Characterisation of urokinase plasminogen activator receptor variants in human airway and peripheral cells

Abstract

Background: Expression of the urokinase plasminogen activator receptor (UPAR) has been shown to have clinical relevance in various cancers. We have recently identified UPAR as an asthma susceptibility gene and there is evidence to suggest that uPAR may be upregulated in lung diseases such as COPD and asthma. uPAR is a key receptor involved in the formation of the serine protease plasmin by interacting with uPA and has been implicated in many physiological processes including proliferation and migration. The current aim was to determine key regulatory regions and splice variants of UPAR and quantify its expression in primary human tissues and cells (including lung, bronchial epithelium (HBEC), airway smooth muscle (HASM) and peripheral cells).

Results: Using Rapid Amplification of cDNA Ends (RACE) a conserved transcription start site (-42 to -77 relative to ATG) was identified and multiple transcription factor binding sites predicted. Seven major splice variants were identified (>5% total expression), including multiple exon deletions and an alternative exon 7b (encoding a truncated, soluble, 229aa protein). Variants were differentially expressed, with a high proportion of E7b usage in lung tissue and structural cells (55-87% of transcripts), whereas classical exon 7 (encoding the GPI-linked protein) was preferentially expressed in peripheral cells (approximately 80% of transcripts), often with exon 6 or 5+6 deletions. Real-time PCR confirmed expression of uPAR mRNA in lung, as well as airway and peripheral cell types with ~50-100 fold greater expression in peripheral cells versus airway cells and confirmed RACE data. Protein analysis confirmed expression of multiple different forms of uPAR in the same cells as well as expression of soluble uPAR in cell supernatants. The pattern of expression did not directly reflect that seen at the mRNA level, indicating that post-translational mechanisms of regulation may also play an important role.

Conclusion: We have identified multiple uPAR isoforms in the lung and immune cells and shown that expression is cell specific. These data provide a novel mechanism for uPAR regulation, as different exon splicing may determine uPAR function e.g. alternative E7b results in a soluble isoform due to the loss of the GPI anchor and exon deletions may affect uPA (ligand) and/or integrin binding and therefore influence downstream pathways. Expression of different isoforms within the lung should be taken into consideration in studies of uPAR in respiratory disease.

Figure 1

Schematic representation of uPAR structure. UPAR consists of seven exons, with an alternative exon 7b previously reported [24]. The classical form is transcribed to give a three domain protein which can be membrane bound, via a GPI anchor. Domain S: Signal peptide removed during processing, Domain G: removed during processing to give GPI anchor at new C-terminus. Different regions of each domain involved in various protein-protein interactions are highlighted, which may give an indication of the functions of variant forms of the receptor. uPA binding involves residues in all three domains (see Figure 4), with key binding regions in D1 and D2 [17,19]. The minimum chemotactic sequence is located in the D1-D2 linker [22], A region implicated in binding integrins α1β3 and α5β1, aiding signalling to vitronectin, which also has chemotactic activity is located in D2 [20]. A region in D3 which binds integrin α5β1 is also highlighted [21].

Figure 2

Identification and characterisation of the uPAR promoter. (A) 5' RACE performed on a panel of tissues/cell types identified a cluster of transcriptional start sites (TSS) which show distinct patterns of usage. TSS are listed relative to ATG of translated sequence. (B) 4000 bp upstream (from ATG) was analysed for potential transcription factor binding sites using four different programs. Those sites identified in two or more different searches are shown.

Figure 3

Identification of 3' variants in a panel of tissues/cell types. 3' RACE was performed on a panel of tissues/cell types using a uPAR-specific forward primer located in exon 4. Frequencies of the most common variants (>5% overall expression) in each cell type based on numbers of RACE clones are summarised (A) and their structures shown (B). "Other" variants include both alternative termination sites and splice variants. Grey blocks highlight 3' UTR in the terminal exon.

