Chronic E-Cigarette Exposure Alters Human Alveolar Macrophage Morphology and Gene Expression

Abstract

Introduction

Alveolar macrophages (AMs) are lung-resident immune cells that phagocytose inhaled particles and pathogens, and help coordinate the lung’s immune response to infection. Little is known about the impact of chronic e-cigarette use (ie, vaping) on this important pulmonary cell type. Thus, we determined the effect of vaping on AM phenotype and gene expression.

Aims and Methods

We recruited never-smokers, smokers, and e-cigarette users (vapers) and performed research bronchoscopies to isolate AMs from bronchoalveolar lavage fluid samples and epithelial cells from bronchial brushings. We then performed morphological analyses and used the Nanostring platform to look for changes in gene expression.

Results

AMs obtained from smokers and vapers were phenotypically distinct from those obtained from nonsmokers, and from each other. Immunocytochemistry revealed that vapers AMs had significantly elevated inducible nitric oxide synthase (M1) expression and significantly reduced CD301a (M2) expression compared with nonsmokers or smokers. Vapers’ AMs and bronchial epithelia exhibited unique changes in gene expression compared with nonsmokers or smokers. Moreover, vapers’ AMs were the most affected of all groups and had 124 genes uniquely downregulated. Gene ontology analysis revealed that vapers and smokers had opposing changes in biological processes.

Conclusions

These data indicate that vaping causes unique changes to AMs and bronchial epithelia compared with nonsmokers and smokers which may impact pulmonary host defense.

Implications

These data indicate that normal “healthy” vapers have altered AMs and may be at risk of developing abnormal immune responses to inflammatory stimuli.

Introduction

Alveolar macrophages (AMs) are innate immune cells that defend the lung against inhaled pathogens and toxins via phagocytosis, secretion of cytokines, chemokines, and growth factors.^1,2 AMs exhibit functional heterogeneity and/or plasticity and have been conventionally divided into classically activated (M1) macrophages, responsible for secretion of proinflammatory cytokines, and alternatively activated (M2) macrophages, which promote tissue remodeling, suppression of immune responses and tumor progression.³ A balance between M1 and M2 macrophages is required for lung health, and AM dysregulation can cause pulmonary diseases.⁴

Electronic cigarettes (e-cigarettes) are battery-powered devices that vaporize e-liquids to deliver nicotine to the lung. The nicotine is carried in a liquid vehicle of propylene glycol and vegetable glycerin along with flavors. In humans, vaping alters the protein content of sputum, impairs neutrophil extracellular trap release (NETosis),⁵ modifies the bronchial epithelial proteome,⁶ and increases levels of neutrophil elastase and matrix metalloproteases, likely due to the direct effects of nicotine on airway leukocytes including AMs.⁷

After exposure to cigarette smoke, AMs release inflammatory mediators including TNF-α and CXCL1 and exhibit impaired phagocytosis of both bacteria and apoptotic cells, leading to increased risk of infection and a failure to resolve inflammation.⁸ AMs also contribute to the chronic inflammation seen in chronic obstructive pulmonary disease patients.⁴ E-cigarettes are perceived as a healthy alternative to cigarette smoke. However, the effects of e-cigarettes on AMs are largely unknown. We and others have recently reported that AMs from ostensibly healthy vapers have increased lipid uptake, suggesting that vaping alters AM function.^9,10 To further investigate this phenomenon, we obtained AMs from healthy nonsmokers, smokers, and e-cigarette users (vapers) and studied their phenotype and gene expression profiles to better understand the consequences of chronic vaping.

Methods

Subject Recruitment and Sample Acquisition

This study was approved by the UNC Investigation Review Board (IRB protocol # 13-2227) and all subjects provided written, informed consent. Fiber-optic bronchoscopy was performed according to American Thoracic Society (ATS) Guidelines. Bronchoalveolar lavage was then performed with the scope wedged into a subsegmental orifice of the right middle lobe. Bronchial brushes were collected in sterile Ringer’s solution and kept on ice following the bronchoscopies until samples could be isolated.⁶

Statistical Analysis

Statistical analyses were performed using the statistical computing environment R version 3.3.2 and GraphPad Prism 6 (San Diego, CA). The significance threshold was p < .05 or q < .01.

Detailed description of the study methods have been provided in the supplement information.

