Cholinergic anti-inflammatory pathway (CAP) bridges immune and nervous systems and plays pleiotropic roles in modulating inflammation in animal models by targeting different immune, proinflammatory, epithelial, endothelial, stem, and progenitor cells and signaling pathways. Acute lung injury (ALI) is a devastating inflammatory disease. It is pathogenically heterogeneous and involves many cells and signaling pathways. Here, we emphasized the research regarding the modulatory effects of CAP on animal models, cell population, and signaling pathways that involved in the pathogenesis of ALI. By comparing the differential effects of CAP on systemic and pulmonary inflammation, we postulated that a pulmonary parasympathetic inflammatory reflex is formed to sense and respond to pathogens in the lung. Work targeting the formation and function of pulmonary parasympathetic inflammatory reflex would extend our understanding of how vagus nerve senses, recognizes, and fights with pathogens and inflammatory responses.

1. Introduction

From 2000, Tracey and colleagues have been working on the mechanisms by which electric stimulation of vagus nerve suppresses activation of NF-κB and production of proinflammatory cytokines in α7 nicotinic acetylcholine receptor- (α7 nAChR-) expressing macrophages and lessens severity of sepsis in animal models [1, 2]. These findings led to establishment of a novel theory of cholinergic anti-inflammatory pathway (CAP) [1]. Most experiments regarding the modulatory effects of CAP on inflammation were tested in the models of sepsis, a syndrome of systemic proinflammatory responses. The experimental results support that spleen is the functional hub of CAP. In 2011, Tracey and colleagues found that innervation of vagus nerve in the spleen, CHAT-expressing T lymphocytes, and α7 nAChR-expressing macrophages forms a neural circuit to finely tune the proinflammatory responses [3]. In this review, we summarized the progress regarding the modulatory effects of CAP on inflammation and pointed out the future research directions towards brain center of CAP, activation of β2 adrenergic receptor, synthesis of acetylcholine in the T lymphocytes, and others.

ALI is a devastating inflammatory disease [4]. It is pathogenically heterogeneous and involves many cells and signaling pathways. From 2007, the modulatory effects of CAP have been tested in a variety of animal models with ALI [5]. Activation of CAP also affects many types of cells and signaling pathways involved in ALI. In this review, we compared the deferential pulmonary inflammatory responses during sepsis (systemic) and ALI (local). Considering different features of modulatory effects of CAP on pulmonary inflammatory responses, we put forward a new working model, pulmonary parasympathetic inflammatory reflex, to extrapolate how vagus nerve through α7 nAChR modulates acute lung infection, inflammation, and injury. The pulmonary parasympathetic inflammatory reflex works locally and may not require spleen. In accordance with this assumption, vagus nerve coupling with α7 nAChR-expressing resident macrophages also modulates intestinal inflammation independent of spleen [6]. Thus, future studying local modulatory effects of CAP on inflammation may be an emerging avenue to explore how vagus nerve senses, recognizes, and responds to pathogens.

2. Cholinergic Anti-Inflammatory Pathway

2.1. Inflammatory Reflex and CAP

Studies suggest that afferent and efferent vagus nerves, α7 nAChR-expressing inflammatory cells, and central vagal nucleus in the brain form an inflammatory reflex that could finely tune inflammation and immunity [2, 7]. Strictly speaking, CAP is the efferent arm of vagal inflammatory reflex and spleen may be the anti-inflammatory hub in this neural circuit [810]. Activation of this pathway provides the host with a fast, discrete, and localized means of modulating the inflammatory and immune responses in variety of animal models [1113].

2.2. The Role of Spleen in Inflammatory Reflex

As Figure 1 shows, in the vagal inflammatory reflex, the sensory neurons may sense the changes of pathogen associated molecular patterns (PAMPs) or damage associated molecule patterns (DAMPs) in the peripheral afferent vagal nerve endings and then feedback to nucleus tractus solitarii (NTS) in the brain stem. After the information is processed in the NTS, the efferent vagus nerve transmits integrated information by action potentials to the celiac ganglion and then delivers in the spleen. Anatomically, the splenic vagus nerve endings are closely in contact with a group of β2 adrenergic receptor- (β2 AR-) expressing T memory lymphocytes (CD4+CDCD62) and release norepinephrine (NE), a sympathetic neurotransmitter. NE activates β2 AR in the T lymphocytes, initiates transcription of choline acetyltransferase (ChAT), and synthesizes acetylcholine (ACh). ACh could activate splenic α7 nAChR-expressing macrophages, inhibit NF-κB activity, promote STAT3 phosphorylation [14], and therefore dampen proinflammatory cytokine production (especially TNF-α and HMGB1) [2, 3, 12, 15, 16].

