Epithelial-mesenchymal transition: Insights into nickel-induced lung diseases

Abstract

Nickel compounds are environmental toxicants, prevalent in the atmosphere due to their widespread use in several industrial processes, extensive consumption of nickel containing products, as well as burning of fossil fuels. Exposure to nickel is associated with a multitude of chronic inflammatory lung diseases including asthma, chronic obstructive pulmonary disease (COPD) and pulmonary fibrosis. In addition, nickel exposure is implicated in the development of nasal and lung cancers. Interestingly, a common pathogenic mechanism underlying the development of diseases associated with nickel exposure is epithelial-mesenchymal transition (EMT). EMT is a process by which the epithelial cells lose their junctions and polarity and acquire mesenchymal traits including increased ability to migrate and invade. EMT is a normal and essential physiological process involved in differentiation, development and wound healing. However, EMT also contributes to a number of pathological conditions including fibrosis, cancer and metastasis. Growing evidence suggest that EMT induction could be an important outcome of nickel exposure. In this review, we discuss the role of EMT in nickel-induced lung diseases and the mechanisms associated with EMT induction by nickel exposure.

1. Introduction

Nickel compounds are environmental and occupational toxicants. Nickel occurs naturally in the atmosphere through volcanic activity, forest fires and weathering of rocks. However, industrial processes such as food processing, mining and refining, as well as extensive use of nickel-containing products such as stainless steel, jewelry and electric equipment, and medical implants increases its atmospheric levels [1–4]. In addition, sewage sludge incineration and fossil fuel combustion significantly contribute to atmospheric nickel levels [5]. While the average nickel concentration in ambient air in the United States is 2.22 ng/m³, in occupational settings, the concentrations of airborne nickel range from 1–60,000 μg/m³ [6]. Humans are primarily exposed to nickel through inhalation, ingestion and dermal contact [1, 4]. Furthermore, smoking cigarettes is an important source of human nickel exposure.

Nickel exposure is associated with a number of health hazards in humans. Pulmonary absorption is a major route of nickel exposure. Therefore, the lung is a major target organ of nickel toxicity and lung and nasal cancers are among the most serious consequences of exposure to nickel compounds [4, 7, 8]. Increase in mortality due to lung and nasal cancers has been reported in nickel refinery workers [6]. Furthermore, nasal cancer risk in nickel refinery workers increased with the duration of employment [9]. Both water-soluble and -insoluble nickel compounds cause neoplastic transformation of cells [10–12]. In animals, nickel compounds induced tumors at the sites of exposure [12]. Therefore, certain nickel compounds have been classified as Group 1 Carcinogens by the International Agency for Research on Cancer (IARC) [3, 13].

In addition to cancers, non-malignant respiratory diseases of nickel exposure have also been well-documented. Nickel exposure is associated with the development of a number of respiratory ailments including asthma, chronic bronchitis, emphysema, pulmonary fibrosis, pulmonary edema and reduced lung function [14–16]. Occupational asthma has been reported in electroplating workers exposed to nickel fumes [17]. Furthermore, inhalation of nickel salts has been experimentally shown to cause bronchoconstriction [18–20]. At occupational exposure levels, nickel sulphate, nickel subsulphate and nickel oxide caused lung inflammation and fibrosis in rats and mice. A positive correlation has been shown between the solubility of nickel compounds and the severity of lung inflammation and fibrosis, with nickel sulphate being the most toxic followed by nickel subsulphate and nickel oxide [6]. Furthermore, studies on nickel refinery workers showed association of nickel exposure with the development of chronic obstructive pulmonary disease (COPD) [21].

Interestingly, a key process that drives the pathogenesis of lung diseases associated with nickel exposure is epithelial-mesenchymal transition (EMT). EMT is a process during which the epithelial cells lose cell-cell adhesion and acquire mesenchymal properties, including increased ability to migrate and invade [22]. EMT is a normal cellular process that plays a major role during development and repair of damaged tissues. However, aberrant EMT induction results in organ fibrosis. Growing evidence suggest EMT induction as an important outcome of nickel exposure [23–26]. Here, we discuss the role of EMT in the development of diseases associated with nickel exposure and the potential mechanisms underlying EMT induction by nickel exposure.

2. Epithelial-mesenchymal transition (EMT) in the pathogenesis of lung diseases

EMT is a process by which the epithelial cells lose their cell polarity and tight junctions, reorganize their cytoskeleton and acquire mesenchymal characteristics including increased invasive and migratory potentials [22]. Downregulation of E-cadherin, an important cell-cell adhesion protein, is a hallmark of EMT. In addition, repression of epithelial markers such as claudins, occludin and desmoplakin, causing the dissolution of adherens junctions and tight junctions, is commonly observed during EMT. Downregulation of epithelial markers is accompanied by upregulation of mesenchymal markers including fibronectin, vimentin, α-smooth muscle actin (α-SMA), desmin, and N-cadherin [22, 27]. Furthermore, several transcription factor families such as SNAIL, ZEB and TWIST, which are identified as EMT master regulators, are also upregulated during EMT. EMT is a reversible process, with the mesenchymal cells capable of undergoing mesenchymal-epithelial transition (MET). During MET, the apico-basal polarity of the cells is established, junctional complexes are assembled and tight junctions are formed. MET generates epithelial cells at different developmental stages and also during metastatic colonization [28, 29].

