The incidence and severity of fibrotic lung diseases increase with age, but very little is known about how age-related changes affect the mechanisms that underlie disease emergence and progression. Normal ageing includes accumulation of DNA mutations, oxidative and cell stresses, mitochondria dysfunction, increased susceptibility to apoptosis, telomere length dysfunction and differential gene expression as a consequence of epigenetic changes and miR regulation. These inevitable ageing-related phenomena may cause dysfunction and impaired repair capacity of lung epithelial cells, fibroblasts and MSCs. As a consequence, the composition of the extracellular matrix changes and the dynamic interaction between cells and their environment is damaged, resulting ultimately in predisposition for several diseases. This review summarizes what is known about age-related molecular changes that are implicated in the pathobiology of lung fibrosis in lung tissue.
Ageing is an irreversible, gradual and systemic deterioration of cellular functions, resulting in increased vulnerability to environmental assaults and subsequently to an increased susceptibility of the elderly body to chronic diseases or even death. It is believed to result from the accumulation of molecular and cellular damage that cannot be repaired by the aged cells, due to evolved limitations in the performance of the cellular mechanisms that ensure somatic maintenance and repair. The lack of a specific cellular ageing programme and the stochastic infliction of damage are in agreement with the perplexing variety of cellular mechanisms that appear to be involved in ageing . All cells suffer upon damage, but proliferating cells are infinitely more vulnerable as they undergo intense cycles of DNA replication. In many cases, DNA damage in highly proliferating cells is counteracted by apoptosis or cellular senescence to avoid the potential propagation of harmful mutations [2, 3]. This phenomenon may play a very important role in the occurrence of abnormal wound healing and fibrosis in the lung as it affects the proliferation of both lung and circulating progenitor cells that are involved in the repair process [4-6].
Lung development in humans starts as early as 4 weeks of gestation, with the alveolarization process lasting at least until 7 years after birth. The complex course of human lung development is tightly regulated by interconnected networks of proliferating cells, growth factors, matrix components and physical forces [7-9]. As individuals age, their lungs undergo several physiological changes, including anatomical, functional and molecular alterations that ultimately result in a progressive decrease of alveolar gas diffusion capacity . Apart from the physiological changes that inevitably occur with age, certain pathological conditions seem to thrive later in life, such as idiopathic pulmonary fibrosis (IPF) and chronic obstructive pulmonary disease (COPD), leading to association of the ageing mechanisms with the underlying pathobiology of these diseases [11-14].
Restoration to an intact epithelium after lung injury is imperative for proper lung function. There are many agents known that can cause injury to lung epithelium, among which are bacterial or viral infections, allergic reactions, inflammation, trauma, exposure to certain chemicals and cancer, while there is also injury of unknown aetiology. As a response to lung injury, several repair mechanisms are activated in order to restore lung homeostasis (reviewed in ). Briefly, lung injury is followed by an acute inflammatory response and a mobilization of epithelial and circulating progenitor cells, which move to damaged areas and differentiate into the appropriate cell types for tissue restoration. The repair response involves a plethora of molecules, including growth factors, chemokines, interleukins, prostaglandins, integrins, matrix components, matrix metalloproteinases, a variety of signalling pathways (including STAT3, Wnt, Rho GTPases and MAP kinase) and the application of biomechanical forces. In contrast to the early response to an acute injury, overcoming a persistent assault relies heavily on the type of injury and the local microenvironment of the wound. For example, cigarette smoking causes recurring cycles of injury and repair in the airway and alveolar epithelial cells that is associated with the emergence of COPD, whereas a persistent injury to the alveolar–capillary membrane is associated with fibrosis [16, 17]. Since both of these diseases associate with ageing, it is reasonable to assume that ageing, by compromising the repair mechanisms, leads to abnormal wound healing and fibrosis. This review focuses on major mechanisms that are known to be implicated in wound healing and are affected by developmental stage or age.
Lung diseases and ageing
Epidemiological studies indicate that ageing is associated with an increased incidence of a variety of chronic lung diseases affecting a significant portion of the population [12, 13]. For example, the occurrence of IPF increases in both prevalence and incidence in the sixth decade of life. Symptoms typically occur at age 50–70 years, and most patients are > 60 years of age at the time of clinical presentation . Currently, only limited therapeutic strategies exist to treat chronic lung diseases, most of which are symptomatic instead of causal.