We have to point out that Figure 1 is a hypothetical model that shows a direct connection of the efferent vagus nerve to the spleen via the celiac ganglion [3]. However, a recent finding has demonstrated there was no neural connection from the vagus to splenic sympathetic by neuroanatomical tract tracing and neurophysiological measurements [17]. Moreover, one study has showed that sympathetic nerves rather than vagi contribute to anti-inflammatory effects revealed by a LPS-challenged splanchnic-nerve or vagus nerve-cut rat model to compare changes of splanchnic sympathetic nerve activity and peripheral blood TNF-α [18, 19]. But this study has its limitations, for example, short experimentation (1-2 h), small sample size, and unknown mechanisms. More recently, Torres-Rosas et al. have found that electroacupuncture controls systemic inflammation in sepsis via the sciatic and vagus nerves and catecholamines from the adrenal glands [20]. Therefore, the controversy regarding anti-inflammatory roles of vagus and sympathetic nerves should be sorted out in the future study.

2.3. Information-Integrating Center of Inflammatory Reflex

It should be mentioned that vagus nerve originates from medullar oblongata, which consists of four nuclei: dorsal nucleus, nucleus ambiguous, NTS, and spinal nucleus of trigeminal nerve [21, 22]. About 80% afferent sensory fibers are contained in the vagus nerve and responsible for transmission of the information to the NTS [21]. For example, animals were intravenously or intraperitoneally challenged with LPS or IL-1β could induce c-fos expression in the nodose ganglia and the NTS [23, 24]. NTS also plays a very important role in projecting information to the nuclei (including the locus ceruleus and dorsal raphe nuclei) of the brain [21]. It is unknown whether and how PAMPs or DAMPS can be recognized by the afferent vagus nerve endings in the lung and how NTS processes the information that is collected from the lung during infection and inflammation.

2.4. Transcription of ChAT in β2 AR-Expressing T Lymphocytes

Upon the vagus nerve stimulation, transcription of ChAT gene in the splenic β2 AR-expressing T lymphocytes might be regulated by cAMP, which is a major second messenger following activation of β2 AR [25, 26]. However, study has also shown that stimulation of efferent vagus nerve induces plasma norepinephrine via the α7 nAChR in a mouse model [27]. This finding raises a possibility that transcription of ChAT and biosynthesis of ACh in the splenic lymphocytes are positively regulated by β2 AR or α7 nAChR. Experimental data have demonstrated that β2 AR-expressing CD3 lymphocytes and α7 nAChR-expressing CD11b/c monocyte or macrophages are present in the spleen. The defects in response to β2 AR or α7 nAChR stimulation or quantity in the spleen lead to the dysfunction of inflammation resolution and postoperative cognition decline [28]. In addition, the β2 agonist is well recognized for its anti-inflammatory property for ALI [29, 30]. Whether this protective effect of β2 agonist on ALI [31] is via activation of splenic β2 AR-expressing T lymphocytes or CAP requires to be investigated.

2.5. Synthesis of Acetylcholine in Nonneuronal Cells

It needs to emphasize the important role of high-affinity choline transporter (CHT1) or choline transporter like proteins (CTLs) in the process of ACh synthesis in the nonneuronal cells (e.g., lymphocytes and lung cells) [3234]. ACh is synthesized from choline and acetyl-CoA by the enzyme choline acetyltransferase (ChAT), and this event may be limited by choline availability [35]. In neurons, loss of CHT-mediated presynaptic choline uptake might result in neonatal lethality [36]. ChAT contains nuclear localization signals and is also localized in the nuclei of neural and nonneuronal cells [37]. Enzymatic activity and nuclear translocation of ChAT are required for its transcriptional enhancement of CHT gene [37]. Pulmonary nonneuronal cholinergic system (including ChAT-, CTLs-, VAChT-, and OCT-mRNA) is downregulated in acute allergic airway inflammation [35], suggesting that synthesis of ACh is regulated locally during inflammation.