EMT is categorized into three subtypes based on their distinct biological settings and different functional consequences. Type 1 EMT is associated with developmental processes such as gastrulation, neural crest formation and heart value formation. During type 1 EMT, the first set of mesenchymal cells known as primary mesenchyme are generated to create new tissues with diverse functions [30]. Type 2 EMT, which generates fibroblasts, is associated with wound healing and tissue regeneration. This type of EMT occurs as part of repair process to reconstruct tissues following trauma or inflammation. Although type 2 EMT is a normal physiological process, chronic inflammation and aberrations in myofibroblast activation could lead to increased deposition of extracellular matrix proteins, resulting in fibrosis and organ destruction [30–32]. Type 3 EMT is associated with the dissemination of cells from the primary tumors, thus playing a major role in epithelial cancer cell metastasis [22].

Emerging evidence implicates EMT in the pathogenesis of a number of lung diseases:

1. Pulmonary fibrosis is characterized by accumulation of myofibroblasts and increased deposition of extracellular matrix proteins. Injury to lung epithelium is an important event in the development of this disease. Upon injury, the epithelial cells undergo EMT, resulting in morphological changes which give rise to fibroblast-like cells [33]. Transforming growth factor-β (TGF-β) levels increase in the injured lung epithelial cells and has been identified as a major driver of EMT in lung fibrosis [33, 34]. TGF-β-exposed alveolar epithelial cells undergo EMT and exhibit fibroblast-like morphology. Furthermore, lung biopsies of idiopathic pulmonary fibrosis patients suggest contribution of EMT to pulmonary fibrosis [34].

2. COPD is a progressive airway obstructive disease, which involves destruction of lung parenchyma (emphysema) and chronic inflammation of large airways (bronchitis) [35]. In addition, pulmonary fibrosis is reported in COPD patients [36]. Emerging evidence strongly suggest EMT as a driver of COPD pathogenesis. In COPD associated with smoking, the epithelium is in an activated state and the structural changes to the underlying reticular basement membrane (Rbm) suggest EMT. Furthermore, increased vascularization of Rbm in large airways indicates type 3 EMT [37–40].

3. Asthma is a chronic disease characterized by progressive airway remodeling. In addition, chronic airway inflammation and airway hyperresponsiveness are major features of asthma [41, 42]. Recent studies have demonstrated the importance of EMT in airway remodeling in asthma, which include airway wall thickening, subepithelial fibrosis, increased smooth muscle mass and angiogenesis [42, 43]. Downregulation of cell adhesion proteins occur in asthmatic epithelium [44]. Furthermore, increased EMT has been observed in asthmatic patients as compared to normal individuals [42, 45].

4. In epithelial cancers, loss of cell-cell adhesion results in dissociation of the primary tumor mass and metastatic dissemination of the cancer cells. EMT is a critical process through which cancers acquire invasiveness and progress to a metastatic state. Increased production of TGF-β is normally observed in cancer cells and several studies suggest TGF-β signaling pathway as a major driver of EMT in cancer progression [46, 47]. Furthermore, recent studies have also identified cell dissemination from premalignant lesions, which suggests that EMT could be involved cancer initiation, in addition to being associated with metastasis [48–51].

Nickel, an EMT inducer, is implicated in the development of a number of lung diseases described above, including pulmonary fibrosis [52, 53], COPD [54], asthma [19, 55] and cancer and metastasis [9, 56]. This suggests that EMT induction by nickel exposure is likely a key mechanism associated with the pathogenesis of nickel-induced lung diseases. The following sections discuss molecular mechanisms underlying EMT induction by nickel exposure.

3. Transforming growth factor-β (TGF-β) signaling

TGF-β is a multifunctional cytokine that belongs to a superfamily of secreted factors. TGF-β is ubiquitously expressed and regulates a variety of physiological processes including embryogenesis, differentiation, cell proliferation, cytoskeletal organization and immune response [57, 58]. Aberrant TGF-β signaling is associated with a number of pathological conditions including autoimmune diseases, fibrotic diseases and cancer [57]. Mammalian cells have three TGF-β isoforms, TGF-β1, TGF-β2 and TGF-β3. TGF-β is synthesized as a large precursor, which is cleaved to release the active mature ligand in a dimeric form [57, 59]. Active TGF-β binds the type I (TGF-βR1) and type II (TGF-βRII) serine/threonine kinase receptors forming a heterotetrameric complex [57, 58]. Activation of the receptor complex by TGF-β leads to phosphorylation of SMAD proteins, which then translocate to the nucleus and regulate the expression of specific genes (Figure 1) [58, 60]. TGF-β signaling is also mediated by non-Smad pathways such as MAP Kinase, PI3K kinase, PP2A phosphatase and Rho GTPases [58, 61].

Figure 1. TGF-β signaling in nickel-induced EMT.