Our understanding of the biology of ageing has advanced remarkably, although the molecular mechanisms linking ageing to chronic lung disease remain unclear. Cellular senescence, oxidative stress, abnormal shortening of telomeres, apoptosis and epigenetic changes affecting gene expression have been proposed to contribute to the ageing process and ageing-associated diseases . Animal studies also support the link between ageing and susceptibility to diseases by demonstrating an increased vulnerability of the aged lung to injury. It has been published by our group that a single episode of lung injury by bleomycin  or lung infection with γ-herpes virus  cause severe progressive pulmonary fibrosis only in naturally aged wild-type mice when compared to young mice. These observations, in some way, challenge the traditional concept that progressive fibrosis is the result of chronic injury.
A combination of several age-related physiological changes might be responsible for a broad spectrum of abnormalities that could lead to different lung diseases with a fibrotic phenotype. In this paper we review the possible complications of age-related: (a) modifications of the extracellular matrix; (b) increase in the apoptosis of alveolar epithelial cells; (c) accumulation of senescent stem cells; (d) telomere dysfunction; and finally (e) epigenetic changes, which individually or in combination can result in alterations of the mechanisms of lung homeostasis and repair, leading to abnormal wound healing and fibrosis.
Extracellular matrix changes induced by ageing
The extracellular matrix (ECM) provides structural support by serving as a scaffold for cells, and as such the ECM maintains normal tissue homeostasis and mediates the repair response following injury. Tension applied through collagen fibrils at the ECM–cell interface might lead to protein synthesis, cell mitotic activity and changes in gene expression via activation of MAPK phospho-relay systems .
Collagen and elastin, the main proteins of ECM, form the scaffold of the alveolar structure and determine the mechanical properties of lung parenchyma. Collagen represents 15–20% of the total dry weight of the pulmonary tissue, with type I and type III collagen adding up to 90% of the total amount. Another protein, fibronectin, forms fibrils that are connected to other matrix components and has been implicated in cell adhesion, migration, epithelial–mesenchymal transition, phagocytosis and cell growth .
The composition of ECM changes during ageing and these alterations undoubtedly might contribute to the physiological decline of lung function. In general, in ageing connective tissue, the collagen type I content of the ECM is increased, whereas the collagen type III content is decreased , along with elastin fibres . The proteoglycan content also appears to decrease with age . The exact mechanism of how the age-dependent changes in ECM components affect lung repair is still unclear, although it is known that fibronectin expression increases in clinical and experimental models of fibrosis [26, 27], which in turn suggests an association with the repair process. In injured lungs, during the early phase of active repair, fibronectin production increases dramatically and this augmentation coincides with fibroblast proliferation. These fibroblasts are responsible for the excessive synthesis and deposition of collagen proteins . Alterations in cell–fibronectin interactions may contribute to abnormal tissue remodelling by stimulating the proliferation of fibroblasts , myofibroblast differentiation  and epithelial–mesenchymal transition, or by just facilitating the deposition of other matrix components, such as collagens . However, it is known that fibronectin undergoes alternative splicing at each of the three fibronectin exons and the ratios of the alternatively spliced mRNAs are affected by age and by the levels of growth factors . Lungs from ageing mice also show a significant increase in the fibronectin isoform extra type III domain A (EDA), combined with an increased susceptibility to bleomycin injury , thus suggesting a connection between age, fibronectin and fibrosis.
As mentioned above, fibronectin EDA is considered necessary for TGFβ1-induced myofibroblast differentiation  and is up-regulated by TGFβ1 through the PI3K–Akt–mTOR and SMAD3 signalling pathways . Recently it was documented that an age-depended increase of TGFβ1 and TGFβR1 expression in old lungs coincides with increased Smad3 mRNA and protein expression. Furthermore, it was shown that Smad3 is phosphorylated and capable of DNA binding only in aged lungs and that it can up-regulate downstream targets, such as PAI-1 . Based on the observation that fibronectin EDA is elevated in patients with IPF , whereas EDA deficiency protects mice from fibrosis , it is reasonable to propose that excessive expression of fibronectin EDA, associated with age, in the lung might promote fibrogenic responses in the setting of lung injury.
