Long-term endogenous acetylcholine deficiency potentiates pulmonary inflammation in a murine model of elastase-induced emphysema

Acetylcholine (ACh), the neurotransmitter of the cholinergic system, regulates inflammation in several diseases including pulmonary diseases. ACh is also involved in a non-neuronal mechanism that modulates the innate immune response. Because inflammation and release of pro-inflammatory cytokines are involved in pulmonary emphysema, we hypothesized that vesicular acetylcholine transport protein (VAChT) deficiency, which leads to reduction in ACh release, can modulate lung inflammation in an experimental model of emphysema. Mice with genetical reduced expression of VAChT (VAChT KDHOM 70%) and wild-type mice (WT) received nasal instillation of 50 uL of porcine pancreatic elastase (PPE) or saline on day 0. Twenty-eight days after, animals were evaluated. Elastase instilled VAChT KDHOM mice presented an increase in macrophages, lymphocytes, and neutrophils in bronchoalveolar lavage fluid and MAC2-positive macrophages in lung tissue and peribronchovascular area that was comparable to that observed in WT mice. Conversely, elastase instilled VAChT KDHOM mice showed significantly larger number of NF-κB-positive cells and isoprostane staining in the peribronchovascular area when compared to elastase-instilled WT-mice. Moreover, elastase-instilled VAChT-deficient mice showed increased MCP-1 levels in the lungs. Other cytokines, extracellular matrix remodeling, alveolar enlargement, and lung function were not worse in elastase-instilled VAChT deficiency than in elastase-instilled WT-controls. These data suggest that decreased VAChT expression may contribute to the pathogenesis of emphysema, at least in part, through NF-κB activation, MCP-1, and oxidative stress pathways. This study highlights novel pathways involved in lung inflammation that may contribute to the development of chronic obstrutive lung disease (COPD) in cholinergic deficient individuals such as Alzheimer’s disease patients.

www.nature.com/scientificreports/ and the BAL fluid obtained by washing the last of the airways with 3 × 0.5 mL of sterile saline solution 41 . For total and differential cell counts, the BAL fluid was centrifuged at 112.03×g for 10 min and the cell pellet was resuspended in 0.2 mL of sterile saline. The total number of viable cells was determined in a Neubauer hemocytometer counting chamber. Differential cell counts were performed on BAL fluid cytocentrifuge preparations (450 rpm for 6 min) (Cytospin, Cheshire, UK) stained with Diff-Quick (Biochemical Sciences Inc., Swedesboro, NJ). At least 300 cells were counted according to standard morphological criteria.
Pulmonary morphometry. After collection of BAL fluid, the anterior chest wall was opened, the lungs were removed en bloc and fixed with 4% formaldehyde for 24 h under a constant pressure of 20 cmH 2 O. The lung was then transferred to 70% ethanol and subjected to conventional histological techniques.

Alveolar diameter evaluated by mean linear intercept (Lm).
For conventional morphometry, an eyepiece with a coherent system of 50 lines, 100 points and a known area was attached to the ocular microscope. Lm, an alveolar diameter indicator, was evaluated by the point-counting technique 43 in 20 non-overlapping lung parenchyma fields per animal with a × 200 magnification, as previously described 43-45 . Pulmonary remodeling. Histological sections were stained for collagen fibers using Sirius-Red (Direct Red 80, C.I. 35780, Aldrich, Milwaukee, USA) and for elastic fibers using Oxidate Weigert Resorcin-Fuchsin. Using the same ocular described above, we evaluated the volumetric ratio of collagen and elastic fibers in the alveolar tissue using a dot counting technique 46 . The volumetric proportion of collagen or elastic was determined by dividing the number of points that reach collagen or elastin by the total number of points that reach the alveolar septa. Measurements were performed at 10-15 lung fields for each animal at a magnification of 400 × and the results were expressed as percentage 44,45 . Immunohistochemical evaluation. Immunohistochemical 47 . For the negative control, the primary antibody was omitted from the procedure and BSA was used instead. Using the point-counting technique described above, we determined the density of positive cells expressing NF-κB, macrophages, and MMP-12 in the lung parenchyma and in the peribronchovascular area in 10-15 fields per animal. Measurements were performed at × 1000 on each slide 48 and the results were expressed as positive cells/area (10 4 μm 2 ). The expression of 8-isoprostane was evaluated using a digital analysis system and specific software (Image Pro Plus v. 4.5 for Windows, Media Cybernetics, USA, https:// www. media cy. com/). Sections were stained with an 8-isoprostane antibody and captured using a microscope (DM2500, Leica, Wetzlar, Germany) attached to a camera (Leica), and images were fed into a computer using Qwin Plus (Leica) software (https:// www. leica-micro syste ms. com/). The area stained with isoprostane (%) was expressed as the amount of isoprostane in a specific frame relative to the total tissue area within that frame and was analyzed in lung tissue and the peribronchovascular area. All the morphometric analysis was performed by two researchers who were unaware of the study groups.
