Lung epithelial cells form a continuous lining of a pseudo-stratified cell layer along the inner walls of airways and form the first line of the defensive barrier for protection from environmental insults and microbial infections [1,2,32]. Various cell types, including basal, ciliated, goblet and alveolar epithelial cells, constitute lung epithelium, and the cellular structure varies considerably depending on their locations within the lung. The integrity and barrier function of lung epithelial cells is maintained by tight junctions which provide strength and regulate permeability. Lung epithelial cells, in a tightly regulated process, sense pathogens and injury, produce cytokines, recruit immune cells and undergo repair under normal conditions [32,33,34]. However, under pathological conditions, the repair process is dysregulated. For example, due to persistent injury or infection, lung epithelial cells can undergo apoptosis or phenotypic transformation to fibroblasts and contribute to inflammation and fibrosis. Such a situation may occur in chronic smoking, leading to lung fibrosis in susceptible individuals [16,32,33].

Thus, multiple cell types (including lung epithelial cells, immune cells and fibroblasts) could contribute to the dysregulated repair reported in lung fibrosis. Numerous altered intracellular signaling pathways, due to repeated injury to the lung epithelium, drive profibrotic processes that culminate in fibrosis and compromise lung function [2,7].

Researchers propose an AOP for lung fibrosis that envisions oxidative stress-induced inflammation as the early molecular mechanism leading to downstream profibrotic cellular mechanisms. Based on the weight of evidence that researchers gathered, oxidative stress was considered as a MIE that leads to increased secretion of proinflammatory and profibrotic mediators (KE1). At the cellular level, a proinflammatory response induces recruitment of inflammatory cells (KE2), which in turn increase fibroblast/myofibroblast differentiation (KE3). At the tissue level, activated fibroblasts and myofibroblasts secrete and deposit extracellular matrix proteins (KE4). This process culminates in lung fibrosis as the adverse outcome.

Exposure to stressors, such as CS, asbestos fibers, silica particles, carbon nanotubes or radiation, perturb the oxidant–anti-oxidant balance, overwhelm anti-oxidant defenses, and have been reported to promote fibrosis in the lung [9,13,36]. Here, researchers have used CS as an example of a stressor that induces oxidative stress [5,8,19,37,38]. CS is a complex, dynamic aerosol that is a mixture of several classes of chemicals and contains a large number of toxicants generated during the combustion process. The US FDA has identified 93 toxicants as hazardous and potentially hazardous compounds in CS and designated them as causative agents for several smoking-related diseases [39]. CS is rich in free radicals such as nitric oxide and contains quinones and semiquinones which are capable of inducing oxidation and reduction. In addition, CS contains many toxicants including volatile organic compounds (e.g., acrolein, crotonaldehyde and formaldehyde), polycyclic aromatic hydrocarbons and heavy metals (e.g., cadmium) that can cause oxidative stress [10].

CS contains high concentrations of oxidants, and the local (lung) and systemic consequences of smoking-induced oxidative stress are widely understood [10]. Airway epithelium, which functions as a physiological barrier to inhaled harmful chemicals, is the primary target of smoking-induced oxidative stress [2,33]. Epithelial injury drives the production of endogenous reactive oxygen and nitrogen species (ROS and RNS) in airway epithelial cells and macrophages [40,41,42,43]. One of the early responses to CS exposure in airway epithelial cells (and several other cell types) is the rapid upregulation of the expression of oxidative stress response genes including, HMOX1, GPX2, GCLC, GCLM, PRDX1 and NQO1, among others [44,45].

Oxidative stress depletes anti-oxidant defenses (e.g., glutathione) in smokers and the expression of the key antioxidant transcription factor Nrf-2 is downregulated [13,21,46]. As a result of these changes, biomarkers of oxidative stress, such as F2-isoprostane and other arachidonic acid metabolites and systemic inflammation markers such as white blood cells, are elevated in chronic smokers [47].