During TGF-β signaling, the active TGF-β binds the TGF-β receptors TGF-βR1 and TGF-βRII. The activation of the TGF-β receptor complex results in phosphorylation of the SMAD proteins, which then translocate to the nucleus and regulate target gene expression. Nickel exposure could activate TGF-β signaling via multiple mechanisms, including TLR4 signaling, TSP-1 upregulation and inhibition of H2S production. Nickel exposure induces TLR4 signaling. which has been shown to enhance TGF-β signaling. The extracellular matrix protein, TSP-1, an important activator of latent TGF-β, is one of the highest upregulated genes in the lungs of mice exposed to nickel, suggesting its role in nickel-induced TGF-β signaling. H₂S is an important negative regulator of TGF-β signaling. Nickel exposure promoted TGF-β signaling by downregulating the enzymes involved in H₂S production, CBS, CSE and 3MST, thereby increasing protein levels of TGF-β1 and phosphorylated SMAD2 and SMAD3.

TGF-β signaling plays an important role in the induction of EMT. TGF-β induces EMT during physiological processes such as embryogenesis and wound healing. However, under pathological conditions, TGF-β-induced EMT could be involved in tumor progression and metastasis, and the development of fibrotic diseases [60, 62, 63]. TGF-β signaling activates the EMT master regulators, SNAIL, ZEB and TWIST [62] and the mesenchymal markers, matrix metalloproteinases (MMPs), α-SMA, vimentin, fibronectin, and N-cadherin [64]. SNAIL, ZEB and TWIST repress E-cadherin expression by directly binding to the E-boxes at its promoter. [65–69]. SNAIL is also a repressor of claudins and occludin, which are involved in maintaining epithelial structure and function [70, 71]. In addition, SNAIL can also activate the expression of the mesenchymal proteins fibronectin, vimentin [72, 73] and N-cadherin [74].

Studies in mice exposed to nickel sulfate aerosol for 72 hours showed increase in TGF-β1 transcript and protein in the bronchoalveolar lavage fluid (BALF). Furthermore, expression of genes associated with TGF-β signaling and extracellular matrix function was significantly increased [46]. TGF-β1 repressed the mouse surfactant-associated protein B (Sftpb) via the TGF-β–responsive region at the Sftpb promoter [46]. Interestingly, Sftpb knockout has been shown to decrease the expression of E-cadherin and increase the expression of vimentin and fibronectin in the lung adenocarcinoma A549 cells. Furthermore, Sftpb knockout enhanced cell survival and migration, suggesting EMT [75]. This suggests that SFTPB downregulation by TGF-β induces EMT. Inhibition of TGF-β using the TGF-β type II receptor–IgG-Fc (TGFbRII-Fc) chimera decreased the nickel-induced total BALF protein levels in mice [46]. TGF-β inhibition also prevented NiCl₂ induced E-cadherin downregulation in both immortalized normal human bronchial epithelial BEAS-2B cells and the lung adenocarcinoma A549 cells [26].

Nano nickel oxide (nano-NiO) induced EMT through activation of TGF-β signaling [76]. The lungs of rats exposed to nano-NiO showed an increase in TGF-β levels, which was associated with pulmonary fibrosis [76, 77]. In addition, several fibrosis associated factors including Smad2 and Smad4 were overexpressed [76]. Furthermore, higher protein levels of collagen I and III were observed in nano-NiO exposed rat lungs, suggesting higher collagen deposition. Collagen I promotes EMT in lung cancer cells via activation of TGF-β3 signaling [78]. Similarly, A549 cells, as well as the human fetal lung fibroblasts exposed to nano-NiO, showed increased TGF-β levels [79]. Furthermore, the nano-NiO-treated A549 cells showed mesenchymal characteristics, including increase in the levels of α-SMA, vimentin and fibronectin and decrease in the levels of E-cadherin, suggesting EMT [79].

Interestingly, although both nano-NiO (20 nm particles) and micro-NiO (1 μm particles) exposure increased the expression of TGF-β and other fibrosis-associated factors in rat lungs, the levels of upregulation were higher in nano-NiO exposed lungs compared to those of the micro-NiO exposed lungs [76]. Furthermore, no increase in collagen I and III was seen in the lungs of rats exposed to micro-NiO. This suggests increased propensity of nano-NiO to cause lung injury and induce EMT, and thus is potentially a more serious health hazard than micro-NiO.

One of the endogenous negative regulators of the EMT process is hydrogen sulfide (H₂S) [80]. Endogenous H₂S plays a number of physiological roles and has been recently identified as a gasotransmitter that is at least as important as the other well-studied gasotransmitters, nitric oxide (NO) and carbon monoxide (CO) [80–82]. In mammalian cells, endogenous H₂S is produced through the metabolism of cysteine. The enzymes involved in this process include the two pyridoxal 5′-phospate- (PLP)-dependent enzymes: cystathionine β-synthase (CBS) and cystathionine γ-lyase (CSE); and the PLP-independent, 3-mercaptopyruvate sulfurtransferase (3MST)/cysteine aminotransferase (CAT) pathway [83–85].