3. Acute Lung Inflammation and Injury and Modulatory Effects of CAP

3.1. Acute Lung Inflammation and Injury

Adult respiratory distress syndrome (ARDS), characterized by ALI, has a mortality of 40% even if the patients receive advanced intensive care [4]. Pneumonia, severe sepsis, and acid aspiration are the most serious causes of ARDS [4, 38, 39]. Gram-negative sepsis derived ALI is characterized by neutrophil alveolitis and increased permeability of the lung microvascular endothelial and alveolar epithelial barriers [4042]. Aspiration of gastric contents is reported to be associated with a 26–36% incidence of ARDS [43, 44]. Aspirated hydrochloric acid may evoke direct damage to the alveolar-capillary membrane and promote adhesion, activation, and sequestration of neutrophils.

3.2. Direct and Indirect Animal Models of Acute Lung Inflammation and Injury

Alveolar epithelial cells are the main target cells in the epithelial respiratory compartment exposed to noxious substances such as E. coli or acid [45]. Injury to the alveolar epithelial barrier is a major determinant of severity of clinical ALI [46, 47]. Our experiments have demonstrated that, at the same dosage, intratracheal challenge of E. coli could induce much severe lung inflammation than intraperitoneal challenge of E. coli. As Figure 2 shows, mice were divided into three groups: control group received PBS; E. coli pneumonia group received an intratracheal challenge of E. coli (107 cfu); E. coli peritonitis ALI group was given an intraperitoneal challenge of E. coli (107 cfu). All mice were also given I125-albumin intratracheally or intravenously to measure lung wet-to-dry weight ratio and lung epithelial and endothelial permeability as previously reported [5, 48]. At 4 h after challenge, three parameters were markedly higher in the E. coli pneumonia compared to E. coli peritonitis ALI.

3.3. Category of ALI

ALI experimental models can be categorized into direct and indirect lung injury based on the route of insults. Acid-induced ALI, LPS-induced ALI, E. coli pneumonia, and other experimental ALI models were considered direct models because the injurious agents (such as HCl, bleomycin, endotoxin, E. coli, and influenza virus) were instilled into the air spaces with initial direct contact with pulmonary epithelium [5, 4951]. Ventilator-induced ALI caused by overstretch of lung epithelial cells is also considered as direct lung injury [5254]. Thiourea-induced lung vascular injury [55], oleic acid-induced ALI [56, 57], peritonitis-induced ALI (including cecal ligation and puncture (CLP)) [16], and transfusion-related ALI (TRALI) (by intravenous MHC I monoclonal antibody) [58] were considered as indirect models because the injurious agents initially interacted with the lung endothelium after intravenous challenge [55].

3.4. Different Effects of CAP on Lung Cytokines: Intratracheal versus Intravenous Insult

The modulatory effects of CAP on proinflammatory cytokines also alter when the challenge route of pathogens is different. Numerous studies have demonstrated that TNF-α is a proinflammatory cytokine and is well regulated by CAP. The spleen is identified as the source of 90% of the serum TNF during endotoxemia and in particular the marginal zone- and red pulp-macrophages of the spleen [10, 59]. Compartmentalization of TNF-α in the blood or alveolus is dependent on route of LPS challenge. For example, intravenous endotoxin significantly increases TNF-α production in the spleen by a factor of 30 as compared with six- and twofold increases in the lung and liver, respectively. Vagus nerve stimulation significantly reduces TNF levels in the spleen (94%) and liver (40%) but not in the lung (20%). However, in a lung injury model by an intratracheal challenge, compartmentalization of TNF-α in alveolus is preserved before alveolar-capillary injury [60]. Once compartmentalization of alveolar TNF-α is lost, injured lung may contribute to a systemic inflammatory response and subsequent multiorgan failure [60]. Similarly, intratracheal LPS induced a significant increase in MIP-2 in BAL fluid, whereas MIP-2 in the plasma was not detectable. In contrast, intravenous LPS caused a marked increase in plasma MIP-2, whereas only a small elevation of MIP-2 concentration in BAL fluid was observed [61]. In a LPS-induced ALI (intratracheal), administration of α7 nAChR agonists could inhibit NF-κB activity in the BAL proinflammatory cells and reduce both TNF-α and MIP-2 levels in the BAL [5]. Vagotomy and deficiency of α7 nAChR worsen lung inflammation [5, 50].

