MiRNAs regulate the expression of genes involved in biological processes relevant to the progression of chronic lung disease, including cellular stress, cell differentiation and apoptosis [30–32]. Here, we aimed to identify miRNAs (and their target mRNAs) associated with emphysema severity in COPD patients. Five miRNAs were identified to be significantly down-regulated in lung from patients with moderate emphysema, compared with lung of mild emphysema patients. In addition, a negative correlation was noted in the expression of these miRNAs and their putative target mRNAs ex vivo. Furthermore, increasing the expression of miR-34c (the miRNA with the largest fold change between the mild and moderate emphysema patient groups) in respiratory cells resulted in decreased expression of predicted mRNA targets in vitro, providing additional functional expression data. Our data suggest modulation of SERPINE1 by miR-34c in vitro and ex vivo. Dysregulation of SERPINE1 by miR-34c could therefore be a potential mechanism involved in emphysema severity and progression. To our knowledge, this is the first study to compare miRNA profiles in the lungs of COPD patients based on emphysema severity.
A major strength of this study was the ability to correlate miRNA-mediated mRNA regulation in the same lung samples. It is known that miRNAs mediate gene silencing through translational inhibition and mRNA degradation. In addition, target mRNA is degraded in more than one third of genes that display translational repression ; hence we have focused on studying changes in mRNA expression as a consequence of altered miRNA expression. MiRNAs that are down-regulated in a certain disease state (e.g. moderate emphysema, compared to mild emphysema) would be expected to upregulate expression of their predicted mRNAs, in an orchestrated process of epigenetic regulation. This relationship of miRNA and mRNA expression has been demonstrated previously in airways from COPD patients  but not in peripheral lung tissue, which we used in this current study. Of the miRNAs identified as differentially expressed in lung between moderate vs mild emphysema patients in this study, miR-34c is particularly relevant to human lung disease, as its expression is increased during normal lung development . We found decreased expression of miR-34c in lungs of patients with moderate emphysema, compared to mild emphysema. Van Pottelberge et al. also found greater than three-fold downregulation of miR-34c in sputum from smokers with COPD, compared to those without COPD, and a direct correlation between miR-34c expression and percent predicted FEV1. Their findings support our study, since there was down-regulation of miR-34c expression in their COPD patients, compared to controls, and also down-regulation in our moderate emphysema patients, compared to mild emphysema, indicating a similar continuum of expression with increasing disease severity. In the setting of reduced miR-34c expression in moderate emphysema lung (vs mild), we were able to confirm increased gene expression of two of its predicted mRNA targets, HNF4 and SERPINE1. HNF4A (hepatocyte nuclear factor 4A) is a transcription factor that regulates the expression of cytokines involved in inflammation, as demonstrated in other organs (e.g. kidney ); its role in the lung is yet to be determined. SERPINE1 (Serpin Peptidase Inhibitor, Clade E (Nexin, Plasminogen Activator Inhibitor Type 1), Member 1), also known as plasminogen activator inhibitor 1 (PAI-1), is a protease inhibitor, and inhibitor of fibrinolysis. PAI-1 knockout mice demonstrate emphysema-like changes in the lung , so our findings of increased SERPINE1 expression in moderate emphysema (vs mild emphysema) may be unexpected, given the role of protease-antiprotease imbalance in the pathogenesis of emphysema. However, SERPINE1 is only one of a number of antiproteases, and has other roles in the lung that could be relevant to the progression of emphysema. For example, SERPINE1 has been shown to be upregulated in the lung or lung cells during hypoxia , exposure to lipopolysaccharide or cigarette smoke extract , or oxidative stress via NF-κB . Importantly, elevated levels of PAI-1 in sputum have been observed in two separate studies of patients with COPD (compared to controls) [39, 40]. Furthermore, sputum PAI-1 levels were higher in more severe COPD (compared to milder COPD), which would be in keeping with our findings of higher SERPINE1 expression in more severe emphysema . Other candidate emphysema severity miRNAs identified in this study – miR-34b, miR-133a/b and miR-149 – have been previously implicated in the pathogenesis of lung diseases. For example, miR-34b expression was decreased in induced sputum from COPD smokers compared to non-COPD smokers . Down-regulation of miR-133a is associated with increased bronchial smooth muscle contraction in patients with asthma . miR-133b negatively regulates the expression of transcription factor, Pitx3, which activates the dopamine receptor (DRD1), a mediator of nicotine addiction in smokers . Overall, emerging evidence of the biological roles of these miRNAs supports the plausibility of their involvement in lung diseases.
A number of potential limitations of this study should be addressed. (i) A physiological measure of emphysema (gas transfer) was used to stratify patients, as in our previous study , rather than pathological examination of lung tissue. The reason for this was that these lung tissue samples were not inflated at harvesting, making morphological assessment of emphysema severity difficult. However, gas transfer is routinely used clinically as a measure of emphysema severity in COPD patients. (ii) Although a relatively small number of samples was used (from 29 patients), the FDRs of the 5 miRNAs were all <0.05; furthermore, the two miRNAs tested (miR-34c and miR-133a) were technically validated independently using qRT-PCR. (iii) No overlap was found between the miR-34c targets that were enriched in ex vivo lung samples and in vitro cell lines. It should be noted that the ex vivo study used Operon microarrays which represented only 14,000 genes. Furthermore, our in vitro study used commercial cell lines, and all of our COPD lung samples were from patients who also had lung cancer. Despite these potential limitations, we found that the expression of five candidate genes in vitro correlated with ex vivo profiles in both in-house  and external datasets (Spira et al.).