H₂S attenuates EMT by inhibiting TGF-β signaling. Inhibition of TGF-β signaling by H₂S is involved in the anti-fibrotic mechanism in several tissues including the lung and kidney [86]. Several studies have shown that the exogenous H₂S donor, sodium hydrosulfide (NaHS), could inhibit EMT and airway remodeling in lung epithelial cells [82, 86–89]. NaHS attenuated EMT induced by the herbicide, paraquat (PQ), in A549 cells by inhibiting the TGF-β1/Smad2/3 signaling pathway [87]. Furthermore, a recent study suggested loss of H₂S-mediated inhibition of TGF-β-Smad signaling as a cause for NiCl₂-induced EMT in A549 cells [88]. NiCl₂-exposed A549 cells underwent EMT, as evidenced by the downregulation of E-cadherin and the upregulation of vimentin and increased migratory ability. Examination of the associated mechanisms showed that nickel exposure downregulated CBS, CSE and 3MST, the enzymes involved in H₂S production. Concurrently, an increase was detected in the protein levels of TGF-β1 as well as in the levels of phosphorylated SMAD2 and SMAD3. However, pretreatment of A549 cells with exogenous NaHS prevented NiCl₂-induced TGF-β1-Smad signaling [88]. These results suggest a novel mechanism of EMT induction by nickel exposure via impaired H₂S signaling (Figure 1).

Nickel is an activator of human TLR4 signaling, which triggers an allergic inflammatory response [90, 91]. NiCl₂ exposure induced TLR4 signaling in the lung cancer A549 and H1299 cells [8]. This study also uncovered the elevated expression of IL-8, TGF-β, MMP2 and MMP9 caused by nickel exposure and demonstrated that TLR4/MyD88 signaling was important in increasing the invasive potential of the cells [8]. TLR4 activation has been shown to enhance TGF-β signaling in hepatic fibrosis [92]. Based on these results, it is plausible that nickel-induced TLR4 activation could be a potential mechanism underlying TGF-β induction in nickel-exposed cells (Figure 1). However, although nickel is an activator of human TLR4, it does not activate mouse TLR4 due to the sequence variations between the human and mouse TLR4 [91]. This suggests existence of additional mechanisms of TGF-β activation by nickel.

TGF-β is produced as an inactive complex, which is activated by the degradation of the prosegments. The extracellular matrix protein, thrombospondin 1 (TSP-1), is an important activator of latent TGF-β [93, 94]. Interestingly, TSP-1 was one of the highest upregulated genes in the lungs of mice exposed to NiSO₄ aerosol [46]. This suggests that TSP-1 could be associated with TGF-β activation upon nickel exposure (Figure 1).

4. DNA methylation

5-methylcytosine (5mC) is an important DNA modification that plays principal roles in a variety of cellular processes. [95]. Nickel-induced aberrant changes to DNA methylation have been well-characterized in a number of studies [25, 96–102]. The DNA methylation alterations caused by nickel exposure is likely a consequence of its ability to inhibit the iron- and 2-oxoglutarate-dependent (2-OG) dioxygenases. Notable among the 2-OG dioxygenases are the Ten-eleven translocation (TET) family of DNA hydroxylases. Removal of 5mC occurs through the TET family dioxygenases, which mediate oxidation of 5mC to 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC) and 5-carboxylcytosine (5caC) [103]. The TET enzymes contain Fe(II) at their catalytic center. Ni(II), being highly similar to Fe(II) in charge and radius, can replace Fe(II) at the catalytic center of the Tet dioxygenases, thus making them highly susceptible to nickel-mediated inhibition [104].

NiCl₂ exposure downregulated E-cadherin in BEAS-2B cells. E-cadherin downregulation was associated with promoter DNA hypermethylation, and DNA methyltrasferase inhibitor 5-aza-2’-deoxycytidine restored its expression [25]. Furthermore, treatment of BEAS-2B cells with antioxidants or reactive oxygen species (ROS) inhibitors and scavengers reversed nickel-induced EMT, and pretreatment of the cells with the antioxidant N-acetylcysteine (NAC) inhibited nickel-induced E-cadherin downregulation. In addition, NAC treatment prevented DNA hypermethylation at the E-cadherin promoter. This suggests a role for ROS generated by nickel exposure in the hypermethylation of E-cadherin promoter [25]. ROS has also been shown to induce hypermethylation of E-cadherin promoter through the induction of SNAIL expression in Huh7 and Hep3B cells [105]. SNAIL directly binds the E-cadherin promoter and recruits the H3K9me2 methyltransferase, G9a, along with the DNA methyltransferase, DNMT1. SNAIL binding of the E-cadherin promoter is required for DNMT1 recruitment, and G9a functions as a bridge between SNAIL and DNMT1 [106].

Nickel sulfide exposure in mice repressed p16 gene expression through promoter hypermethylation [98]. The tumor suppressor p16 is an EMT repressor, which regulates EMT by transcriptionally activating miR-141 and miR-146b-5p, negative regulators of AUF1, an RNA binding protein. AUF1 stabilizes ZEB1 mRNA, thus increasing its protein levels and thereby functioning as an EMT inducer [107]. Ectopic expression of p16 in the highly invasive breast cancer cell line upregulated epithelial markers and downregulated mesenchymal markers [107] Therefore, it is plausible that nickel-exposure-induced p16 downregulation could contribute to EMT.

5. Histone modifications

Besides the TET family proteins, another category of major epigenetic regulators among the iron- and 2-oxoglutarate-dependent dioxygenases are the Jumonji C (JmjC)-domain containing family of histone demethylases [108]. In eukaryotes, DNA is wrapped around the four core histone proteins, H2A, H2B, H3 and H4, to form the fundamental structural unit of chromatin, the nucleosome. The N-terminal tails of histones are subject to a variety of post-translational modifications including acetylation, methylation, phosphorylation and ubiquitination [109–112]. Specific histone modifications are associated with the active or silent states of a particular genomic locus. For example, histone acetylation is associated with transcriptional activation, and histone methylation is associated with both gene activation and silencing. While histone H3 lysine 4 trimethylation (H3K4me3) is associated with gene activation, H3K9me3, H3K9me2 and H3K27me3 are associated with gene silencing [113, 114].