Results

AM Appearance and Phenotype Differ Between Nonsmokers, Smokers, and Vapers

We performed research bronchoscopies on 17 nonsmokers, 13 smokers, and 11 vapers. There was no significant difference in age, body mass index, Forced vital capacity (FVC), Forced Expiratory Volume in 1 second (FEV₁), peripheral blood, or bronchoalveolar lavage fluid cell counts between the groups, and both sexes were represented (Supplementary Table S1). We compared the morphology of bronchoalveolar lavage fluid-derived AMs. After exposure to Diff-Quik stain, nonsmokers’ AMs displayed the classic macrophage appearance (Figure 1, left) and smokers’ AMs were darker in color (Figure 1, middle), while vapers’ AMs were intermediate in terms of staining (Figure 1, right). However, vapers’ AMs consistently had cytoplasmic inclusion bodies that were absent from nonsmokers’ and smokers’ AMs (see arrows; Figure 1). To quantify these changes, we wrote series of ImageJ scripts to measure cellular morphological parameters (Supplementary Figure S1, A). Smokers’ AMs showed the greatest density, nonsmokers were the lightest and vapers’ AMs fell in between these two cohorts (Supplementary Figure S1, B). Furthermore, AM circularity and cell perimeter were significantly altered in smokers compared with both nonsmokers and vapers (Supplementary Figure S1, C and D), nucleus size remained unaltered across the three groups, while cell area and the ratio of cell area:nucleus area were significantly decreased in smokers and vapers compared with the nonsmokers (Supplementary Figure S1, E–G).

Figure 1.

Alveolar macrophages (AMs) obtained from smokers and vapers exhibit different cell morphologies to those obtained from nonsmokers. Representative images from cytospin slides of BALF-derived AMs from nonsmokers, smokers, and vapers at low (top) and higher (bottom) magnifications. Arrows highlight the increased cytoplasmic inclusions in vaper’s macrophages. BALF = bronchoalveolar lavage fluid.

We next performed immunocytochemistry to probe for inducible nitric oxide synthase (iNOS) expression as an M1 phenotype marker, and CD301a expression as an M2 marker (Supplementary Figure S2, A). AMs were identified by morphology and M1 and/or M2 marker expression was quantified. Based on this paradigm, vapers had a significant increase in the number of iNOS-positive AMs compared with either nonsmokers or smokers (Supplementary Figure S2, B). These cells also exhibited a corresponding decrease in CD301a staining compared with nonsmokers (Supplementary Figure S2, B). Smokers’ AMs displayed a significant decrease in CD301a staining compared with nonsmokers (p ≤ .001) and had a significant increase in the iNOS and CD301a coexpressing phenotype when compared with nonsmokers (p ≤ .05; Supplementary Figure S2, B).

Vapers’ AMs Have an Altered Gene Expression Profile Compared With Nonsmokers and Smokers

We then performed gene expression analysis on AMs from a subset of subjects (see Supplementary Table S2) using the Nanostring PanCancer-Immune panel of 770 genes. Despite the lack of smoking habit-dependent clustering (Figure 2, A), we observed a significantly number of differentially expressed genes (DEGs) in AMs from both vapers and smokers, relative to nonsmokers (Figure 2, B; Supplementary Table S4). Interestingly, there was little commonality between DEGs in AMs from smokers and vapers. Indeed, vapers’ AMs had a greater proportion of downregulated DEGs (132-down, 42-up) compared with smokers (120-up, 13-down; Figure 2, C) with the commonly altered genes between being mostly upregulated (34-up, 8-down; Figure 2, D). For example, the cytokines IL-19, IL-24, and the chemokine CXCL12 were all downregulated while EBI3, CCL15, and IL-17F were upregulated.

Figure 2.

Smokers and vapers AMs exhibit unique, differentially expressed genes. (A) Gene count data were log₂-transformed, plotted in a heatmap and hierarchically clustered. Individuals were annotated by cohort (nonsmokers [N], smokers [S], and vapers [V]) and sex (male [M] and female [F]). Subjects were divided into groups (Group 1—cyan, Group 2—magenta, and Group 3—green) based on the hierarchical clustering analysis. (B) Volcano plot showing significantly up and downregulated genes in smokers (red) and vapers (black), relative to nonsmokers. Values with a log₂ fold change <−10 or >10 or a −log₁₀q value >10 were omitted for clarity but were factored into all analyses. (C) Venn diagram showing up- (red) and downregulated (blue) genes between smokers (left) and vapers (right), relative to nonsmokers. (D) Heatmap showing differentially expressed genes commonly altered in both smokers and vapers. Genes were considered significant if the fold change >±2, and the corrected p value (q value) was <.1. N = 9 nonsmokers, 10 smokers, and 6 vapers. AMs = alveolar macrophages.