3.5. The Modulatory Effects of CAP on Animal Models of ALI Are Dependent on PRR (Table 1)

By analyzing Table 1, we can conclude that activation of CAP might affect the development of lung infection, inflammation, or injury in a PRR- (pattern recognition receptors-) dependent manner. For example, nicotine administration worsens Gram-positive bacterial pneumonia (TLR2) [50] and influenza viral pneumonia (TLR3, TLR7, or RIG-I-MAVS) [6264] but improves Gram-negative bacterial pneumonia or LPS-induced ALI (TLR4) [65]. It has to be noted that activation of α7 nAChR universally suppresses TLR2, TLR3, TLR4, or TLR9 agonist (rather than live pathogens) induced TNF-α production in monocytes [66]. These findings suggest that vagus nerve through α7 nAChR responds to PAMPs or DAMPs differently. Therefore, vagus nerve may play pleiotropic roles in modulating lung infection and inflammation.

3.6. Opposite Effects of CAP on Lung Infection and Inflammation

The discovery that splenectomy inactivates CAP strongly supports that spleen determines the function of CAP [9]. In the splenectomized animals, nicotine therapy worsens animals with lethal polymicrobial sepsis [9]. This finding suggests that once CAP is dysfunctional, activation of α7 nAChR would paradoxically compromise immunity and worsen lung infection. Traumatic brain injury or stroke might cause functional impairment of CAP and activation of α7 nAChR worsened Gram-negative bacterial pneumonia [59, 67].

4. Pulmonary Parasympathetic Inflammatory Reflex

4.1. Vagus Nerve Helps Sensing, Recognizing, and Responding to Pathogens

We have to mention that classical CAP theory was mostly tested in the experimental models of sepsis (intravenous LPS) or CLP peritonitis animal models in which spleen is required for dampening inflammation; however, these models only present mild lung inflammation which is manifested by less impressive intra-alveolar inflammation and hyaline membrane formation [68]. Different from the regulatory effects of the classical CAP on sepsis, vagus nerve might modulate lung infection and inflammation using new machinery: pulmonary parasympathetic inflammatory reflex [69], and the spleen may not be involved in this regulatory mechanism.

As illustrated in Figure 3, the pulmonary parasympathetic inflammatory reflex may consist of three components: the afferent arc residing in the distal airway or alveolus; the NTS information-integrating center in the brain stem; and the efferent arc innervating the distal lung epithelial cells. Vagus nerve endings are reported to innervate the distal airway of the lung, possibly in the alveoli [70, 71] (though it is entirely unclear how efferent fibers traveling in the vagus nerve might exert influence upon the alveolar region), where varieties of sensors or PRR in the vagal afferent arc are located. Via this apparatus, mechanical, chemical, biological, and other stimuli in the alveoli can be sensed. Sensory neurons express TLR3, 4, 7, and 9, which can recognize different pathogens [7274]. Lung neuroendocrine cells also are complex airway sensors, which are predominantly innervated by vagal afferent fibers derived from the nodose ganglion [75]. The information is transmitted via the afferent arm to NTS, a processing center, which is capable of differentiating types of infection, inflammation, or challenges. After processing, the active potentials are remitted from NTS to the alveoli via the vagal efferent arc. The vagal nerve endings could synthesize and release ACh, which in turn activates α7 nAChR in the proinflammatory cells (e.g., macrophages and neutrophils) or epithelial cells to regulate the production of proinflammatory cytokines via NF-κB or other signaling pathways.

