The expression and functional activity of uPAR is of interest in both cancer and respiratory disease. The current study aimed to characterise the key regulatory regions of UPAR, identify all common splice variation and quantify expression in the normal lung and in specific lung and peripheral cells at the mRNA and protein levels. We have identified a localised TSS with some cell specificity and confirmed expression of uPAR in lung and peripheral cells. At the mRNA level, multiple uPAR splice variants were identified including alternative E7 (E7b) and deletions of E3, E5+6 and E6, and their patterns of expression in different tissues/cell types characterised. Primary peripheral cells (PMN and PBMC) expressed multiple exon deleted forms of membrane bound uPAR, whilst lung cells including epithelium and airway smooth muscle expressed a greater proportion of an alternative soluble uPAR with and without an exon 4+5 deletion. Protein analyses confirmed expression of multiple differentially expressed forms of uPAR in all cell extracts, and soluble uPAR was detected in the supernatants of cultured HASM, HBEC, THP1 and BEAS2B cells. Specificity of the Western blotting analyses was confirmed by siRNA. Our data provide a novel insight into the molecular mechanisms that potentially regulate uPAR expression and activity in the airways and the periphery which has implications regarding the potential role of uPAR in airway disease biology.
Previous analyses in cell lines suggested the TSS of UPAR is 52 bp upstream of the ATG, with a minor start site around -77 bp in U937 (monocytes) and HeLa (epithelial) cells . We have now confirmed and extended these initial analyses using tissues and cells of the respiratory and immune systems; namely lung, HASM, HBEC, PMN and PBMC. In all samples the TSS was located between -42 and -77, however there were clear cell specific locations within this interval i.e. -46 to -52 is the dominant TSS in most cells and lung tissue but not in HBEC where the -75 to -78 locus is dominant. The functional significance of the cell specific TSS usage remains to be resolved although it is interesting to note that functional studies using promoter-reporter approaches have suggested that most transcriptional activity is found in the first 220 bp upstream of the ATG in HeLa, HCT116 and RKO (colon cancer) cells [29–31]. Functional studies have implicated AP1 (-122) and SP1-like motifs (spanning -139 to -154) as being important for basal uPAR transcription in colon cancer and HeLa cells [29, 30]. We used data-mining to characterise relevant transcription factor binding sites in the 4 kb promoter region identifying a number of different potential pro-inflammatory transcription factor binding sites including AP1, STAT, GATA and CCAAT/EBP sites using two or more databases. Only a single AP1 (-236) site and no SP1 sites were replicated, despite their known importance, indicating that some sites may be missed using our replication strategy, however, the AP1 (-122) and SP1 (-151) sites identified in functional studies were detected by individual databases. uPAR expression has been shown to be regulated by a number of growth factors at the transcriptional level, which include TGFβ1 utilizing the AP1 site at -236 . TGFβ1 is a pro-fibrotic cytokine that has been implicated in airway remodelling in asthma and therefore it is tempting to speculate that TGFβ1 induction of uPAR may be a significant mechanism involved in airway remodelling.
The second objective of the current study was to determine uPAR splice variation in airway and peripheral cells using real-time PCR. Two donors were used for each cell type (but not tissues), to allow for inter-individual differences. We observed expression of the previously identified uPAR E7b soluble splice variant , which is predicted to result in the loss of the GPI anchor, in the lung and periphery. Interestingly the primary peripheral cells expressed low levels of this variant (as a fraction of total uPAR) compared to the lung tissue and cells. This suggests that peripheral cells could retain more of their uPAR at the cell membrane (GPI-bound) whilst lung structural cells produce more soluble, secreted uPAR. Binding of uPA to membrane bound uPAR initiates many intracellular signalling pathways, resulting in outcomes including differentiation, proliferation and cell motility , whilst soluble uPAR can act as a chemoattractant for hematopoietic cells . It is therefore possible that airway cells might express more soluble uPAR to allow them to attract cells into the airways, whilst the peripheral cells express mainly the surface form of the receptor to allow them to respond to external signals.