We also obtained bronchial brushings of predominantly epithelial cells by bronchoscopy and performed a similar analysis with the PanCancer-Immune gene panel (Supplementary Figure S3 and Tables S3 and S4). Again, the subjects did not cluster according to smoking and vaping habits (Supplementary Figure S3, A). False-discovery rate-adjusted DEGs showed significant difference in both vapers and smokers, compared with nonsmokers (Supplementary Figure S3, B). However, a different pattern in bronchial brushings compared with AMs and the bronchial brushings from smokers showed a greater proportion of affected genes (77-down, 5-up) than vapers (27-down, 24-up; Supplementary Figure S3, C). All common genes between smokers and vapers in the bronchial brushings samples were downregulated (14-down, 0-up; Supplementary Figure S3, C). The common DEGs between smokers and vapers are shown in Supplementary Figure S3, D. Again, cytokines and chemokines were altered in both smokers and vapers bronchial brushings with CXCL1, CXCL10, and IL-8 all being downregulated. In addition, we also studied a custom panel of 24 genes relevant to airway physiology. These genes were not significantly altered in vapers’ AMs (Supplementary Figure S4). However, MUC5AC was significantly upregulated in bronchial epithelia from both smokers and vapers, while the protease cathepsin B (CTSB), as well as activating transcription factors 4 and 6, were significantly downregulated in vapers (Supplementary Figure S4).

A full list of both AM and bronchial brushing gene fold changes and q values are included in Supplementary Table S4. Since we observed more changes in vapers’ AMs (Figure 2), we focused our analyses on these cells. We observed alterations in gene expression of chemokines, cytokines, and their receptors in vapers’ AMs (Supplementary Figure S5). Whilst IL-3, IL-17F, and TNFRSF13C were upregulated 73-, 16-, and 2-fold, respectively, all other genes, including IL-6, were downregulated. These DEGs clustered mostly into two groups (Supplementary Figure S5, A). In contrast, 31 chemokine and cytokine DEGs were upregulated in smokers’ AMs, suggesting a markedly different inflammatory response (Supplementary Figure S5, B). Accordingly, DEGs for both AMs and bronchial brushing samples were grouped using the PANTHER classification system (Supplementary Figure S6). In every dataset, “cellular processes” were the most involved pathway, followed by “response to stimulus.” Vapers’ AMs and smokers’ bronchial brushings were both characterized by downregulation, whilst both vapers’ and smokers’ macrophages had a number of upregulated processes including metabolism and immune responses (Supplementary Figure S6).

In Vitro Vape Exposure Alters Protein Levels in M1-Polarized THP-1 Cells

To verify some in vivo AM gene expression results at the protein level, we exposed THP-1-derived macrophages to e-cigarette vapor in vitro. Since we observed increased M1-type macrophages in vapers’ lungs (Supplementary Figure S2, B), we polarized THP-1 cells toward the M1 phenotype (see online Methods). We exposed M1-THP-1 cells to varying doses of Banana Pudding-flavored vape puffs. Cells were then lysed and Western blots were performed. We initially chose MUC1 and CD70, which were down- and upregulated in vapers’ AMs, respectively (see Supplementary Table S4). Similar to the human studies, in vitro exposure to a Banana Pudding-flavored vape puffs led to similar down- and upregulation of MUC1 and CD70, respectively (Supplementary Figures S7, A, B and S8). Interleukin-1 receptor-like 2 protein (IL1RL2), exhibited a nonsignificant, but ~30-fold increase in vapers, was also significantly increased in banana pudding-exposed M1-THP-1 cells (Supplementary Figures S7 and S8).

Discussion

By light microscopy, we observed an altered morphological appearance, which included a moderate increase in cell density and an increase in cytoplasmic vacuoles (Figure 1). These findings are very consistent with recent work by Madison et al. who found that vaped mice had lipid laden macrophages that were characterized by “intracytoplasmic inclusions” and the recent case reports of lipid laden macrophages in hospitalized humans.¹¹ We also found that vaping altered iNOS and CD301a expression, which was indicative of an altered M1 and/or M2 phenotype (Supplementary Figure S2). Increased M1 or iNOS phenotype has previously been shown in smokers’ AMs,¹² and our data (Supplementary Figure S2) are consistent with this, suggesting that iNOS is a valid M1 marker. Under normal conditions, AMs from healthy individuals are polarized toward an M2 phenotype.¹³ Interestingly, we found that vaping lead to an increase in iNOS expression, which was significantly greater in vapers than both smokers and nonsmokers, indicating an M1 phenotype (Supplementary Figure S2). Similarly, CD301a expression was significantly reduced in vapers. Our data were consistent in terms of showing upregulation of CCL15 (chemokine prevalent in M1 macrophages), and downregulation of CCL18, CCL26, CD209, CXCR4, and TGF-B (chemokines, chemokine receptors, and C-type lectin prevalent in M2 macrophages)^13,14 (Figure 2, D; Supplementary Table S4).