4.2. Recruitment of α7 nAChR-Expressing Cells and Nonneural ACh in the Lung

During lung infection and inflammation, alveolar macrophages produce MIP-2, a key chemokine, which could attract neutrophil migrating into the alveoli [76]. These infiltrated neutrophils also express α7 nAChR. Apart from release of ACh from the vagal nerve endings in the distal airway, lung epithelial cells, immune cells, and neuroendocrine cells also produce nonneuronal ACh [32, 35, 77]. α7 nAChR in bronchial epithelial cells can be upregulated by the stimulation of nicotinic agonists [78, 79]. This positive feedback between acetylcholine and α7 nAChR might facilitates maintenance of concentration of acetylcholine in the alveoli.

4.3. The Role of Local CAP Is Emerging

In addition, the theory of classical CAP has also been challenged by recent researches. In a rat model, one group reported that vagal efferent neurons in the rat neither synapse with splenic sympathetic neurons nor drive their ongoing activity using vagal terminals anterograde and Fast Blue labeling technology and electrophysiological stimulation [17]. A recent study has shown that gastrointestinal CAP plays a protective role in a mouse of postoperative ileus [6]. In this study, denervation of spleen and depletion of T lymphocytes could not deactivate the protective property of vagus nerve stimulation. Anterograde labeling revealed that vagal efferents closely make contacts between cholinergic myenteric neurons and resident α7 nAChR-expressing macrophages. Therefore, the protective effects are attributable to local vagal nerve innervation and resident macrophages independent of spleen [6].

5. Modulatory Effects of CAP on Cells Involved in Lung Inflammation

5.1. Cells Involved in Acute Lung Inflammation

Macrophages (monocytes), neutrophils, mononuclear cells, epithelial cells, endothelial cells, hematopoietic stem cells, mesenchymal stem cells, endothelial progenitor cells, T lymphocytes, and fibroblasts play roles in the different phases of lung inflammation and repair [4, 5, 49, 68, 76, 80, 81].

5.2. Modulatory Effects of CAP on Different Cells

Activation of α7 nAChR could modulate inflammatory responses in variety of types of cells and affect the development of inflammatory models (Table 2).

5.3. Modulatory Effects of Activation of α7 nAChR on Inflammatory Cells May Be Dynamic

On average, 20–25% of total cells are α7 nAChR-expressing cells in the bone marrow (BM), blood, spleen, lymph nodes, and Peyer’s patches [82]. Lung α7 nAChR+Gr1 or α7 nAChR+CD11b+ granulocytes (neutrophils and monocytes) were increased to 40% after being infected with E. coli [50], suggesting that more granulocytes migrate into the lung. In addition, vagus nerve stimulation significantly attenuates CD11b+ cells in the spleen during sepsis [83]. These findings support that α7 nAChR-expressing proinflammatory cells can dynamically migrate among lung, spleen, and other organs during different stages of inflammation [84]. This dynamic movement of α7 nAChR-expressing cells might facilitate them being activated by acetylcholine released from the vagus nerve.

6. Modulatory Effects of CAP on Signaling Pathways

6.1. Signaling Pathways Involved in Acute Lung Inflammation

It was reported that activation of p38 MAPK, AKT1, and NF-κB in neutrophils contributes to ALI [85]. Lack of AKT1 could worsen acid, LPS, or bacteria induced acute lung infection and inflammation [8688]. LPS activates the STAT kinases, Src, and JAK. LPS treatment could activate STAT3 in the resident lung cells and recruited inflammatory cells [89]. In a rat model of intrapulmonary deposition of IgG immune complexes, STAT3 activation was dramatically suppressed by depletion of neutrophils or lung macrophages, resulting in reduced gene expression of IL-6 and IL-10 in whole lung tissues [90].

