We have previously published that EMT could be an active process in human smokers' airways,  based on the demonstration of Rbm fragmentation and cleft formation and the presence of cells in both Rbm and basal epithelium staining for established markers of EMT [12, 15] such as S100A4 and MMP-9. These changes were present in airway biopsies from asymptomatic smokers with normal lung function, but were most pronounced in currently smoking COPD patients. This confirmatory study, although with limited numbers, focussed on this most relevant clinical phenotypic group, but is likely to be generalizable also to smokers with normal lung function.
Although we have provided rare, human data in human COPD, suggesting the potential for EMT,  this conclusion has been open to reviewer scepticism that the markers of EMT that we used, although regarded as 'classic' in the literature, [12, 15] were not qualified with further markers, especially showing that the Rbm cells express epithelial markers as well as mesenchymal markers. It is also been suggested that the Rbm cell data could be confounded by infiltration from below by inflammatory cells, which could at least potentially be positive for S100A4 and MMP-9. Our COPD subjects in this study were on as needed "relief" bronchodilator medication only, and it is unlikely that the discussed appearances are due to confounding by therapy, which we have deliberately tried to avoid as much as possible.
In the present study we have followed up our previous findings by demonstrating that cells in the Rbm positive for mesenchymal markers did frequently stain with an epithelial marker. Further, the present study illustrates that there are cells in the Rbm that double stain for cytokeratin-(s) and the "EMT marker" S100A4, representing further evidence of EMT in smoking-related COPD. Furthermore, our staining of tissue comprehensively for markers of a variety of inflammatory cells indicates that our data were not confounded by significant infiltration of the Rbm by inflammatory or immune cells.
Although only a relatively small number of cells in the Rbm could be demonstrated to stain with cytokeratin, it is possible that more (indeed most) of these Rbm cells are actually of epithelial origin, as it is well documented that during development of EMT, epithelial cells lose epithelial markers as they gain mesenchymal markers. Therefore, a cell which has gone through a complete transition will not express epithelial markers. It is, however, reassuring from the point of view of our hypothesis about active EMT in the airways in smokers and COPD, that we have been able to demonstrate at least a reasonable number of double-stained cells in the Rbm. Similarly, the frequent presence of cells in the basal epithelium staining for S100A4 is further evidence of EMT in these COPD airway biopsies. The data are consistent with a model where migratory cells gain the S100A4 mesenchymal marker early, but then rapidly loose epithelial markers as they migrate through the Rbm.
We found that there were very few CD8+ and CD4+ T lymphocytes and no CD19 B-cells in the Rbm and basal epithelium. Cells positive for CD19 were mainly in the lamina propria and hardly detectable in the basal epithelium or Rbm. We also confirmed that our Rbm cellular S100A4 staining was not confounded by infiltrating dendritic cells or Langerhans cells for which sections were stained for CD11c (dendritic cell/inflammatory cell marker)  and S100 (human Langerhans cell marker) . Although they were certainly present, there were relatively few such cells in the basal epithelium and Rbm compared to S100A4-stained cells. Most of the cells positive for the inflammatory cells studied were observed in the lamina propria below the Rbm, and not in the Rbm itself. EMT is a process defined by loss of epithelial markers and gain of mesenchymal markers. A function of this is that there is no single definitive marker for EMT. S100A4 is widely published as representing a reasonable marker for EMT [8–12]. We have also shown that these S100A4 positive cells in the Rbm also double-stain for another mesenchymal protein, vimentin,  widely used as an EMT marker [12, 15].
EMT has only recently been recognised in the human lung or airway, [6, 8] and now we provide further confirmatory evidence for its active presence in COPD airways . However, EMT is well described in lung embryogenesis,  metastatic malignant disease  and as part of the repair process in renal disease following tissue injury . Active EMT is indicated by the degradation of underlying epithelial basement membrane and transition of epithelial cells into mesenchymal cells with migratory potential, such that they move away from the epithelium in which they originated into deeper tissue [12, 15].
Active airway EMT in COPD could potentially relate to subsequent fibrotic activity in the sub-epithelial tissue, or just reflect severity of COPD pathology. COPD is a complex disease with physiologically a mixture of a variable emphysema component and intrinsic airway narrowing. Pathologically the airways are always abnormal and it is this component we have sampled from large airway biopsies. Airway obstruction is thought to be mainly in the small airways, but the pathology affects the whole bronchial tree [2–5]. There are in fact very few data available on fibrosis or alterations in fibroblast/fibrocyte populations in the airway in COPD. Our data taken together suggest that remodelling changes can be more exaggerated in established COPD and especially in current smokers [28, 29]. This area of pathology needs more specific study and remodelling changes need relating to changes in airway calibre and stiffness. Even if there are no immediate physiological consequences, there may well be implication for other COPD-associated processes and especially airway malignancy where the concurrence of EMT may be a significant part of tumour pathogenesis.
EMT has been described in metastatic malignant disease especially when associated with S100A4 expression. Epithelial cell nuclear expression of S100A4 strongly correlates with both active EMT and metastatic disease in the oncology literature [8, 30, 31]. Increased expressions of S100A4 and MMP-9 are observed in human non-small cell lung cancer (NSCLC) and have significant correlations with clinical and biological behaviour of such cancer cells [8, 30, 31]. Epidermal growth factor receptor (EGFR) is also over-expressed in many types of cancers, including NSCLC and we have shown that this is co-expressed in the epithelium as part of the whole 'COPD phenotype signal' . It has been shown that clinical responsiveness to EGFR inhibition using the "biologic" erlotinib is directly linked to the degree of EMT in NSCLC . Our findings of smoking-related active EMT may help understanding of why lung cancer is so common in smokers, but especially so in those with COPD where we have found EMT to be most marked.
Our study is of course limited by the fact that it was carried out in living human subjects and is based on 'static' pictures. Further study may require COPD model systems in animals and/or human cell culture. This might allow direct, vital imaging to observe dynamic real-time changes in cell phenotype and cellular migration during EMT.
It has been argued that Rbm fragmentation in smokers and COPD biopsies in our studies represents an "artefact" due to tissue processing. Against this is the fact that clefts contained cells, and also, to some extent vessels . Further, tissues from normal healthy controls  and asthmatics (data not shown) were processed in exactly the same way, but rarely showed any such Rbm fragmentation.