Other components of the ECM that might have important implications in the ageing lung are the expression levels and activity of matrix metalloproteinases (MMPs) and their inhibitors, TIMPs. Comparative studies between young and old mice show a significant increase in Mmp2, Mmp9 and Timp2 mRNA expression in old lungs. However, only Mmp9 activity was found to be enhanced in old lungs, probably associated by the low expression of its inhibitor, Timp1. The increase in MMP expression could lead to increased susceptibility to injury, leading to increased leukocyte migration and more tissue damage in the injured lung .
Taken together, these observations indicate that ageing leads to changes in the expression of TGFβ and ECM composition that might have important implications in the lung repair process.
Aged mesenchymal stem cells in repair
Bone marrow-derived cells can be divided into haematopoietic (B-HSC) and mesenchymal (B-MSC) stem cells. The accurate characterization of B-MSCs has been a complicated issue, since there are no specific cell surface markers. Enrichment of B-MSCs from crude bone marrow suspensions is achieved by selection for a plastic-adherent population that expresses neither haematopoietic nor endothelial cell surface markers but is positive for the expression of adhesion and stromal markers . In lack of a defined panel of unambiguous markers distinguishing B-MSCs, a criterion for establishing B-MSC phenotype is to use adherent cells isolated by cell sorting that: (a) express CD44, CD73, CD90 and CD105; (b) lack the expression of haematopoietic markers such as CD45, CD34 and CD31; and finally (c) in a tri-lineage differentiation assay, confirm their plasticity by the ability of the cells to differentiate into adipocytes, osteocytes and chondrocytes . An interesting aspect of B-MSC biology is that they can migrate to the lung, responding to a lung injury [35, 36], and participate in lung repair by modulating the immune response and the mechanism of lung repair . It is known that B-MSCs can traffic in vitro into the lung after bleomycin treatment  and differentiate into lung epithelial cells [35, 36] under the proper stimuli. Both bone-marrow stem cell populations are implicated in the pathogenesis of IPF and both change with age. Although the majority of studies are focused on B-MSCs' role in wound healing and fibrosis, probably due to their therapeutic potential, there is a small group of adherent B-HSCs (fibrocytes) that are implicated in the pathobiology of pulmonary fibrosis. Fibrocytes express stem and leukocyte cell markers, such as CD45 and CD34, that can traffic to the lungs in response to Cxcl12 in a bleomycin injury murine model and can produce type I collagen [39, 40]. High levels of circulating fibrocytes have been associated with age-related susceptibility to lung fibrosis in a mouse bleomycin model  and poor prognosis in IPF . A third stem cell population, lung-resident mesenchymal stem cells (luMSC), is also implicated in the protection from lung fibrosis through inhibition of T cell proliferation after injury , but it is not known how this function is affected by ageing.
B-MSCs from elderly people have different morphology, increased production of ROS and oxidative damage , DNA-methylation changes affecting cell differentiation , slower proliferation rate in culture [45, 46] and shorter telomeres , and a large proportion of them stain positive for senescence-associated β-galactosidase . Ageing mice are characterized by a senescence-related increase in fibrocyte mobilization (and a parallel decrease in B-MSCs) and higher serum levels of Cxcl12. B-MSCs from old mice are characterized by a quiescent state with low metabolic activity and are primarily in the G0 phase of the cell cycle. This quiescent state is maintained by both extrinsic and intrinsic mechanisms and has been postulated to be a way of preserving their long-term proliferative potential and genomic integrity. The problem that arises is that quiescent B-MSCs escape DNA damage checkpoints and several repair pathways that are cell cycle-dependent, and that results in the accumulation of DNA damage during ageing, ultimately leading to rapid stem cell depletion or exhaustion. The critical effect of accumulation of DNA damage in ageing MSCs has recently been demonstrated, using a progeroid mouse model with defects in DNA repair, due to a deficiency in the excision repair cross-complementation group1 (Ercc1) protein. Adoptive transfer of wild-type muscle derived-MSCs in the Ercc1-deficient mice was sufficient to halt premature ageing and increase lifespan . In an insightful study, Conboy and colleagues showed that an old mouse with declining organ stem cell capacity had undergone a rejuvenation of aged muscle and liver progenitor cells when it was surgically joined with a young mouse in order to share circulatory systems . Several studies since have demonstrated that both physiological ageing and pathological senescence can affect these functions. In a rat model for cardiomyopathy, human MSCs from aged donors did not perform as well as ones from young donors . MSCs from old donors fail to differentiate in vitro into neuroectodermal cells , and early passage MSCs are more efficient in promoting the proliferation and maintenance of haematopoietic progenitor cells . In one case, the administration of stem cells from young mice restored cardiac angiogenesis in senescent mice when stems cell from old mice failed to do so . The superiority of very young B-MSCs can be explained by several aspects of their biology . Briefly, B-MSCs of fetal origin express the pluripotency stem cell markers (Oct-4, Nanog, Rex-1, SSEA-3, SSEA-4, Tra-1-60 and Tra-1-81), have longer telomeres and greater telomerase activity and express more human telomerase reverse transcriptase. Fetal MSCs were also more readily expandable and senesced later in culture than their adult counterparts.