Measures of cytokines. In 5 additional animals from each group described above, the lungs were removed and rapidly frozen to perform cytokine measurements on the lung homogenate. The Bradford protein assay (Bio-Rad Laboratories, Hercules, USA) was used to measure total protein as described 49 . A Milliplex mouse plex cytokine assay kit (Merck Millipore, Billerica, USA) was used to test samples for the presence of MCP-1 (monocyte chemoattractant protein-1), IL-6, IFN-γ, MIP-2 (macrophage inflammatory protein-2) and IL-10. The assay was read in the Bio-Plex suspension array system, and the data were analyzed using the Bio-Plex Manager version 4.0 software. Levels of the analyzed cytokines were obtained using standard curves ranged from 32,000 to 1.95 pg/mL, as previously described. Results of all cytokines were expressed as pg/mg of protein.
Statistical analysis. Statistical analysis was performed using SigmaStat software (SPSS Inc., version 10, California, USA, https:// systa tsoft ware. com/ produ cts/ sigma plot/). Normality was assessed by the Kolmogorov-Smirnov test and data were expressed as mean ± SEM. Parametric data were analyzed by two-way ANOVA (emphysema and VAChT deficiency), followed by the Holm-Sidak post hoc test. The level of significance was adjusted to 5%.

Results
On day 28, animals were weighted and both VAChT KD HOM -SAL ( www.nature.com/scientificreports/ lar in both genotypes, suggesting that cholinergic deficiency does not make tissue elastance worse (Fig. 1A). No significantly differences were found in airway resistance among experimental groups (data not shown). We examined the role of VAChT deficiency in the susceptibility to elastase-induced lung inflammation in BAL fluid and lung. WT-PPE and VAChT KD HOM -PPE groups showed increased number of macrophages ( Fig. 1B) and neutrophils (Fig. 1C) when compared to control mice that received saline (P < 0.05). Interestingly, while VAChT KD HOM -PPE mice showed increased number of lymphocytes (   Fig. 2N). However, the effect was similar in both genotypes, suggesting that cholinergic deficiency does not worsens alveolar destruction (Fig. 2N,O).
To assess whether cholinergic deficiency interfered with cytokines release, we analyzed levels of IL-6, MIP-2, MCP-1, IL-10 and IFN-γ in the lung homogenate of the four groups studied ( Table 1). All cytokines were Isoprostane, a marker of oxidative stress 47 , was evaluated by immunohistochemistry. We found that WT-PPE and VAChT KD HOM -PPE mice showed increased 8-isoprostane-PGF-2α staining in the lung parenchyma (Fig. 3L) compared to saline groups (P < 0.05). VAChT KD HOM -PPE also showed an increase in 8-isoprostane-PGF-2α staining (P < 0.01) in peribronchovascular area compared to VAChT KD HOM -SAL (Fig. 3K), which was much stronger than that observed for WT-PPE mice. Representative photomicrographs showing slice of lung stained for 8-isoprostane are shown in Fig. 3M-T.
Endogenous VAChT deficiency did not interfere with lung remodeling in mice instilled with elastase, although increased MMP-12 positive cells. Pulmonary remodeling is a characteristic of emphysema, and some features of remodeling are the deposition of extracellular matrix fibers. Both WT-PPE and VAChT KD HOM -PPE groups showed increase of collagen and elastic fibers deposition in both peribronchovascular tissues (Fig. 4A,G-K,Q-T) and pulmonary parenchyma (Fig. 4B-F,L-P) compared to saline groups (P < 0.001 for all comparisons). However, there was no significant difference between genotypes. These results suggest that cholinergic deficiency did not interfere in the deposition of extracellular matrix in the pulmonary parenchyma and in the peribronchovascular axis.
In Fig. 5, data from positive cells to MMP-12 in peribronchovascular area and lung parenchyma are shown. We found that WT-PPE and VAChT KD HOM -PPE showed increased number of positive cells to MMP-12 compared to respective saline groups (P < 0.01). However, in both pulmonary compartments, animals with cholinergic deficiency and emphysema (VAChT KD HOM -PPE) showed more augment of MMP-12 expression than observed in wild-type (WT-PPE) (P < 0.05).

Discussion
The present study investigated whether long-term endogenous cholinergic deficiency is involved in emphysema development induced by elastase. The major finding in the present study is that VAChT deficiency increases pulmonary inflammatory responses induced by elastase, without affecting the emphysema development, lung function, and pulmonary remodeling. These results suggest that VAChT levels, and consequently ACh release, can modulate lung inflammation in an emphysema model reinforcing previous data that ACh has an important protective role against pulmonary inflammation in different models of pulmonary diseases 37,38 .