The JmjC domain histone demethylases (KDM2 to KDM7), target a number of both activating histone methylation marks, including H3K4me2/3 and H3K36me1/2/3, and silencing histone methylation marks, including H3K9me3 and H3K27me3 [115]. Similar to TET enzymes, the JmjC domain histone demethylases are sensitive to nickel exposure due to the ability of nickel to inhibit their activities by replacing iron at the iron-binding sites of these enzymes [116, 117]. Nickel inhibited the activity of the H3K9 demethylase, JMJD1A/KDM3A in a dose-dependent manner, with approximately one molecule of Ni(II) inhibiting one molecule of JMJD1A. The inhibitory effect was significantly amplified in the absence of iron. However, the addition of iron ions could not rescue JMJD1A activity, suggesting that Ni(II) binds the enzyme with a higher affinity as compared to Fe(II) [117].

SPRY2, a tumor suppressor gene, is an important suppressor of ERK1/2 activation, and its silencing is often observed in a number of human cancers especially in the metastatic stages [118, 119]. Nickel exposure inhibited SPRY2 expression in BEAS-2B cells, which was associated with increased H3K9me2 levels at its promoter [116]. SPRY2 repression in nickel-exposed BEAS-2B cells resulted in ERK activation and neoplastic transformation. JMJD1A overexpression and knockdown studies showed that JMJD1A, which directly binds the SPRY2 promoter was essential for its expression. SPRY2 repression is often observed in a number of human cancers. This suggests downregulation of SPRY2 through JMJD1A inactivation as one of the mechanisms underlying nickel carcinogenesis [116]. SPRY2 has also been demonstrated to negatively regulate TGF-β signaling and TGF-β-induced EMT by inhibiting ERK activation [118]. These studies suggest that nickel could induce EMT by inhibiting JMJD1A demethylase activity. Furthermore, nickel exposure caused spreading of the repressive H3K9me2 to adjacent active chromatin regions via inhibition of the DNA binding activity of the insulator binding protein CTCF, resulting in the downregulation of a number of genes [113]. It is likely that nickel-induced inhibition of H3K9me2 demethylase activity contributes to H3K9me2 spreading and aberrant gene silencing.

KDM6B/JMJD3 is a histone demethylase that mediates the removal of repressive histone modification H3K27me3. TGF-β exposure upregulated KDM6B expression in the mammary epithelial NMuMG cells, resulting in H3K27me3 demethylation at SNAIL1 promoter. This resulted in the transcriptional activation of SNAIL and induction of EMT [120]. Interestingly, NiCl₂ exposure increased KDM6B expression in the human embryonic kidney cell line HEK293T and renal carcinoma cell line 786-0. There was a concomitant decrease in the levels of H3K27me3 in both these cell lines [121]. Therefore, it is plausible that KDM6B upregulation by nickel could be associated with EMT induction.

Lysine-specific demethylase (LSD1/KDM1A) is a transcriptional repressor that demethylates H3K4me1 and H3K4me2 [122]. LSD1 is associated with the silencing of epithelial genes. LSD1 overexpression occurs in a number of solid tumors and levels of LSD1 expression correlates with tumor aggressiveness and EMT induction [122–127]. Nickel exposure has been shown to upregulate SNAIL in BEAS-2B cells [24]. SNAIL represses epithelial genes by recruiting LSD1, which recognizes H3K4 methylation through its interaction with the protein SFMBT1, resulting in its demethylation [128]. Interestingly, the protein levels of LSD1 remained unaltered due to nickel exposure. However, the levels of LSD1 acetylation significantly decreased [127]. In A549 cells, LSD1 protein is acetylated by the histone acetyl transferase MOF, which prevents its binding to the epithelial gene regulatory elements. Upon nickel exposure, MOF was significantly downregulated, resulting in LSD1 deacetylation. This caused increased binding of LSD1 at the epithelial genes, E-cadherin and KRT8, and caused H3K4me2 demethylation. Overexpression of MOF could counteract the nickel-induced EMT. Although MOF depletion facilitated nickel-induced EMT induction, in the absence of nickel exposure, depletion of MOF could not induce EMT in A549 cells. This suggests that in addition to MOF downregulation, other signals provided by nickel exposure are important for EMT induction [127].

NiCl₂-induced EMT in the lung epithelial BEAS-2B cells is associated with increased ZEB1 expression [24]. In the epithelial cells, the ZEB1 promoter is marked by both the activating histone modification, H3K4me3, and the repressive histone modification, H3K27me3, thus existing in a bivalent chromatin environment [24]. Bivalent genes are normally transcriptionally silent, but are considered to be in a poised state [129]. When appropriate signals are received, the promoter bivalency resolves to monovalency through loss of the activating or the repressing marks, leading to the corresponding gene expression changes [129]. Nickel exposure resolved the bivalent chromatin at ZEB1 promoter to a monovalent status by inducing loss of H3K27me3, resulting in active gene transcription [24].

