Background
Asthma is a chronic inflammatory condition of the airways that affects around 300 million people worldwide [
1]. The airway epithelium, derived from the endoderm, is the first structural barrier to the inhaled environment in the airway mucosa. In asthma, the airway epithelium has an altered phenotype displaying altered cell cycle kinetics and increased numbers of basal cells [
2,
3]. In addition, the lamina propria of asthmatic donors contains resident fibroblasts, derived developmentally from the mesoderm, that have been shown to exhibit an invasive and synthetic phenotype [
4‐
8]. How these alterations in cellular phenotype occur in the disease is unknown but it is clear from the many genomic studies that asthma involves both genetic and environmental components.
The epigenetic landscape is essential in determining cell fate through histone modification and DNA methylation patterns that regulate the expression of genes integral to cellular development and differentiation [
9‐
11]. Covalent modifications of the histone N-terminal tails can regulate gene expression and include acetylation, methylation, phosphorylation, and ubiquitination [
12,
13]. Histone acetylation and phosphorylation are associated with a more open chromatin structure and gene expression, whereas histone methylation and ubiquitination can work both in a gene repressive and expressive manner depending on the target residue [
13‐
17]. The enzymes responsible for the addition/removal of these modifications include: histone acetyltransferases (HATs)/deacetylases (HDACs), protein kinases/phosphatases, histone methyltransferases (HMTs)/demethylases (HDMs), and ubiquitin ligases/deubiquitinating enzymes (DUBs) [
13,
17]. DNA methylation is facilitated by DNA methyltransferases (DNMTs) that add a methyl group to cytosine bases, forming 5-methylcytosine (5-mC) [
12]. Addition of this mark at a gene promoter is generally associated with transcriptional repression and gene silencing [
12,
18]. Furthermore, the epigenome is adaptable; it has the capability to respond to and be modified by environmental factors [
10]. The outcome of this interaction depends on the environmental stressor and can be a normal physiological response or deregulation of the epigenome producing an abnormal phenotype [
10,
19].
Abnormal epigenetic control of gene expression has been identified in both fibroblasts and epithelial cells in numerous pathologies [
20‐
25]. However, very little is known about the expression and regulation of epigenetic modifying enzymes in asthma. Indeed, dysregulation of epigenetic mechanisms in asthma has been identified in a variety of cells but most studies have been performed in tissues from outside of the lung [
26]. While dysregulation of enzymes involved in histone acetylation was identified in the airways of asthmatics, there is still disagreement on the exact enzymes responsible [
27‐
30]. We have additionally identified unique DNA methylation patterns in airway epithelial cells (AECs) from asthmatic donors [
31] yet research on the variability of the enzymes responsible for these changes is lacking.
To further elucidate the mechanisms driving the epigenetic alterations observed in the asthmatic airways, a better understanding of the gene expression profiles of epigenetic modifying enzymes in airway tissues is required. We hypothesize that the gene expression of epigenetic modifying enzymes is cell-specific and influenced by asthma status in tissues derived from the airways. Specifically, the aim of this study was to identify if the expression profiles of epigenetic modifying enzymes is cell- and disease-specific by profiling 82 genes across 5 families of epigenetic enzymes in AEC and fibroblasts from healthy and asthmatic donors. We identified 24 cell-specific and 7 disease-specific differentially expressed genes (6 in AECs and one in fibrolasts). Although not all of the disease-specific genes passed correction, we were able to identify a corresponding change in AURKA protein expression in asthmatic compared to healthy individuals.
Discussion
This is the first study to evaluate the gene expression levels of histone and DNA modifier enzymes in AECs and airway fibroblasts derived from human lung tissue. We found significantly higher expression for 24 of these enzymes in AECs compared to airway fibroblasts from healthy individuals. Further, we demonstrate that AURKA is differentially regulated in AECs from asthmatic compared to healthy donors. In addition, we identified a corresponding increase in AURKA protein expression in AECs from asthmatic compared to healthy donors, further supporting our findings. Even though AECs and fibroblasts reside in close proximity within the airway mucosa, the function of each cell is very different. These data support the notion that epigenetic modulation of gene expression may be important for cell type specificity, and may potentially influence susceptibility to diseases such as asthma.
Multiple studies have documented differential DNA methylation in relation to tissue and cell specificity, and how this is altered in diseases [
31,
39‐
41]. Yet very few studies have focused on the global expression of enzymes responsible for DNA methylation expression. In our study, DNMT3a and MBD2 were both elevated in AECs compared to airway fibroblasts. DNMT3a is not only integral for mammalian development but also responsible for
de novo DNA methylation [
42]. It is possible that the elevated DNMT3a seen in AECs may reflect the cell’s geographical position. The airway epithelium is constantly in contact with external environmental factors thus must be responsive and adaptable to incoming stimuli. Elevated DNMT3a allows the cell to methylate genes
de novo in response to these environmental stimuli. The increased expression of MBD2 may be a response to the increase in DNMT3a as MBD2 is a transcriptional repressor which binds methylated DNA [
43]. To further support this theory, the complex which MBD2 forms to repress gene expression is not strongly bound to the DNA [
43] suggesting a transient visit as would be expected from a responsive reaction.