6.2. Modulatory Effects of Activation of α7 nAChR on Signaling Pathways

Activation of α7 nAChR in macrophages, monocytes, and other immune cells may downregulate production of proinflammatory cytokines and attenuate the inflammatory responses by several possible mechanisms: NF-κB activation, JAK-STAT3 pathway, and PI3K-AKT1 pathway. Activation of α7 nAChR by its agonists in monocytes and macrophages could reduce nuclear translocation of NF-κB and the transcription of proinflammatory cytokines [1, 2, 91, 92]. In the sepsis and lung injury mouse models, administration of α7 nAChR agonist also suppresses activation of NF-κB [5, 16]; however, one study showed that vagus nerve electrical stimulus could attenuate the proinflammatory cytokine responses in vivo but did not decrease the NF-κB activation [93]. Whether activation of α7 nAChR affects TLR4 signaling pathway, for example, MyD88, TRIF, IRAK, TARF, and other adaptor proteins in the proinflammatory cells, needs further study. The modulatory effects of activation of α7 nAChR on signaling pathways are summarized in Table 3.

6.3. Spatial and Temporal Effects of α7 nAChR Activation on p-STAT3

Anti-inflammatory effect of nicotine in murine macrophages acts through the recruitment of JAK2 to the α7 nAChR and subsequent phosphorylation of JAK2, thereby initiating the anti-inflammatory STAT3 cascade [14]. JAK2 inhibitor AG490 inhibited the anti-inflammatory effect of GTS-21 in the human PBMCs (peripheral blood mononuclear cells) [94], suggesting that p-STAT3 mediates inhibitory role of activation of α7 nAChR. However, in an endothelial cell line, GTS-21 significantly reduced STAT3 activation by phosphorylation and DNA binding [95]. In the splenocytes or myocardium tissue, cardiac troponin I (TnI) induced STAT3 activation and IL-6 is inhibited by nicotine [96]. In macrophage cell line, both α7 nAChR activation and inhibition of JAK2 blunt STAT3 phosphorylation. Inhibition of STAT3 phosphorylation mimicked the α7 nAChR signaling, inhibiting NF-κB and cytokine production in macrophages. These findings suggest the proinflammatory role of p-STAT3. In addition, unphosphorylated STAT3 might compete with NF-κB. Inhibition of STAT3 protein expression enhanced cytokine production and abrogated α7 nAChR signaling [97].

6.4. Modulatory Effects of α7 nAChR Activation Might Involve CREB and c-fos

It has been assumed that interaction between α7 nAChR and adenylate cyclase 6 increases intracellular cAMP, a secondary messenger, which in turn promotes phosphorylation of CREB. P-CREB translocates into the nucleus and initiates transcription of c-fos, an early response gene. Activation of c-fos could inhibit NF-κB activity and production of proinflammatory cytokines [11, 59]. So far, there is no scientific evidence to prove this hypothesis, but some previous findings indicate that it might be testable. For example, in epithelial cells, α7 nAChR physically binds adenylate cyclase 6 [98]. In response to LPS stimulation, Fos−/− macrophages and mice showed significantly enhanced production of TNF-α, IL-6, and IL-12 p40 but reduced production of the anti-inflammatory cytokine IL-10 compared with wildtype controls. Activation of c-fos inhibits NF-κB activity [99].

7. Concluding Remarks

How nervous system, especially vagus nerve, modulates inflammation and immunity has been a puzzle for many years. In past decade, a large body of evidence has shown that the classical CAP could systemically modulate proinflammatory responses via spleen. More recently, the regulatory role of local CAP is emerging and challenging. The immediate questions we have to answer are the following. How vagus nerve senses the PAMPs or DAMPs in the airspaces of the lung? What sensors and receptors are used by sensory nerve endings during this process? To where and how does vagus nerve send the pathogenic signals? How are pathogenic signals being integrated or transformed in the brain center? What are targeting cell population of vagus nerve in the deferent stages of infection and inflammation? How signaling pathways are finely tuned by vagus nerve spatially and temporally during infection and inflammation? In summary, the overall task of this review is to extend our understanding of how nervous and immune systems work collaboratively during infection and inflammation.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.


This work is supported by The National Natural Science Foundation of China (Grant nos. 81270139, Xiao Su; and 81300058, H.Y.W.); The Key Project of Science and Technology of Shanghai (Grant no. 12JC1408900, Xiao Su); One Hundred Person Project of the Chinese Academy of Sciences (Grant no. Y316P21209, Xiao Su); The Knowledge Innovation Program of the Chinese Academy of Sciences (Grant no. Y114P11209, Xiao Su).