediators. Interestingly, RvD3 was selectively increased with EVS (vide infra). Multivariate analysis of LM profiles obtained from human or mouse vagus profiles demonstrated a strong association between different species (Supplemental Fig. 1D); the sphere in the three-dimensional score plot represents 95% confidence. Principal component analysis (PCA) confirmed that RvD6, RvE3, and RvD4 were associated with mouse vagus, whereas RvD5, RvE1, MaR1, and NPD1 were associated with human vagus (Supplemental Fig. 1D).
Electrical stimulation enhances vagus production of SPM and reduces eicosanoids
We next investigated whether EVS ex vivo also led to LM production. After 20 min of electrical stimulation, we found a specific group of SPM was increased. PCA confirmed that mouse vagus nerve subjected to EVS clustered separately compared with control (Fig. 2A). In multivariate analysis, RvD4, RvE1, RvD3, and PDX were associated with EVS (Fig. 2B). Also, quantitation of the increase in SPM gave a statistically significant increase of ∼3× the sum of RvD3, RvD4, and RvE1. Interestingly, prostanoids and thromboxane were reduced by EVS (Fig. 2C), as were LTC4, LTD4, and LTE4 (Fig. 2D, Supplemental Table II). These findings identify LM of human and mouse vagus as well as, to our knowledge, the first evidence of vagus SPM production. Together, the present findings identify SPM as vagal products that are known controllers of host response to inflammation and infection (2, 5, 14).
Vagus from human and mouse also produces PGD2, PGE2, and PGF2α (Figs. 1, 2) as well as LT. LTB4 is a potent chemoattractant, and CysLT (LTC4, LTD4, and LTE4) are appreciated for their production by mast cells and role as slow-reacting substance of anaphylaxis in allergic reactions (23). However, CysLT may also possess physiologic functions in neural and endocrine systems, as in pineal gland control of hormone release (23). Because CysLT are potent smooth-muscle constrictors and stimulate vascular permeability (23), their vagus production is of interest and may contribute to neural reflex pathways that can modulate organ function. Novel SPM such as PCTR1, regulated by vagal stimulation of ILC3 to control infection (10), along with maresin conjugates in tissue regeneration and resolvin conjugates in tissue regeneration (12), were not present in either mouse or human vagus, in contrast to LTC4, LTD4, and LTE4. EVS of mouse vagus increased SPM that included LXB4, RvE1, RvD3, and RvD4 (Fig. 2A–C). This was accompanied by decreases in both PGs and CysLT (Fig. 2C–E). These findings indicate that vagus stimulation increases proresolving mediators that can directly stimulate resolution of inflammation and infections by virtue of their actions on phagocytes and reduce chemokines, cytokines, and proinflammatory LM as well as enhance microbial killing and clearance (2). Also, Rv (e.g., RvE1) reduce pain via SPM receptors on neurons (7).
In PGE synthase-1 (mPGE1) knockout mice, vagus stimulation is abolished, implying that absence of PGE2 is critical to the cholinergic anti-inflammatory pathway (24). In resolution of contained exudates, PGE2 signals LM class switching, increasing SPM (2). Vagus nerve also responds with cytokine-specific neural signals (25) that can contribute to systemic inflammation. Additional regulators of inflammation resolution that possibly may be vagus controlled include hypoxia-inducible factors, purinergic signaling, and miRNAs (26–28), which interact with SPM (2). Vagus-stimulating devices in arthritis patients target the inflammatory reflex, reducing TNF-α, IL-1β, and IL-6 (29).
Our results demonstrate that isolated human vagi produce specific SPM, suggesting that EVS may activate resolution of inflammation via SPM and downregulation of PGs and LT. Excess PG and LTB4 are known to contribute to chronic inflammation (23). Network mapping in the immune system (Figs. 1, 2) can highlight species differences in physiologic and pathologic networks (30). The present results demonstrate species differences with human SPM, in that, with EVS, human vagus produced MaR1 and RvD4 (Fig. 1, Supplemental Table I). RvD4 is produced by both human and mouse vagus, suggesting proresolving functions are intact in both species. Hence, these results document vagus proresolving capacity, with human and mouse vagus directly producing LX, Rv, and PD in amounts commensurate with their potent pico- to nanogram actions (2) that can impact multiple organs and immune cells. They also demonstrate that EVS increases SPM and diminishes PGs and LT that are known to contribute to chronic inflammation and allergic responses (23). Together, these results, to our knowledge, identify a new vagus proresolving reflex that may be targeted via electrical stimulation to improve disease treatments in which Rv and unresolved inflammation are involved as well as to possibly improve overall health status.
Acknowledgements
We thank Mary Halm Small for expert assistance in manuscript preparation.
Footnotes
This work was supported in part by National Institutes of Health Grant R01GM038765 (to C.N.S.).
The online version of this article contains supplemental material.
Abbreviations used in this article:
- CysLT
cysteinyl LT
- DC
direct current
- EVS
electrical vagus stimulation
- LC-MS/MS
liquid chromatography–tandem mass spectrometry
- LM
lipid mediator
- LT
leukotriene
- LX
lipoxin
- MaR
maresin
- NPD1/PD1
neuroprotectin D1
- PCA
principal component analysis
- PCTR
PD conjugate in tissue regeneration
- PD
protectin
- PDX
10S,17S-diHDHA
- Rv
resolvin
- SPM
specialized proresolving mediator.
References
Disclosures
The authors have no financial conflicts of interest.