In conclusion, ageing and extensive in vitro culture of B-MSCs defines telomeric length, pluripotency, proliferating potential and, overall, their ability to execute their regenerative role [54, 55].
‘Ageing’ telomeres and fibrosis
Telomeres [from the Greek words ‘telos’ (end) and ‘meros’ (part)] are special chromosomal end-structures (TTAGGG repeats) of various lengths that are required to ensure chromosomal stability and correct cell division [56-58]. Due to the ‘end-replication-problem’, telomeres shorten with each cell division and eventually reach a critical length that is no longer functional . Telomere shortening, by way of either ageing/replicative senescence [60, 61] or DNA-damageing agents [62-66], resembles double-stranded DNA breaks and thus activates a DNA damage response that results in senescence or apoptosis . Telomere length can be affected by several normally occurring age-related phenomena, such as elevated levels of reactive oxygen species (ROS)  and multiple cell divisions , but it can also be affected by factors that are unrelated to ageing but have a cumulative effect over the years, such as cigarette smoking  or obesity .
The combination of exhaustion in the progenitor cell reservoir and the progressive accumulation of senescent cells could account for the organ decline associated with ageing . Ageing somatic cells cannot maintain telomere length but their replicative capacity can be restored by ‘engineering’ telomeres with the over-expression of telomerase reverse transcriptase (hTERT), a subunit of the telomerase enzyme [71, 72]. On the contrary, human fetal tissues show high telomerase activity, presumably due to high numbers of undifferentiated progenitor cells . Although telomerase mutations lead to dysfunctional telomerase and the consequent acceleration of telomere shortening that is normally observed with ageing, short telomere length, independently of a functional polymerase, is sufficient in causing a degenerative disease similar to dyskeratosis congenita .
According to current data, up to 15% of familial [75-77] and 3% of sporadic IPF cases  are associated with mutated telomerase genes. In the study by Alder and colleagues, an astounding 97% of idiopathic interstitial pneumonia (IIP) patients were found to have shorter telomeres in their peripheral blood and alveolar epithelium in comparison to the median length of aged-matched control subjects . In an independent study, 20% of familial and sporadic cases of pulmonary fibrosis without a mutation in the telomerase genes showed significantly (low 10th percentile of the controls) decreased telomeres . Remarkably, these patients were indistinguishable from the rest, leading to the conclusion that telomere shortening alone can cause IPF. Shorter telomeres in alveolar epithelial cells type II (AEC2) of IPF patients coincide with lower levels of polymerase and increased apoptosis markers , suggesting that AEC2 cells from these individuals cannot cope with the increased demand of proliferation that is necessary for repopulating the injured alveolar epithelium surface . In a recent study  involving heterozygotes for TERT mutations and their families, it was clearly demonstrated that pulmonary fibrosis is age-dependent (not present in carriers aged < 40 years but reaching 60% in carriers of ≥ 60 years), in line with an age-dependent reduction of telomere lengths in the circulating leukocytes. Another interesting outcome of this study was that telomere length is affected even in those family members that do not carry a mutation, suggesting a pattern of inheritance for shortened parental telomeres.
To conclude, telomere shortening is a component of the age-related predisposition to idiopathic organ fibrosis, and individuals with shorter telomeres may have increased risk of developing idiopathic interstitial lung disease.