ZEB1 is a negative regulator of E-cadherin [24]. The nickel-exposed BEAS-2B cells with high ZEB1 expression exhibit EMT characteristics, including downregulation of epithelial markers, upregulation of mesenchymal markers and metastatic phenotype [24]. Interestingly, the loss of H3K27me3 was stable, and chromatin environment at the ZEB1 gene did not revert to epithelial state even after the termination of nickel exposure, resulting in continued ZEB1 overexpression post exposure [24]. This resulted in persistent mesenchymal phenotype even after the termination of exposure. Nickel exposure caused persistent changes to the active histone modification H3K4me3 as well [130]. These results suggest that epigenetic alterations that persist after the cessation of exposure could be a common phenomenon associated with nickel exposure.

ZEB1 is a negative regulator of the microRNAs (miRNAs) associated with epithelial phenotype, miR-200 family and miR-205. On the other hand, miR-200/205 are negative regulators of ZEB1. Thus, ZEB1 and miR-200/205 exist in a double negative feedback loop [131, 132]. In BEAS-2B cells, ZEB1 overexpression due to nickel exposure caused downregulation of miR-200/205, which likely caused further increase in ZEB1 expression. Continued downregulation of miR-200/205 even after the termination of nickel exposure contributed to the continuous overexpression of ZEB1 resulting in the persistent EMT phenotype. However, ZEB1 knockdown after the termination of nickel exposure abrogated miR-200/205 downregulation and re-established the epithelial phenotype in BEAS-2B cells. These results suggest ZEB1 as an important mediator of nickel-induced persistent EMT [24].

Bivalent chromatin at ZEB1 gene has also been observed in the non-CSC population of the human mammary basal epithelial cells (HMECs). EMT signals such as TGF-β quickly converted the bivalent chromatin to an active chromatin configuration, and the cells enter CSC state [133]. These findings suggest that the bivalent chromatin configuration at the ZEB1 promoter enables its rapid upregulation upon receiving appropriate environmental cues. Nickel exposure could be one such environmental cue, which activates ZEB1 gene in order to trigger the wound healing process via EMT induction. However, persistent ZEB1 overexpression could result in pathogenesis. The molecular basis of irreversible alteration to the chromatin at ZEB1 promoter, which enables its persistent overexpression, remains to be investigated.

6. MicroRNAs (miRNAs)

MicroRNAs (miRNAs) are non-coding RNAs, which are 18–25 nucleotides in length. miRNAs negatively regulate gene expression either by targeting the 3′ UTR of target mRNAs and causing their degradation or translation inhibition. Regulation of EMT associated transcription factors by miRNAs are well documented. miR-203, miR-30a and miR-34 family downregulate SNAIL expression [134–136]. ZEB family genes are well-studied targets of miR-200 family and miR-205. TWIST1 is also a direct target of several miRNAs, including miR-15, -16, -186, -381 and -495 in various cell types [137–141]. Furthermore, miR-9 is a repressor of E-cadherin. In addition to the miRNAs that directly target the EMT transcription factors, several miRNAs that target other transcription factors such as MYC and SP1 are also described as pro- or anti-metastatic [134]. These studies suggest that deregulation of miRNAs could have significant consequences in the development of EMT-associated diseases and cancer metastasis.

Prominent among the EMT-regulating miRNAs are the miR-200 family and miR-205, which are negative regulators of EMT master regulators ZEB1 and ZEB2. miR-200/205 are downregulated by nickel exposure, which contributed to nickel-induced EMT [24]. In the nickel-exposed BEAS-2B cells, miR-200/205 downregulation is likely caused by ZEB1 upregulation, since ZEB1 knockdown restored the levels of miR-200/205 [24] (see section ‘Histone modifications’ for more information on ZEB1-miR-200/205 interaction). TGF-β treatment has been shown to upregulate several EMT associated transcription factors including ZEB1 and ZEB2 [142, 143]. Therefore, TGF-β signaling activated by nickel exposure could also contribute to miR-200/205 downregulation and EMT.

Upregulation of miR-21 is seen in multiple cancers, including cancers of the lung, and elevated expression of miR-21 promoted EMT [144, 145]. In A549 cells, miR-21 inhibition increased E-cadherin expression while miR-21 mimics decreased it [144]. Exposure to nickel nanoparticles (nano-Ni) caused upregulation of miR-21 in mouse lungs [53]. Nano-Ni exposure also caused increased expression of fibrosis-associated factors TGF-β1, p-Smad2, COL1A1, and COL3A1. Furthermore, severe lung inflammation and fibrosis were detected in the lungs [53]. In the miR-21 knockout mice, nano-Ni-induced upregulation of pSmad2, COL1A1 and COL3A1 was significantly diminished, although TGF-β1 levels remained unaltered. Furthermore, pulmonary inflammation and fibrosis decreased in miR-21-knockout mice suggesting a role for miR-21 in nano-Ni-induced EMT [53]. Though the mechanism of miR-21 overexpression by nano-Ni exposure remains unclear, studies have shown that it can be upregulated by TGF-β [146–149]. Therefore, TGF-β induced by nano-Ni exposure could be associated with miR-21 upregulation and fibrogenesis.

miR-152 is a negative regulator of DNA methyltransferase 1 (DNMT1) [150]. Downregulation of miR-152 is associated with EMT induction [151]. Nickel sulfide (NiS) exposure downregulated miR-152 in human bronchial epithelial 16HBE cells through promoter DNA hypermethylation. While ectopic expression of miR-152 inhibited cell proliferation in the NiS-transformed cells 16HBE, anti-miR-152 promoted it [152]. This suggests that NiS-induced miR-152 downregulation could potentially induce EMT.