In vitro exposure of lung epithelial cell cultures to CS depletes intracellular glutathione levels [48]. Additionally, the ROS-activated transforming growth factor (TGF)-beta 1 signaling also suppresses Nrf2 expression. The downregulation of Nrf2 activity is reflected in a decreased expression of antioxidant enzymes such as catalase, superoxide dismutase 1 (SOD1) and 2 (SOD2), heme oxygenase 1, γ-glutamylcysteine synthetase and glutathione peroxidase 1 (GPX1), which further exacerbates oxidative stress [49,50,51,52]. A recent review summarized the extensive evidence on oxidative stress in various models of bronchial epithelial cells from exposure to CS preparations [53].

In a bleomycin mouse model of lung fibrosis, increased oxidative stress was evident from the downregulation of Nrf-2, which also resulted in the proliferation of lung fibroblast and the secretion of collagen [54]. Restoring the redox balance by treatment with activators of Nrf-2, such as dimethyl fumarate and sulforaphane, mitigated features associated with fibrosis in mice treated with bleomycin [55,56].

A breach of the lung barrier function from exposure to inhaled toxicants results in increased epithelial permeability and cell death, and triggers inflammation by lung epithelial and the resident immune cells [33]. Oxidative stress triggers inflammatory responses in the lungs [57,58]. Exposure to CS induces the release of inflammatory cytokines from lung epithelial cells and immune cells [30,59,60].

For example, cultured bronchial epithelial cells, on exposure to CS extracts, release interleukin-8 (IL-8) which is a chemotactic factor for neutrophils [61]. Further, bronchoalveolar fluid (BAL) samples from smokers contain higher levels of IL-8 and increased counts of neutrophils [62]. Several other reports also indicate that the release/expression of proinflammatory mediators (such as tumor necrosis factor-alpha (TNF-alpha), chemokine ligands GRO-1 and CCL-2, and multiple interleukins) is secreted from alveolar epithelial cells and lung macrophages in response to oxidative stress arising from lung injury (KE1) [34,63,64,65,66,67,68,69].

Two similar KEs (event ids #149 and #901) which describe inflammatory responses exist in the AOP knowledge base. In the context of this AOP, the event id# 149, “Increase, Inflammation” is proposed as the KE 1 as it is more appropriate because of the context of inflammation following lung epithelial injury.

Following lung injury, immune cells are recruited to the lung, and those cells produce mediators of inflammation, including ROS and proinflammatory cytokines [70,71,72]. Treatment with CS extracts releases the chemotactic proteins ICAM1, MCP, GM-CSF from lung epithelial cells [42,73,74,75,76]; this promotes further recruitment and activation of pro-inflammatory cells (KE2). Treatment of NHBE cells with CS preparations results in alterations in cytokines and the associated pathways, together with changes in xenobiotic and oxidative stress pathways [77,78]. Additionally, increases in the infiltration of neutrophils, dendritic cells, and lymphocytes were demonstrated in animal models of lung inflammation, fibrosis, emphysema, and CS-exposed in vitro systems [72,79,80,81,82]. While KEs #901 and 1497 address mechanisms of recruitment of inflammatory cells, because of its relevance to the AO, the KE #1497 is used in the current AOP as the KE2.

TGF-beta 1 pathway is widely recognized as the main inductor and regulator of fibrosis. TGF-beta 1 plays a central role in fibrotic processes in the lung by inducing epithelial–mesenchymal transition (EMT), driving the proliferation of fibroblasts and differentiation towards myofibroblasts (KE3) [32,35,83,84]. An existing KE in the AOP knowledge base captures the cellular events in the KE3, and the KE #1500 is proposed as the KE3.