In addition to the exon7 variants we also identified a series of internal exon deletion variants expressed within the context of both the membrane bound and soluble receptor. These included exon 3, 4+5, 5+6 and 6 deletions. Exon deleted forms of classical uPAR were found most frequently in the primary peripheral cells, PMN and PBMC. Exon 6 deletion was the most widely expressed. Loss of exon 6 would disrupt the structure of D3, particularly as it includes one of the key cysteines involved in disulphide bonding. This would be expected to lead to reduced uPA binding affinity  as two amino acid residues shown to be involved in uPA binding are located in this exon . Expression of an exon 4+5 deleted variant of uPAR has previously been described and was shown to be associated with shorter disease-free survival in breast cancers using a real-time PCR strategy , however the assay used did not distinguish between exon 7 variants. In the present study, we did not detect this deletion in combination with classical exon 7 using real-time PCR (although a single clone was obtained from PBMC by PCR), whereas in combination with alternative exon 7b this variant was detected in most tissues/cell types, notably the structural cells HASM, HBEC and BEAS2B. Based on the known structure of uPAR, this variant might express a soluble form of the receptor lacking the key chemotactic sequence  and D2, which includes integrin binding domains . Some uPA binding ability conferred by D1 may be retained  although many uPA binding residues will be lost  and the structure of the protein  will be compromised. It is possible that this form might act as a dominant-negative inhibitor of uPAR action, by sequestering uPA and preventing its binding to full length membrane-bound or soluble forms of the receptor.
Repeated epithelial wounding/repair leading to airway remodelling has been implicated in asthma pathogenesis . Epithelial cells have the capacity to repair by cell spreading, migration and proliferation, all integrin-dependent processes. Integrins have therefore been suggested to play a prominent role in wound repair in asthma and the expression of α3β1 (laminin receptor) and α5β1 (fibronectin receptor) integrins have been shown to be upregulated at epithelial wound edges [35, 36]. uPAR interacts with integrins (mainly α3β1 and α5β1) and has been shown to influence cell adhesion and migration on ECM proteins [37, 38]. Therefore the identification of splice variants that lose this ability to bind integrins (i.e. exon 5 deletions) may be of significance in normal and disease mechanisms.
At the protein level, multiple forms of uPAR were detected in cell lysates. Overall, our data suggest that multiple forms of uPAR are expressed in the panel of cells studied and there is some evidence of differential expression. This corresponds to the findings of Brooks et al. , who showed predominantly lower molecular weight forms of uPAR to be expressed in peripheral neutrophils, whilst higher molecular weight forms were more highly expressed in lung eosinophils. These data are also in keeping with our mRNA analyses which suggested a higher proportion of exon deletion variants are expressed in the periphery, potentially resulting in lower molecular weight proteins.
Although the majority of high abundance predicted uPAR variants might be expected to be detected using either the D1 or D2 antibody, the patterns of expression observed were markedly different. This could reflect differing affinities of the two antibodies for different variants. It should be noted that the epitope for the D1 antibody (amino acids 52–60) corresponds to one of the potential glycosylation sites for the protein (Asn52) . The presence of a large carbohydrate side-chain may prevent binding of the antibody, resulting in detection of fewer forms by this antibody as we have observed. Additionally, the D1 antibody will not detect D2/3 proteolytic fragments, explaining the presence of fewer low molecular weight proteins detected using this antibody. An attempt has been made to mirror patterns of protein expression as determined by Western blot to real-time PCR results, with some success for the D1 antibody results e.g. the presence of a 50 kDa protein potentially representing the alternative exon7b observed in all cell types except PMN and PBMC, as well as a 45 kDa form seen only in PBMC and PMN which may represent single exon deletions of classical uPAR. Concordance between the two assays may be affected by differing efficiencies of the real-time PCR assays or antibody specificity. However, it will also reflect genuine biological differences, as post-translational control mechanisms including glycosylation and proteolytic cleavage play a key role in uPAR protein maturation [15, 25]. Our data reflect this complex regulatory pathway.
All cultured cells tested showed expression of soluble uPAR. This assay does not distinguish between classical uPAR released from its GPI anchor, proteolytic fragments and alternatively spliced (exon 7b) soluble uPAR. However, all of these cell lines expressed relatively high proportions of alternative uPAR(E7b) mRNA and the level of mRNA correlates with soluble uPAR protein expression e.g. BEAS2B and THP1 had elevated alternative uPAR(E7b) mRNA and soluble uPAR protein. Therefore, the elevated levels of uPAR in the sputum of asthma and COPD subjects may reflect expression of soluble uPAR from both epithelial and smooth muscle cells.