Apoptosis, wound healing and age
Apoptosis is an evolutionarily conserved cell suicide programme that is strictly regulated and executed through finely controlled signalling pathways. Apoptosis is essential for embryogenesis, development and the maintenance of tissue homeostasis in adult individuals, including cellular turnover, control of proliferative expansion of cells or response to virus infection . Apoptosis is executed by the concerted action of a set of cysteine proteases known as caspases, which are activated by several converging pathways. The extrinsic pathway is induced by the stimulation of members of the tumour necrosis factor-receptor family; the cellular stress pathway involves the alteration of mitochondrial permeability and subsequent cytochrome c release, and the intrinsic apoptosis pathway is induced by direct DNA damage, hypoxia or survival factor deprivation .
Regardless of the activation pathway, apoptosis plays a role during injury and wound healing. Higher apoptotic responses increase the magnitude of injury and the possibility of forming scar tissue. On the other hand, apoptosis is necessary for the removal of inflammatory cells during wound healing. In IPF, a paradox of apoptosis occurs, with enhanced apoptosis of type II lung epithelial cells in combination with resistance to apoptosis in fibroblasts and myofibroblasts. This opposite receptiveness to apoptosis is believed to be critical in the pathogenesis of lung fibrosis . A potential factor in the age-related susceptibility to IPF is the accelerated apoptosis of type II lung epithelial cells. Increased apoptosis of irreplaceable postmitotic cells and loss of function has been described during ageing for heart, muscle and brain . Ageing human lungs show loss of epithelial cells and homogeneous enlargement of the alveolar spaces. During injury, enhanced apoptosis of alveolar epithelial cells may result in ineffective re-epithelialization of a damaged alveolar wall, promoting a fibrotic tissue-repair response . Studies by our group demonstrated that type II lung epithelial cells from ageing lungs in mice are highly susceptible to apoptosis during injury, and that increase in apoptosis was associated with expression of markers of endoplasmic reticulum (ER) stress, BiP and Xbp1, that were not expressed in the injured lungs of young mice . Recent evidence suggests that the ER may represent a critical organelle for the induction and regulation of apoptosis. Under stress conditions, such as age-associated oxidative damage to the Ca2+ ATPase pump, ER can induce cytochrome c release from mitochondria, thus initiating apoptosis in muscle cells . In addition, ER can release procaspase-12, which, upon activation by m-calpain, cleaves and activates caspase-3 .
Defective wound healing in the ageing lung can also be associated with increased apoptosis of epithelial cells by extrinsic pro-apoptotic pathways. Data from animal models show that activation of the Fas pro-apoptotic pathway, as well as TGFβ over-expression, causes apoptosis of lung epithelial cells [90, 91]. Fas-mediated apoptosis effectors and TGFβ have been found to increase with age in several organs, including skin and lung [19, 92]. There is evidence suggesting that increase in the release and activation of TGFβ from alveolar epithelial cells could be associated with oxidative stress. Oxidative stress, provoked by the excess of free radicals, is considered to be among the major factors contributing to ageing. In ageing, systemic imbalance among the sophisticated antioxidant system (eg superoxide dismutases, glutathione) and ROS results in the generation of excess in free radicals that overwhelms the cellular antioxidant defences. The consequences, among others, may include direct damage to DNA and apoptosis . Interestingly, oxidative stress has been also associated with IPF. Thus, for example, mitochondrial generation of ROS has been suggested to be linked to increased cellular oxidative stress and apoptosis of alveolar epithelial cells . Moreover, the commonly used murine model of bleomycin-induced pulmonary fibrosis is characterized by an early induction of ROS-dependent alveolar epithelial cell apoptosis, which is impeded by caspase inhibitors . However, it is unknown whether excessive oxidative stress associated with ageing increases the risk of developing IPF.
It is also known that mitochondrial dysfunction promotes apoptosis and may be a central mechanism driving mammalian ageing . At the mitochondrial level, apoptosis is orchestrated by the release of mitochondrial proteins that regulate pro-apoptotic molecules, such as caspases. The maintenance of mitochondrial function during stress conditions is established by the balance between pro-apoptotic and anti-apoptotic members of the Bcl-2 family of proteins, as well as levels of heat shock proteins (HSPs), which are a family of chaperones that are up-regulated following exposure to environmental stress and help to stabilize and refold protein intermediates, or facilitate the degradation of proteins that become irreversibly damaged. Furthermore, they prevent the induction of apoptosis by interacting with several apoptotic mediators, including Bax, Bak, cytochrome c and caspase-3 . Over-expression of HSPs is reported in keloid tissue, associated with increased with proliferation of fibroblasts and matrix synthesis . However, mice over-expressing HSP-70 show amelioration of bleomycin-induced lung fibrosis , possibly associated with protection against apoptosis of lung epithelial cells.