Emphysema is one of the COPD manifestations. Our data validated previous findings that a single dose of elastase instillation can induce a significant increase in alveolar diameter, and decrease tissue elastance, which are the main feature of emphysema 51 . Elastase instillation also induced an increase in macrophages and neutrophils in BAL fluid and an increase in MAC2 positive macrophages in lung tissue and peribronchiolar area. We also found increased levels of cytokines in the lung, as well as in oxidative stress and NF-κB positive cells in lung tissue and peribronchiolar area. Increased collagen and elastic fibers deposition in lung parenchyma and peribronchovascular area suggest a process of lung remodeling. Elastase instillation is not the most physiological way to induce emphysema, especially compared to human emphysema, which is most induced by cigarette smoke. However, this model show similar characteristics to the lungs of patients with COPD 1,6,14,52 and has been used in other experimental studies of emphysema 44,45,53 .
Considering the multiple biological functions related to ACh and the importance of this mediator in inflammation, we hypothesized that changes in endogenous cholinergic neurotransmission affect the pathophysiology Table 1. The effects of reduction in VAChT levels on pulmonary cytokines. Data represent the mean ± S.E.M of four to five animals per group. The levels of IL-6, IL-10, IFN-ϒ, MCP-1 and MIP-2 (pg/mg) in lung homogenate were increased in animals submitted to the elastase protocol compared to the saline groups. MCP-1 levels were increased in VAChT KD HOM-PPE group compared to WT-PPE. *P < 0.05 compared to WT-SAL and VAChT KD HOM-SAL. **P < 0.05 compared to WT-PPE. www.nature.com/scientificreports/  www.nature.com/scientificreports/ in the peribronchovascular area. MAC-2 expression in macrophages has been shown to suggest that these cells were activated by inflammatory stimuli 44,45,53 . Macrophage plays an important role in COPD since these cells induce the release of several proteases involved in lung destruction and remodeling 7,54 . In addition, lymphocytes, especially CD8 + and neutrophils, are also involved in COPD, as they can release pro-inflammatory cytokines and proteases. The bronchoconstriction action of ACh in muscarinic receptors has been intensely studied in lung diseases 30 . A role for ACh in nicotinic receptors has been recognized in acute models of inflammation. Nicotinic receptors are expressed in bronchial and alveolar epithelial cells, as well as in inflammatory cells, such as macrophagic cells, neutrophils, and lymphocytes [55][56][57] . Binding of ACh to α7 nicotinic receptors (nAChR) inhibits the production of TNF-α, MIP-2 and other inflammatory cytokines 19,28,34,58 . More related to COPD, Zhang et al. 59 showed that the nAChR gene is a susceptibility variant for the development of COPD. In addition, Budulac et al. 60 suggested that single nucleotide polymorphisms in the nAChR cluster are indirectly involved in the development of emphysema, interfering with smoking, increasing nicotinic dependence in humans. Recently, it was shown that the use of an agonist of nAChR7 suppressed the release of inflammatory mediators by human peripheral blood mononuclear (PBMCs) from COPD patients 61 .

WT-SAL
We found that VAChT-mutant mice showed a two-fold increase in MCP-1, an inflammatory protein involved in the recruitment of macrophages. MCP-1 is upregulated in patients with COPD 13 and is involved in mucus hypersecretion and influx of macrophages into the lung 14,15 . The major cell that produces MCP-1 is epithelial cells and macrophages and the last one is increased in VAChT animals that received PPE. In turn, macrophage infiltration is also regulated by MCP-1 release 62 . Therefore, we hypothesized that macrophages are the major source of MCP-1 in this model and the effects of VAChT deficiency in MCP-1 can be due to increase in macrophage. Also important, macrophage is the most important immune cell to express the nicotinic receptors involved in the anti-inflammatory effects of cholinergic system 17 . In addition, MCP-1 has been increased in mice with signal transducer and transcription activator (STAT3) deficiency 63 , a possible pathway involved in the anti-inflammatory cholinergic system 28 . In this regard, we previously showed that mutant mice to VAChT showed reduced expression of tyrosine kinase (JAK-2) in lung 64 , that probably inhibits STAT3 pathway, which can maybe explain the increased levels of MCP-1. Conversely, IL-6 and MIP-2, cytokines also involved in the recruitment of macrophages and neutrophils to the lung and has been found to be increased in the lung of patients with COPD 65,66 were not differentially affected in VAChT-mutant mice.