7. Long non-coding RNAs (LncRNAs)

LncRNAs are non-coding RNAs, which are over 200 nucleotides in length. Recent studies have begun to uncover the role of lncRNAs in regulating EMT and metastasis. A number of lncRNAs such as HOTAIR, H19 and MALAT1 are identified as EMT promoters [153–157]. Some lncRNAs function as competing endogenous RNAs (ceRNAs), which impede binding of miRNAs to their targets. LncRNA-ATB upregulated ZEB1/2 and promoted hepatocellular carcinoma metastasis by functioning as a ceRNA for the miR-200 family members. Similarly, HOTTIP upregulated SNAIL and promoted EMT and metastasis in esophageal carcinoma cells by functioning as a ceRNA for miR-30b, a SNAIL repressor [158]. Furthermore, MEG8 suppressed the expression of MIR34a and MIR203 by facilitating EZH2 mediated H3K27me3 enrichment at their promoters, resulting in the upregulation of SNAIL [159].

Although a few lncRNAs, such as NRG1 [160] and MEG3 [161], are implicated in nickel-exposure-induced tumorigenesis, their role in nickel-induced EMT has not been thoroughly investigated. Nickel exposure downregulated MEG3 through promoter DNA hypermethylation, resulting in malignant transformation of BEAS-2B cells [161]. MEG3 overexpression reversed the nickel-induced cell transformation. Lung tissues from lung squamous cell carcinoma (SCC) also showed a significant downregulation of MEG3, suggesting a role for MEG3 inhibition in lung tumorigenesis [161]. A study on gastric cancer tissues also suggested the anti-tumorigenic effects of MEG3. Interestingly, this study showed that MEG3 upregulated E-cadherin while downregulating the mesenchymal markers N-cadherin, Snail and β-catenin [162], thereby decreasing the cell’s ability to invade and migrate. Moreover, this study showed that the anti-EMT effect of MEG3 was mediated by its ability to negatively regulate miR-21, a promoter of EMT [162].

MEG3 is also downregulated by nano-NiO exposure [163]. Nano-NiO exposure induced EMT, causing collagen deposition in A549 cells and pulmonary fibrosis in rats in a TGF-β dependent manner. MEG3 overexpression decreased TGF-β1 expression and suppressed EMT. Therefore, MEG3 downregulation has been implicated in the EMT induction and fibrosis by nano-NiO exposure [163]. Contrarily, MEG3 has also been shown to be essential for TGF-β-induced EMT in the lung cancer cell lines, A549 and LC-2/ad, where it recruits repressive epigenetic marks to the regulatory regions of E-cadherin and MIR200 [164].

8. Hypoxia-inducible factor-1 (HIF-1) signaling

Cellular response to low oxygen is mainly mediated by HIF-1 signaling [165]. HIF-1 is a heterodimeric transcription factor consisting of a hypoxically inducible α subunit and a constitutively expressed β subunit. Under normoxia, the proline residues in the oxygen-dependent degradation domain (ODDD) of HIF-1α are hydroxylated by prolyl-hydroxylases (PHDs). The hydroxylated proline residues are recognized by von Hippel-Lindau protein (pVHL), which mediates HIF-1α ubiquitination, leading to its proteosomal degradation [166]. Hypoxia prevents HIF-1α hydroxylation due to inhibition of PHD activity, resulting in lack of interaction with pVHL and HIF-1α stabilization. The stabilized HIF-1α translocates to the nucleus, dimerizes with HIF-1β and activates target genes by binding the hypoxia-response elements (HREs) at their promoters [167–169] (Figure 2). HIF-1α is stabilized in solid tumors due to low oxygen availability and is associated with a number of cancers [170–173].

Figure 2. Activation of HIF-1 signaling in nickel-exposed cells contributes to EMT.

Under physiological conditions, HIF-1α is rapidly degraded via hydroxylation of its oxygen-dependent degradation domain by prolyl-hydroxylases (PHDs). Upon nickel exposure, Ni(II) replaces Fe(II) at the iron-binding site of PHDs causing their inactivation. This leads to the stabilization of HIF-1α, which translocates to the nucleus, dimerizes with HIF-1β and binds the HREs at the promoters of target genes, resulting in their activation. HIF-1 could directly activate EMT master regulators, SNAIL, ZEB and TWIST, by directly binding their promoters. HIF-1 could also upregulate lysyl oxidase genes Lox and LoxL2, which interact with SNAIL and cause E-cadherin repression.