Several groups have reported age-associated alterations in the expression of members of the Bcl-2 family of proteins, as well as of various HSPs. For example, Chung and Ng found a significant increase in levels within the gastrocnemius muscle in old rats , which might represent a compensatory mechanism in response to a pro-apoptotic cellular environment. These investigators have also shown a significant increase of Bax and a decrease of Bcl-2 expression in the gastrocnemius muscle of old rats, with no change in the expression of Bak . In the heart, the anti-apoptotic mitochondrial protein Bcl-2 shows a strong tendency to decrease with age . Because Bcl-2 can prevent cytochrome c release, these observations provide a potential mechanism underlying the increase in apoptosis observed in ageing tissues. In IPF, low levels of Bcl-2 are observed in lung epithelial cells, but fibroblasts have high levels of HSP and Bcl-2. This is in agreement with the paradoxical response to apoptosis observed in the fibrotic lung .
In conclusion, both the association of apoptosis with age-related susceptibility to injury in the lung and other organs and the cell/tissue specificity of this process warrant further investigation.
Epigenetic changes related to ageing and fibrosis
‘An epigenetic trait is a stably heritable phenotype resulting from changes in a chromosome without alterations in the DNA sequence’ . In concordance, epigenetics is ‘the study of the mechanisms of temporal and spatial control of gene activity during the development of complex organisms’ . There are three main mechanisms of epigenetic regulation: (a) DNA methylation; (b) histone modifications (acetylation, methylation, ubiquitinylation, phosphorylation and sumoylation); and (c) non-coding RNAs. It is important to understand that all these mechanisms coexist and cross-react to form a complex network of transcriptional and post-transcriptional regulators of gene expression. Various epigenetic changes have emerged as key mechanisms by which cells change during development and cellular differentiation, and as a response to environmental stimuli and stress. Tight control of this network is necessary for all biological processes and any dysregulation of the network is associated with disease [104, 105] and ageing .
DNA methylation involves the addition of a methyl group to the 5 position of the pyrimidine ring of a CpG dinucleotide by DNA methyltransferases (DNMTs) . Clusters of CpGs form CpG islands or CpG island shores, whose methylation status can modify the activity of quiescent transposable elements or the expression of genes through either direct blocking of transcription factors or chromatin remodelling . In general, hypermethylation of promoter-associated CpG islands results in transcriptional silencing, while hypomethylation results in transcriptional activation. Nevertheless, there are some exceptions, such as the promoter of the human telomerase catalytic subunit (hTERT), where the hypermethylation of the promoter CpGs is associated with higher gene expression . Irregular methylation of CpGs is a trait of multiple diseases, including lung cancers and IPF [110, 111].
The amino-terminal parts of histones, and especially lysine and arginine residues, are sites for a variety of post-translational modifications that affect their interactions with DNA, the other histones, chromatin-binding proteins (transcriptional factors, RNA polymerase, etc.) and chromatin-remodelling complexes. A lot of studies focus on the methylation and acetylation of the N-terminus lysines of histones H3 and H4, while research on ubiquitylation and sumoylation is still limited. The methylation of histones is catalysed by histone lysine methyltransferases (HLMTases). Acetylation and deacetylation are catalysed by histone acetyltransferases (HATs) and histone deacetylases (HDACs), respectively. Usually, acetylation of histones in promoters and other control elements marks transcriptionally active genes, while methylation is a hallmark of transcriptionally inactive loci. However, there are many exceptions to the above rule, eg H3K27me3 is associated with silencing but H3K4me3 is associated with transcriptional activation. Histone modifications are often interdependent. Methylated H3K9 can inhibit H3 acetylation at several lysine residues . Histone modification and DNA methylation are also interdependent and can act in unison to maximize their impact on chromatin regulation. For example, DNMTs can bind HDACs to achieve histone deacetylation and repress transcription . Furthermore, histone modification status can define the de novo methylation of local DNA [114, 115], while DNA methylation can be used as a pattern for reconstructing histone modifications after DNA replication .