The anti-inflammatory effect of ACh on α7nAChR is associated with inhibition of NF-κB translocation to the nucleus and consequent inhibition of cytokines released from macrophages and other cells 67,68 . NF-κB is involved in the pathophysiology of COPD and is also increased in the lungs of patients with COPD 69 . We found that www.nature.com/scientificreports/ elastase treatment in mutant mice induced an increase in NF-κB positive cells only in the peribronchovascular area compared to the WT-PPE group, suggesting that this signaling could be one of the mechanisms involved in the amplification of pulmonary inflammation observed in emphysematous and mutant mice. Interestingly that we found this effect in peribronchovascular area and not in lung parenchyma. This can be attributed to the fact that mice have a more pronounced inflammation in this region than in lung tissue or around airways different from what is observed in human lung with COPD 70,71 or because the main source of ACh in lungs is the airway epithelial cells that can produce ACh by a neuronal and non-neuronal mechanism 27 . However, is important to note that in this study, the expression of NF-kB was evaluated in inflammatory cells that can be in this model, macrophages, lymphocytes, and neutrophils since they were detected in BALF.
Other mechanisms may also be involved in the anti-inflammatory effects of ACh such the effects of α7nAChR activation on the JAK2 and activation of the STAT3, thereby reducing the release of proinflammatory cytokines by the induction of SOCS3 63 .
Oxidative stress plays an important role in the development of COPD and induces deleterious effects on the respiratory tract of patients with COPD 72,73 . Instillation of elastase in VAChT-mutant mice induced an increase of isoprostane-8 staining in the peribronchovascular area that was not observed in wild-type animals submitted to the same protocol. These data suggest that increased oxidative stress may be a pathway that partially explains the data obtained in animals with VAChT reduction. Noteworthy, Roy et al. 74 demonstrated that mice with VAChT deficiency in cardiomyocytes show increased oxidative stress in the heart.
Tissue remodeling can be defined by changes in the amount, composition, and organization of the extracellular matrix structure and it is a common feature of repair of tissue damage observed in COPD 75 . Interesting, that there is more alveolar destruction in VAChT-KD-SAL compared to WT-SAL, although both saline groups have less alveolar destruction compared to animals that received elastase. One possibility to explained it is the long-term deficiency of cholinergic tone which induces, per se, an inflammatory milleau as previously showed by Pinheiro et al. 64 , although in the present study we did not found statically difference in lung inflammation between WT-SAL and VAChT-KD-SAL.
As cholinergic deficiency aggravates pulmonary inflammation, we expected that it would also affect pulmonary remodeling. However, morphological analysis of the lung revealed that long-term cholinergic deficiency did not affect the destruction and remodeling of the parenchyma in this model of emphysema. Our data also showed that tissue elastance changes induced by instillation of elastase is not worse in VAChT deficiency than in WT control mice. It makes sense since changes in extracellular pulmonary fibers are one of the most important determinants of pulmonary compliance changes observed in emphysema 75 .
Several evidence suggest that remodeling appears in response to inflammation and lung injury, however the cause-effect of inflammation, remodeling, alveolar destruction and lung function is controversial 76 . Ito et al. 77 demonstrated that lung function and abnormal compliance observed in a mouse emphysema model were associated with collagen remodeling. In this case we did not found any alteration in collagen deposition or lung function between mutant mice and wild-type with emphysema. Another study that investigated pulmonary alterations in a papain-induced emphysema model 40 observed increased number of macrophages starting one day after papain instillation while alveolar destruction, remodeling, and changes in elastance were evident only after day 15. The morphological changes were suggested to be more related to increased MMP-12 expression than to inflammation. In this regard we evaluated MMP-12 positive cells in both peribronchovascular area and in lung tissue and contradicts our hypothesis, cholinergic deficiency in emphysematous mice induced an increase in positive cells for MMP-12.
MMP-12 activity is associated to the destruction of alveolar walls and the use of MMP inhibitors in emphysema have been suggested 78 , however, Manoury et al. 79 showed that mice deficient to MMP-12 did not improve the lung fibrose induced by bleomycin. One limitation of the present study is that we did not evaluate the MMP activity, and we looked to one point during the emphysema development (28 days). Therefore, the specific relationship between MMP-12, collagen deposition, alveolar destruction, and lung inflammation in VAChT emphysematous mice was not totally clear and maybe a time-course study will be necessary to better understand these findings. Together, our data indicates that the increase in inflammatory response was not the main determinant of lung function or alveolar destruction in this model.
In conclusion, we have shown that reduction of cholinergic signaling increases lung inflammation in a model of emphysema at least in the elastase instillation models, ACh exerts a protective anti-inflammatory effect without affecting emphysema and tissue remodeling. To the best of our knowledge, this is the first time that the role of VAChT in pulmonary inflammation in a model of emphysema has been reported. Although these data reveal a new pathway involved in the pathophysiology of COPD, further studies investigating how neuronal or non-neuronal cholinergic signaling contribute to the increase of pulmonary inflammation in emphysema are warranted. Endogenous cholinergic dysfunction in the long term, a situation commonly observed in several diseases, including heart failure, dysautonomia and Alzheimer's disease 36 may facilitate the development of COPD.