The outcome of an epigenetic change can be variable depending on the particular modification that occurs. Methylation of lysine and arginine residues on histone tails is facilitated by enzymes which are specific to both residue and site yet the outcome can activate or repress transcription [
13]. In contrast, histone acetylation, commonly associated with gene expression, is regulated by enzymes that have been described as promiscuous in their substrate specificity [
14]. We identified differential expression of enzymes involved in both histone methylation and acetylation in AECs compared to airway fibroblasts. Of the 6 enzymes involved in histone methylation, half target the activating mark H3K4me; SETD3 methylates while KDM5B and KDM1A demethylate H3K4. This may indicate that AECs preferentially utilize H3K4 methylation over others to control gene expression. A similar observation was seen with histone acetylation as 5 HATs and 6 HDACs were identified. Three of the HDACs that were elevated in AECs comprise 75% of the class I HDAC family of enzymes important in controlling proliferation, differentiation, and tissue development programs [
44]. Higher expression of the majority of the class I HDAC family of enzymes in epithelial cells may be a reflection of their considerable specialization as they have the capacity to differentiate and develop into a variety of epithelial cell types, which requires manipulation of the processes mentioned above.
We found elevated expression of 3 histone kinases and 2 DUBs when we compared AECs to airway fibroblasts. Although histone phosphorylation is commonly associated with gene activation, histone ubiquitination can result in both permissive and repressive states depending on the residue. However, all of the resulting histone modifications from the 5 above enzymes are associated with gene expression. This suggests there may be an imbalance in the regulation of these activating marks in AECs, potentially indicating lower levels of cellular transcriptional activity in airway fibroblasts compared to AECs.
Through its interaction with β-catenin, CREBBP has recently been identified as a pivotal component of the machinery maintaining an undifferentiated and proliferative state [
45]. Inhibition of this interaction facilitates β-catenin and EP300 pairing which is thought to control cell differentiation [
45,
46]. Our findings of decreased gene expression of CREBBP in AECs from asthmatics may indicate a divergence away from a proliferative state towards an initiated, but incomplete differentiation pathway. This imbalance of proliferation/differentiation mechanisms may contribute to the phenotypically immature epithelium seen in asthmatic airways.
In the context of disease, aurora kinases have been linked to spermatogenic arrest, chromosomal instability, and tumorigenicity in pathologies such as infertility, chronic inflammation, and a wide range of cancers [
47‐
49]. AURKA is capable of phosphorylating H3S10, a site implicated in both gene activation and cell division [
15,
50]. In a murine model of wound repair, rapid and sustained phosphorylation of H3S10 was associated with wound healing in intestinal epithelial cells [
51]. Further, although the mechanism is not fully clear, phosphorylation of H3S10 is a critical component of chromatin compaction during mitosis [
52]. Given that AECs from asthmatics are mitotically dyssynchronous [
53], show defects in cell cycle regulation [
54], and exhibit abnormal proliferation and delayed wound repair [
55‐
57], our finding of increased AURKA expression may indicate aberrant regulation of these processes in asthma.
We identified elevated mRNA expression of the histone methyltransferase SMYD3 in airway fibroblasts from asthmatics. SMYD3 is integral to cell cycle regulation through interactions with RNA polymerase II and methylation of H3K4 [
58]. In addition to gene activation through H3K4 methylation, SMYD3 is capable of gene repression through H2K20 methylation [
59], suggesting a complex role for this enzyme. However, although differences in gene expression were seen, we were unable to replicate these findings at the protein level possibly indicating a further level of transcriptional control.
While we found many cell-specific and some disease specific changes in the enzymes involved in epigenetic modification, there are limitations to our study. We used a cell culture model that does not necessarily represent the complexity of cell – cell interactions known to be integral to airway mucosal homeostasis. However, a cell culture model allowed us to identify differences in the epigenetic modification families in relatively undifferentiated epithelial cells and fibroblasts under controlled conditions. Although we examined gene and protein expression of the epigenetic modifiers associated with asthma, we did not assess the activity of these enzymes, which has been shown to differ in disease. In addition, due to the sample size, we were unable to examine sex differences within our samples. Lastly, we did not look at the targeted epigenentic changes as a result of the differential expression of the epigenetic modifying enzymes and further studies would need to be performed to solidify the functional effects of the cell and disease specific changes we described in our cohort.
Acknowledgements
Not applicable.