Because of the complexity and large volume of epigenetic activity that a cell has to deal with during its lifespan, there is a significant possibility of errors that would impact chromatin architecture and gene expression. Dysregulation of epigenetic processes can underlie several diseases  and be associated with ageing . Earlier studies on the epigenetic changes during ageing which suggest a pattern of gradual global demethylation of CpG islands , and multiple studies since, have supported the hypothesis that ageing is associated with a general relaxation of epigenetic control. Once a critical number of epi-mutations is accumulated, the cell, tissue or organ starts to malfunction. This is in concurrence with the ‘dys-differentiation’ theory that was proposed by Cutler, suggesting that ageing for cells is, in reality, a time-dependent drifting away from their proper state of differentiation . A recently emerged paradigm in ageing suggests that ageing may also derive from the declining multipotency of adult stem cells, which undergo not only changes in their total numbers and irreversible alterations but also suffer a broad spectrum of epigenetic alterations, including DNA methylation and histone modification changes [119, 120].
Despite the huge volume of publications that address the association of ageing with several epigenetic modifications, very little has been published in relation to wound healing and fibrosis. In a recent high-throughput approach on DNA methylation by Rabinovich and colleagues, it was shown that the lung genomes of IPF patients have a DNA hypomethylation profile that clearly distinguishes them from controls . The genes that were associated with the differentially methylated CpG islands are involved in apoptosis, morphogenesis and cellular biosynthetic processes. Through genome-wide approaches that provide an invaluable global view of disease-related molecular alterations and gene-focused approaches, there is hope to elucidate the molecular pathways that define the pathobiology of fibrosis. Gene-focused studies have shown that Thy-1 cell surface antigen, which suppresses myofibrobastic differentiation of lung fibroblasts, is epigenetically down-regulated in fibroblastic foci by hypermethylation  and histone modifications , while Cyclooxygenase-2 (COX-2, PTGS2), regulating the production of the anti-fibrotic prostaglandin E2, is epigenetically down-regulated by hypoacetylation of the COX-2 promoter in lung fibroblasts from IPF patients . Likewise, the antifibrotic angiostatic chemokine γ-interferon (IFNγ)-inducible protein of 10 kDa (IP-10) is down-regulated in lung fibroblasts from patients with IPF by histone hypoacetylation and increased histone H3 methylation, facilitated by recruited histone methytransferases and heterochromatin protein 1 (HP-1) . Furthermore, methyl CpG-binding protein 2 (MeCP2) is implicated in myofibroblast differentiation and pulmonary fibrosis, by regulating the expression of the α-smooth muscle actin (α-SMA) gene through binding to the methylated CpG promoter islands , and pro-apoptotic p14 (ARF) is down-regulated in many IPF fibroblasts through promoter hypermethylation, presumably resulting in their increased resistance to apoptosis . Clearly there is a lot be done before there is an accurate evaluation of the impact of epigenetic modifications on fibrosis.
MiRNAs are short (∼22 nt) endogenous non-protein-coding ribonucleotides that, despite their abundant and ubiquitous nature, escaped discovery until the early 1990s . Through a mechanism of antisense complementarity to the 3′UTR site of their target mRNAs, they regulate gene expression by promoting mRNA cleavage or by repressing translation, both mediated by the RNA-induced silencing complex (RISC) . On rare occasions they can promote post-transcriptional gene up-regulation . The pairing of a mature miRNA with its target mRNA is achieved by a specific seed region located at nucleotides 2–8 on the miRNA . MiRNAs form an extremely flexible regulatory system, which can be fine-tuned on multiple levels. One miRNA can regulate hundreds of genes and one gene can be regulated by multiple miRNAs. The formation of the miRNA/mRNA duplex depends on several parameters, such as the accessibility of the miRNA binding sites, RNA secondary structure, the position of the binding site in the 3′UTR and its distance from the stop codon, and reciprocal interference between different miRNAs and RNA binding proteins sharing the same target .