Lung fibrosis is characterized by excessive deposition of ECM (KE4) in the lung parenchyma by different stressors [60,85,86,87]. The abnormal ECM deposition remodels the normal lung tissue architecture by replacing epithelial tissue with fibroblasts, compromising gas exchange and leading to respiratory failure and death [88]. Multiple stressors and increased oxidative stress can additionally induce the release of TGF-beta 1 and collagen secretion from human lung fibroblasts via cellular influx through chloride channels. Activated TGF-beta 1, in turn, induces the expression of ECM proteins such as fibronectin, collagen I, collagen III, α-smooth muscle actin (ACTA2), vimentin and MMP-9 from fibroblasts [84,89,90,91,92], thereby leading to the excessive deposition of ECM [13]. TGF-beta 1 also induces the secretion of plasminogen activator-1, which blocks plasminogen activation, fibronectin proteolysis, and fibroblast apoptosis [93]. Additionally, TGF-beta 1 further promotes the remodeling of ECM and EMT in ROS-dependent mechanisms [94]. The KE #68 in the AOP knowledge base captures these tissue-level events.

In summary, inhaled toxicants/oxidants drive lung fibrosis (AO) by creating persistent oxidative stress (MIE) and the subsequent secretion of proinflammatory and mediators (KE1), which feed on each other. Unresolved proinflammatory response induces the recruitment of inflammatory cells (KE2) and establishes a state of chronic inflammation, which in turn promotes fibroblast proliferation and myofibroblast differentiation (KE3). This chronic inflammation causes recurrent lung injury, abnormal wound healing, and fibrosis in susceptible individuals. At the tissue level, activated myofibroblasts deposit increased ECM proteins (KE4) and subsequently lead to lung fibrosis as the AO.

It is well established that oxidative stress and inflammation co-exist and constitute feedback loops. Substantial knowledge exists on the role of persistent oxidative stress promoting proinflammatory profibrotic changes in the lung. In this AOP, researchers propose that oxidative stress, as an MIE, causes lung injury (KE1) and drives the subsequent KEs (2–4) and the AO. The relationship between inhaled toxicants (e.g., cigarette smoking) oxidative stress and inflammation through KERs A and B, is amply demonstrated. Similarly, the role of profibrotic mediators (KER C) and ECM deposition (KER E) to the AO are well described in the AOP 173 [25]. These KERs are not discussed further in this manuscript. Here researchers propose a new KER, KER D in which the MIE directly leads to events in the KE #3 (Increased fibroblast proliferation and myofibroblast differentiation) through a new KER, KER D [30].

Lung fibroblasts play an important role in the remodeling of the extracellular matrix and subsequent progression to lung fibrosis [95]. Exposure of CS extracts induces senescence in lung fibroblasts, and subpopulations of senescence-resistant fibroblasts have been shown to undergo increased proliferation, mobility and secretion of TGF beta-1. Treatment with antioxidants blocked CS extract-induced proliferation of senescence-resistant fibroblasts, but not the other responses [96].

TGF beta-1 is an important multifunctional cytokine which regulates fibrosis and is tightly regulated through multiple mechanisms, including ROS [84]. Several reports indicate that ROS directly activates TGF beta 1 and induces its release from macrophages and airway epithelial cells [21,97], and thus regulates TGF-beta 1-induced profibrotic effects [98].

TGF beta-1 is synthesized and secreted into the extracellular space as a large latent complex containing mature dimers of TGF beta-1 bound to latency-associated peptide (LAP). One of the key mechanisms of activation of TGF beta-1 involves its release from LAP, through a redox mechanism [99]. Nitrosative stress was also reported to interfere with the ability of LAP to bind and inactivate TGF beta-1 [100]. ROS generated from ionizing radiation have been shown to release TGF-beta 1 from the latent form in an isoform-specific manner. The latent form of TGF beta 1 was hypothesized to contain a redox switch, and the oxidative modification of LAP-beta 1 was shown to activate TGF beta-1 [101]. Similarly, asbestos-derived ROS activate TGF beta-1, which was blocked by antioxidant enzymes (SOD and catalase), deferoxamine (chelates iron), or anti-oxidants [102,103].

