Oesophageal adenocarcinoma is a devastating disease that has been rising year on year over the past three decades and is the 6
th highest cause of cancer mortality in the UK, accounting for around 5% of all cancers [
1,
2]. The escalating incidence is thought to be a result of the combination of an obesity epidemic, an aging population, and
H. pylori eradication [
3‐
5]. The disease is curable by surgery or endoscopic therapy if diagnosed at a very early stage [
6] but usually, diagnosis is made at an advanced stage with the presence of lymph node and distant metastases [
5]. There are few clear prognostic indicators of susceptibility to developing oesophageal adenocarcinoma although patients with Barrett's oesophagus are thought to be more at risk to developing oesophageal adenocarcinoma. However, the progression from Barrett's oesophagus to dysplasia and subsequent adenocarcinoma is unpredictable and poorly understood [
7]. The lack of prognostic indicators results in presentation of patents at late disease stages, resulting in poor five year survival rates and patients usually succumb to disease re-occurrence [
5,
8]. For a significant majority, surgery is not beneficial and in such patients with distant metastases, survival is limited to 9 months [
9‐
11]. If the situation is to change then a deeper understanding of tumour growth and metastases is needed to identify new treatment targets.
The ETS domain transcription factor family consists of a group of 27 proteins in humans that all contain the conserved ETS DNA-binding domain and share a core DNA binding specificity centred around the sequence GGA
A/
T[
12,
13]. The PEA3 subfamily includes three transcription factors, PEA3 (also known as ETV4 and E1AF), ER81 (also known as ETV1) and ERM (also known as ETV5). These proteins all contain three conserved domains with sequence identity of 95%, 85% and 50% in the ETS, acidic and Ct domains respectively [
14]. This similarity potentially allows for an overlap in PEA3 subfamily function through acting on a common set of target gene promoters. Indeed due to their conserved DNA binding domain, significant overlap in promoter binding has been observed more generally amongst ETS domain transcription factors [
15,
16]. The PEA3 subfamily plays an important role in embryogenesis, especially in neurogenesis [
17] and also in mammary gland development [
14,
18,
19]. In the adult, PEA3 subfamily members are generally expressed at lower levels and in a more restrictive manner [
14] but ETS domain proteins, and especially the PEA3 subfamily are associated with carcinogenesis, especially tumour metastases and their overexpression often indicates adverse prognosis [
14,
20]. This has been shown to be the case in breast cancer, colon cancer, ovarian cancer and gastric cancer [
14]. More recently, high expression levels of ER81 have been shown to occur in prostate cancer as a result of chromosomal translocations of the ER81 gene into loci with high promoter activity in prostate cells [
21,
22]. PEA3 expression often correlates with enhanced invasive properties and hence is associated with metastasis. For example, in gastric cancer and colon cancer cells, PEA3 inhibition reduces cell invasion
in vitro[
23,
24]. Conversely, PEA3 over-expression induces an invasive phenotype in breast and ovarian cancer cells [
25,
26]. Similarly ER81 over-expression enhances the invasive capabilities of prostate cancer cells [
22]. The invasive phenotypes of cells with high PEA3 subfamily expression are thought to be due in part to their ability to regulate the expression of matrix metalloproteases (MMPs) [
20]. MMP1 has been shown to be an adverse marker in oesophageal adeoncarcinoma [
27,
28]. In colon and gastric cancer cell lines, PEA3 has been shown to regulate
MMP-1 and
MMP-7 expression [
23,
24]. A potential link between PEA3 and MMP7 expression was also suggested in studies on oesophageal squamous carcinoma cells [
29]. MAP kinase signalling is also important in PEA3 activation [
30,
31] in part through driving its dynamic sumoylation [
32]. Importantly MAP kinase signaling synergises with PEA3 in
MMP activation as demonstrated by enhanced MMP-9 and MMP-14 production in response to EGFR signaling in ovarian cancer [
25]. These observations indicate that PEA3 subfamily members are likely central regulators in carcinogenesis and are potential therapeutic targets.
A unifying view of PEA3 function in cancer is therefore that it is a regulator of
MMP expression in response to ERK MAP kinase pathway signaling. However, to date few studies have connected these molecular events together in a single system and the potential role of PEA3 subfamily members in oesophageal adenocarcinoma has not previously been investigated. Indeed, none of the wider ETS domain transcription factor family has been implicated in oesophageal adenocarcinoma, although Ets-1, Ets-2 and Elk-1 have been shown to be over-expressed on squamous oesophageal cancers [
33‐
35]. Here, we show that high PEA3 expression is a frequent occurrence in oesophageal adenocarcinoma. In oesophageal adenocarcinoma cell line models, PEA3 plays a role in promoting invasion and is also important for oesophageal cell proliferation. Molecularly, the invasive properties are likely due to the activation of
MMP-1 expression. Furthermore we also show an important role of the ERK pathway in promoting PEA3 activity and ensuing invasion. In adenocarcinoma tissue, the co-occurrence of PEA3 family member expression correlates with enhanced
MMP-1 expression. Active ERK signaling correlates with enhanced stage suggesting an important role in promoting metastasis via PEA3 and ER81. These results indicate that the ERK-PEA3-MMP-1 axis identified in oesophageal cancer cells is also likely to be operative in oesophageal adenocarcinoma tissue. This pathway could potentially be targeted by drug inhibition with a view to improve prognosis.