MicroRNAs are very abundant in a variety of organisms and cell types and their importance as regulatory elements of all possible aspects of human development and physiology is steadily rising. A comprehensive description of all miRNA functions is beyond the scope of this review, but to mention a few, miRNAs are involved in stem cell differentiation, haematopoiesis, cardiac and skeletal muscle development, neurogenesis, insulin secretion, cholesterol metabolism and the immune response . There are 1600 sequences identified as miRNAs in the human genome (http://www.mirbase.org/cgi-bin/mirna_summary.pl?org=has ). A remarkable aspect of miRNA biology is the observed diversity of distinct expression patterns, depending on the cell type and the developmental stage of that particular type. Notably, not all miRNAs are created equally, with some reaching very high copy numbers while others remaining fairly rare per cell, comprising a profile that is tightly controlled on a transcriptional and post-transcriptional level.
During mammalian development, miRNA expression undergoes significant shifts . In a study performed on lungs from newborn, post-weaning and adult female mice, 14 miRNAs of the 484 examined varied to a statistically significant extent . In another study, 44 miRNAs were found to be differentially expressed between neonatal and 2 month-old mice , while little difference was documented for miRNA expression profiles between the lungs of 6 month-old and 18 month-old mice, a condition that may contribute to tissue homeostasis. That suggests that in adults any significant shift from that expression profile would signify pathological conditions . Contrarily, in a study with human PBMCs, the majority of the miRNAs studied so far had decreased with age . In the same study, 21 miRNAs were found to be differentially expressed between fetal and adult human lungs. Interestingly miR-26b, which is characterized as pro-fibrotic in liver fibrosis , is increased in adult lungs, while the anti-fibrotic miR-335  is only expressed in newborns. Mir-154, which is highly expressed in neonatal mouse lungs and fetal human lungs, is also over-expressed in IPF , suggesting the implication of developmental pathways in the pathogenesis of the disease, probably during wound healing [141, 142].
In a recent review, Vettori and colleagues  have gathered from the literature over 40 miRNAs that have been linked to fibrosis in various organs and diseases. The dysregulation of many miRNAs can be limited in a specific organ or disease, although there are cases where a certain miRNA can be dysfunctional in several fibrotic disorders. Most of these miRNAs directly induce or protect from fibrosis by targeting TGFβ canonical and non-canonical pathways [144-146], connective tissue growth factor (CTGF) [147, 148] and ECM structural proteins or enzymes involved in ECM remodelling [149-151]. Some of them are involved in mechanisms that indirectly affect fibrogenesis, such as epithelial–mesenchymal transition (EMT) [152, 153] or the proliferation and resistance to apoptosis in myofibroblasts . Several of these miRNAs (see Table 1) show a different level of expression from fetal or neonatal to young adult lungs [134, 155], which is imperative for proper lung function, as shown for let-7d . Of the 46 microRNAs that were differentially expressed in the lung of IPF patients , 18 showed age-dependent expression, with significant differences between lungs from early developmental stages and adults (Table 1). More importantly, their expression pattern in IPF is reminiscent of the expression pattern in fetal and neonatal stages, arguing in favour of the pathological re-activation of developmental processes in IPF .
Table 1. Age-dependent expression, in lung tissue, of microRNAs that are implicated in fibrosis (column 1) and IPF (column 6)
The growing number of miRNAs and the regulation of their expression offer an exciting and challenging layer in the control of lung fibrosis, with great potential for therapeutic intervention.
The process of ageing is characterized by progressive functional impairment and reduced capacity to respond appropriately to environmental stimuli and injury. Although the incidence and severity of fibrotic diseases increase with age, there has been relatively little investigation into the mechanisms of the effects of ageing on susceptibility to these disorders. Normal ageing mechanisms include cumulative DNA mutations, oxidative and cell stresses, mitochondria dysfunction, increased susceptibility to apoptosis, telomere length dysfunction and differential gene expression as a consequence of epigenetic changes and miRNA regulation (Figure 1). These natural ageing processes cause dysfunction and impaired repair capacity of lung epithelial cells, fibroblasts and MSCs. As a consequence, the composition of the extracellular matrix changes and the dynamic interaction between cells and their environment is damaged. A better understanding of these ageing biological process and repair responses may provide novel and better therapeutic approaches for these devastating diseases.
The authors would like to thank Marta Bueno for her help with the design of the figure.
All the authors contributed to the revision of the literature, writing of the manuscript and approval of the final version.