Separately, TGF beta-1 itself was reported to induce ROS production and regulate the expression of anti-oxidant and pro-oxidant genes leading to further exacerbation of oxidative stress [13,104,105]. The anti-oxidant N-acetyl cysteine (NAC) reversed several aspects of TGF beta 1-induced EMT in rat alveolar epithelial cells, including restoration of intracellular glutathione levels and cellular morphology. The treatment with NAC also reduced the levels of TGF beta 1-mediated increases in mesenchymal markers such as α-smooth muscle actin and vimentin [94]. It was also reported that continuous treatment with aqueous extracts of CS induced the expression of TGF beta-1 and resulted in phenotypic changes consistent with fibroblasts [106].

Findings from murine models also lend support to the role of oxidative stress in lung fibrosis. A recent study indicates that the ROS-Nrf2 pathway mediates the development of TGF beta-1-induced EMT through the activation of Notch signaling, although its role in fibrosis remains unknown at present [107]. Furthermore, in a knockout-mouse model (deficient in the p47 subunit of NADPH oxidase) BAL cells were unable to produce ROS in response to bleomycin treatment. It was observed that in the bleomycin-treated group, collagen deposition was absent in the lungs, indicating the importance of endogenous ROS in lung fibrosis [108].

Additional evidence for ROS as drivers of lung fibrosis comes from the bleomycin-induced mouse models. First, inhibition of bleomycin-induced oxidative stress with anti-oxidant treatments suppressed increases in inflammatory cell counts and blocked lung fibrosis. The authors measured the levels of malondialdehyde (MDA), extracellular SOD, total antioxidant capacity and ROS in lung tissue and of NF-kB-mediated inflammatory genes expression of TNF-alpha, IL-1 beta and IL-6 in these animal models [109,110]. Second, glutaredoxin, an enzyme involved in redox signaling, reverses a post-translational modification known as protein S-glutathiolation and restores cysteine residues on target proteins [111,112]. Recently, it was shown that mice deficient in the glutaredoxin gene (GLRX) show excessive collagen deposition following either treatment with bleomycin or recombinant TGF beta-1. Further, transgenic mice that conditionally overexpress GLRX suppressed TGF beta-1-induced lung fibrosis [113]. Thus, several lines of evidence support the role of ROS and oxidative stress in the development of lung fibrosis (KER D).

Multiple stressors, including CS, cause imbalance in oxidants and reductants, chronic inflammation, and fibrotic responses in the lung [36,114,115,116]. Additionally, lung fibrosis increases the risk of lung cancer [8]. With cigarette smoking being a significant risk factor, a focused AOP, such as this one, may aid in identifying and evaluating novel targets for early diagnosis and treatment options for lung fibrosis that are more efficacious.

Combustion-related toxicants are major drivers of smoking-related diseases, which develop with a latency of several decades [10]. Emerging and traditional non-combusted tobacco products do not generate combustion-related toxicants. In addition, aerosols of the inhaled potentially reduced risk products are chemically less complex and contain significantly lower levels of oxidants [117], as well as not inducing oxidative stress-response genes [44]. Non-combustible products have been hypothesized to have a lower risk of smoking-related diseases [118,119]. This approach is consistent with the guidance issued by the US FDA that for tobacco products to gain approval through the FDA’s modified risk tobacco product review must demonstrate or be reasonably likely to result in reduced risk, and a reduced public health burden of cigarette smoking [120]. In that context, AOPs are suggested as potential tools to assess the relative risk of cigarettes and alternative tobacco products [121]. Thus, this AOP should guide the development of fit-for-purpose alternative assays for the evaluation of candidate reduced risk tobacco products.

It should be noted that smoking cessation continues to be the best option to reduce harm from smoking [122], although tobacco harm reduction could include complete substitution of smoking with less harmful tobacco products for those who are unwilling to quit [123].