Figure 4

Protein sequences of uPAR exon 7 splice variants. Alternate splice variants which were identified and confirmed by real time PCR in the current analyses are shown. Alternate exons are labelled back and blue; red amino acids are encoded over an exon boundary. The signal peptide, removed during processing (not included in numbering of mature peptide) is highlighted green. The region highlighted pink is removed during processing to give GPI anchor at the new C-terminus. Domains are defined by the end cysteines involved in disulphide bridges (purple). Peptide D2A (yellow, domain 2) binds integrins αvβ3 and α5β1, aiding signalling to vitronectin and also has chemotactic activity [20], whilst a region highlighted in domain 3 (yellow) binds integrin α5β1 [21]. The minimum chemotactic domain (turquoise) binds FPRL1 and encourages chemotaxis of many cell types [22]. uPA binding regions determined by phage display and peptide array are highlighted grey [19], whilst residues involved in the uPA binding determined by alanine scanning mutagenesis are highlighted red [17].

Figure 5

Protein sequences of uPAR internal exon splice variants. Alternate splice variants which were identified and confirmed by real time PCR in the current analyses are shown. A signal peptide removed during processing (not included in the numbering of mature peptide) is highlighted green and the terminal domain (pink) is removed during processing to give a GPI anchor at the new C-terminus.

Figure 6

Location of Taqman primers and probes. Variants shown were identified in one or more clone and in one or more cell type during RACE/PCR analyses. Location of primers and probes generating variant specific TaqMan assays are indicated.

Figure 7

Expression of uPAR splice variant mRNA in different tissues/cell types. A series of real-time PCR (TaqMan) assays was used to measure the expression of different splice variants of uPAR in an extended panel of tissues/cell types (lung, brain, HASM, undifferentiated HBEC, BEAS2B, PMN, PBMC, THP1). Expression of each variant is shown as mean + SEM of three PCR replicates, for two donors or biological replicates as appropriate. Data are shown as 2^-ΔCt normalised to HPRT1 and relative to a suitable plasmid positive control containing the specific splice variant cDNA (designated 100%). (A) total uPAR, (B) total classical uPAR (exon 7), (C) classical uPAR exon 6 deletion, (D) classical uPAR exons 5+6 deletion, (E) classical uPAR exon 3 deletion, (F) total alternative uPAR (exon7b), (G) alternative uPAR exon 4+5 deletion. Classical uPAR exon 5 and 4+5 deletions and alternative uPAR exon 5 deletion were not detected.

Figure 8

Expression of different uPAR splice variant mRNAs in different tissues/cell types normalised to total uPAR. A series of real-time PCR assays was used to measure the expression of different splice variants of uPAR in an extended panel of tissues/cell types (lung, brain, HASM, undifferentiated HBEC, BEAS2B, PMN, PBMC, THP1). Expression of each variant is shown as mean + SEM of three PCR replicates, for two donors or biological replicates as appropriate. Data are shown as 2^-ΔCt normalised to total uPAR and relative to a suitable plasmid control (designated 100%). (A) total classical uPAR (exon 7), (B) classical uPAR exon 6 deletion, (C) classical uPAR exons 5+6 deletion, (D) classical uPAR exon 3 deletion, (E) total alternative uPAR (exon7b), (F) alternative uPAR exon 4+5 deletion. Classical uPAR exon 5 and 4+5 deletions and alternative uPAR exon 5 deletion were not detected.

Figure 9

Expression of uPAR protein in different cell types. Western blotting of cell lysates was performed for an extended panel of cell types (HASM, undifferentiated HBEC, BEAS2B, THP1, PMN, PBMC) plus recombinant uPAR (ruPAR) using domain I specific (A) and domain II specific (B) antibodies to identify different variants. A β-actin antibody was used as a loading control (C). An ELISA assay, with a sensitivity of 30 pg/mL, was used to detect soluble uPAR in the culture supernatant where appropriate (D).

Figure 10

siRNA knockdown of uPAR in human bronchial epithelial cells. HBEC were treated with siRNA (negative control, uPAR specific (S032) or pooled siRNA panel (mix)) for 24 hours before RNA and protein extraction. Western blotting was performed using two different anti-uPAR antibodies and an anti-β-actin antibody. Recombinant uPAR (ruPAR) was run as a control (A). Real-time PCR was performed using a total uPAR assay and normalised using HPRT as a housekeeping gene (B). Densitometry was performed on Western blots (regions measured shown boxed) to allow semi-quantitative measurement of knockdown. Data are shown before (C) and after (D) normalisation for β-actin expression. The anti-D2 antibody detected multiple different weight species. Densitometry was performed for individual bands and normalised for β-actin expression (D).