PHDs are 2-oxoglutarate (2-OG)-dependent dioxygenases, which require Fe(II) and ascorbate as cofactors [174]. Ni(II) can compete with Fe(II) and replace it at the iron-binding site of PHDs causing their inactivation. This results in HIF-1α stabilization [117, 175, 176]. Therefore, nickel exposure stabilizes HIF-1α even under normoxia [170]. HIF-1 has been well-documented to regulate a large number of cellular processes including cell survival, proliferation, motility, metabolism, extracellular matrix function and angiogenesis [177–180]. Emerging evidence suggest a major role for HIF signaling in the induction of metastasis. A number of studies have demonstrated the role of HIF-1 in inducing EMT in multiple cell-types [181–184]. EMT can be induced by increased HIF-1α expression either due to hypoxia or due to the constitutive expression of HIF-1α mutant containing a deletion in the ODD domain [185, 186]. Moreover, HIF-1α knockdown in hypoxic cells reversed both the expression of EMT markers and the metastatic ability [185, 187, 188].

The effect of HIF-1 signaling on EMT is mediated by a number of downstream targets of HIF-1 such as the EMT master regulators, SNAIL, ZEB and TWIST. HIF-1 directly targets these EMT factors by binding the HREs at their promoters and activating their expression (Figure 2) [184–186, 189–191]. Suggesting major roles for the EMT master regulators in mediating HIF-1 signaling, ZEB1 inhibition abolished HIF-1α-induced EMT in metastatic colorectal cancer cells [189]. Similarly, TWIST knockdown in lung and hypopharyngeal cancer cells abolished the EMT phenotype and inhibited EMT induced by hypoxia or HIF-1α overexpression [185, 186]. Inhibition of SNAIL by shRNAs reduced HIF-1α-induced EMT in gastric CSCs [192]. These studies also suggest that the downstream effectors of HIF-1-induced EMT could be cell-type specific. In addition, HIF-1 can also regulate EMT markers in an indirect fashion. HIF-1α promoted EMT and fibrogenesis in mice renal epithelial cells by upregulating lysyl oxidase genes Lox and LoxL2 [193]. The lysyl oxidases have been shown to functionally interact with SNAIL, causing E-cadherin repression in MDCK cells [194].

Stabilization of HIF-1α by nickel compounds is well documented [25, 170, 195–197]. NiCl₂ exposure significantly increased the protein levels of HIF-1α, which was associated with E-cadherin repression in BEAS-2B cells [25]. ShRNA-mediated HIF1α knockdown in BEAS-2B cells diminished NiCl₂-induced E-cadherin downregulation and fibronectin upregulation, suggesting a role for HIF-1 signaling in NiCl₂-induced EMT in the lung [25]. Nano-NiO exposure increased the levels of HIF-1α and TGF-β1 in human fetal lung fibroblasts, which contributed to the development of pulmonary fibrosis [77]. Exposure of human fetal lung fibroblast cells to HIF1-α inhibitor 2-deoxy-d-glucose (2-DG) reduced the nano-NiO-induced overexpression of TGF-β1. On the other hand, TGF-β-Smad inhibitor treatment diminished the nano-NiO-induced increase in HIF-1α protein levels [77]. This suggests a synergistic interaction between the HIF-1 and TGF-β signaling pathways in the development of nickel-induced pulmonary fibrosis [77]. Although the role of HIF-1 signaling in nickel-induced EMT has not been extensively investigated, given the well-established role of HIF-1 signaling in EMT induction, it is reasonable to speculate that HIF-1α stabilization by nickel exposure could play a major role in EMT induction.

9. Conclusions and future perspectives

Evidence from recent studies clearly indicate the importance of EMT induction in the etiology of lung diseases associated with nickel exposure. TGF-β is an important regulator of a number of pathogenic processes [198]. It is clear that the activation of TGF-β signaling by nickel exposure is an important event in nickel pathogenesis. Although some mechanisms such as impaired H₂S production [88] and activation of TLR4 signaling [8] could begin to explain the underlying mechanisms, further investigation is necessary to fully understand this critically important process. Similarly, HIF-1 signaling, which could directly regulate the expression of EMT associated factors including ZEB, TWIST and SNAIL, is an important EMT inducer. The mechanisms underlying nickel-mediated HIF-1α stabilization is well understood. Future studies will reveal the importance of HIF-1 signaling in nickel-induced lung diseases.

In addition to TGF-β and HIF-1 signaling, epigenetic deregulation constitutes an important phenomenon that is associated with nickel-induced EMT. The mutagenic potential of nickel compounds is low and accumulating evidence suggest aberrant epigenetic changes as major drivers of nickel toxicity. Consequently, nickel-induced epigenetic changes have been extensively characterized by a number of studies [113, 130]. However, epigenetic changes related to EMT induction is only beginning to be understood. Although inactivation of 2-OG dioxygenases potentially plays a major role in the aberrant epigenetic changes caused by nickel exposure, the mechanisms associated with nickel-induced dysregulation of the epigenome are not fully understood. Bivalency resolution via loss of H3K27me3 at the ZEB1 promoter plays an important role in EMT induction by nickel. However, the underlying mechanisms remain unclear. Bivalent chromatin, which poises genes for activation upon receiving appropriate stimuli, is present in all cell types. However, it is more prevalent in the embryonic stem (ES) cells, where the developmental gene promoters are marked by both active and repressive epigenetic marks [111, 129, 199]. EMT is involved in a number of early developmental processes such as gastrulation and neural crest formation. Therefore, it is reasonable to speculate that nickel exposure could have significant deleterious consequences during early development, and this warrants in-depth investigation in the future. Thus, while the current studies have shed important light on the contribution of EMT to nickel-induced lung diseases, there are a number of open questions and exciting avenues for future research.