As with the iNOS and/or CD30a phenotyping, our Nanostring analysis indicated that vaping exerted a greater effect on gene expression in AMs than smoking. Indeed, 174 genes were significantly altered in vapers’ AMs, with the majority being downregulated. In contrast, 133 genes were significantly altered in smokers’ AMs, with the majority being upregulated (Figure 2). A number of cytokines and chemokines were altered in vapers’ AMs, including downregulation of IL-6 and upregulation of IL-3. In the bronchial brushings, we found 51 DEGs in vapers’ bronchial epithelia, compared with 82 DEGs in smokers’ epithelia, suggesting a different response to vaping in bronchial epithelia versus AMs. Broadly speaking, our data were similar to our previous proteomic study, where we found slightly more changes in smokers’ versus vapers’ bronchial epithelia, including upregulation of MUC5AC but not MUC5B in bronchial epithelia (Supplementary Figure S4). Airway MUC5AC levels inversely correlate with FEV₁, suggesting that vapers with elevated MUC5AC levels may be at risk of impending airway obstruction.¹⁵

We also observed that signal transducer and activator of transcription 3 and 6 (STAT3 and STAT6),¹⁶ which regulate airway remodeling, were upregulated in vapers’ but not smokers’ bronchial epithelia (Supplementary Table S4), while unfolded protein response-associated ATF4 and ATF6¹⁷ genes were downregulated. These observations suggest that long-term vaping may cause ultrastructural changes to the airways, and impair airway inflammatory and/or oxidative stress responses. Chemoattractant proteins CXCL1, CXCL10, and IL-8 were also significantly downregulated in both vapers’ and smokers’ bronchial epithelia, suggesting that vaper’s epithelia may not mount an appropriate immune response when challenged.

Despite the significant changes in DEGs, we did not observe any hierarchical clustering based on the three cohorts (Figure 2). AMs are highly plastic and may exhibit multiple phenotypes per individual,¹⁸ which may explain why they failed to cluster. According to the smoking diaries, vapers reported using 11–265 puffs per day over a 2-week period. Moreover, since we wanted to study the airways of “real” vapers, they were allowed to use their own device and their own e-liquids. Given the large number of e-liquids available (>7700) and the range and customizability of the devices, it is highly likely that each vaper in our study used a different e-liquid, device, and device settings. In addition to not controlling for e-cigarette type, further limitations of our study include the relatively small sample size and that many of our vapers were former smokers. Given that (1) most gene expression reverts to baseline levels after smoking cessation¹⁹ and (2) the changes we observed in vapers were significantly different to smokers, it is likely that the observed effects in vapers were not caused by former smoking. However, we cannot exclude the possibility that prior smoking history may alter the lung’s susceptibility to vaping.

An association between vaping with increases in chronic bronchitis-like and asthma symptoms has previously been reported.²⁰ Our current data indicate that altered AMs occur not just in hospitalized e-cigarette users, but also in relatively healthy vapers. In conclusion, we have observed significant changes in AM phenotype and gene expression in vapers that were greater than those seen in smokers. Taken together, these data indicate that vaping has a marked effect on the lung that is different from smoking and indicative of airway remodeling, altered inflammatory status and, possibly, immunosuppression. These changes may leave the lung more vulnerable to infection and the development of pulmonary disease. Thus, vaping should be avoided and not declared safe, or even safer than cigarette smoking, until extensive toxicological studies and more detailed studies have been performed.

Supplementary Material

A Contributorship Form detailing each author’s specific involvement with this content, as well as any supplementary data, are available online at https://academic.oup.com/ntr.

Funding

This work was funded by National Institutes of Health (NIH) and Food and Drug Administration (FDA) grant HL120100, NIH/NHLBI grant HL135642, and NIH and FDA grant HL153698. Research reported in this publication was in part supported by NIH and the FDA Center for Tobacco Products (CTP). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH or the Food and Drug Administration. We thank Martha Almond, Mary E. Braun Martino, Carol Robinette, and Heather Wells for technical assistance.

Declaration of Interests

All authors have no conflicts to disclose.