In the pancreas, injury and inflammation, or oncogenic mutation of KRAS, lead to a phenomenon called acinar-to-ductal metaplasia (ADM). Although ADM is reversible once injuries are resolved, oncogenic KRAS prevents this, and such lesions can progress to pancreatic intraepithelial neoplasia (PanIN), and, following additional oncogenic changes, to pancreatic ductal adenocarcinoma (PDA). The secondary events involved in progression to PDA are less well defined; two papers in Cancer Cell now indicate a role for epidermal growth factor receptor (EGFR) signalling early in tumorigenesis.

Ardito et al. observed activation of the EGFR pathway in areas of ADM and PanIN in a mouse model with pancreatic-specific expression of oncogenic Kras (KrasLSL-G12D/+;Ptf1aCre/+ mice; hereafter referred to as KrasG12D mice). In this model, Egfr knockout (KrasG12D;EgfrKO) almost completely abrogated tumorigenesis. Furthermore, KrasG12D mice had upregulation of endogenous Egfr and of two of its ligands (transforming growth factor-α (Tgfa) and amphiregulin (Areg)), indicating that EGFR activation in this context might require ligand shedding. To investigate this, the authors examined KrasG12D mice that lacked the TGFα and AREG sheddase ADAM17; these mice phenocopied the KrasG12D;EgfrKO mice. In addition, KrasG12D mice lacking either Egfr or Adam17 were resistant to ADM and PanIN induced by pancreatitis, and EGFR, ADAM17, TGFα and AREG were upregulated in human chronic pancreatitis.

requirement for EGFR in KRAS-driven pancreatic tumorigenesis

Navas et al. also investigated the role of EGFR in KRAS-induced pancreatic cancer, using a slightly different mouse model in which pancreatic acinar cells specifically expressed oncogenic KrasG12V (KrasLSL-G12Vgeo/+;Elas-tTA/tetO-Cre mice; hereafter referred to as KrasG12V mice). In pancreatic cell explants from these mice, ADM was enhanced by the addition of exogenous TGFα or EGF, and knockout of Egfr reduced ADM. PanIN lesions from KrasG12V mice and from humans, as well as human pancreatitis samples, had high EGFR expression. Similar to Ardito et al., these authors knocked out Egfr in the KrasG12V mice and found that loss of EGFR completely abrogated PanIN development. Together, both studies support a requirement for EGFR in KRAS-driven pancreatic tumorigenesis.

The tumour suppressor p53 is often mutated in human PDA, so both groups investigated the role of EGFR in KRAS-mutant mice lacking p53. Ardito et al. observed delayed tumour initiation in response to blocking EGFR signalling by pharmacological and genetic methods in KrasG12D mice with conditional p53 knockout (Trp53KO). However, treatment with the EGFR inhibitor erlotinib (plus gemcitabine) did not improve the survival of KrasG12D;Trp53KO mice with detectable PDA (compared with gemcitabine alone), indicating that later stages of PDA progression might be EGFR-independent. Navas et al. examined KrasG12V mice with conditional Egfr and Trp53 knockout, as well as KrasG12V;Trp53KO mice treated with erlotinib, and found that in both models EGFR loss delayed PDA formation, but did not eliminate it.

Both groups also examined the activation of pathways downstream of EGFR. Ardito et al. found that a threshold level of ERK activation, achieved by activation of both KRAS and EGFR, was required for PanIN formation. Navas et al. noted increased activation of the PI3K–AKT pathway and signal transducer and activator of transcription 3 (STAT3). Futher studies should determine the relevance of these pathways to human PDA progression.

It is intriguing that, as demonstrated in mouse models by Navas et al. and by clinical data, KRAS and EGFR activation seem to be mutually exclusive in lung or colon tumours, indicating a pancreatic-specific role for KRAS and EGFR cooperation. Furthermore, a limited benefit of erlotinib in advanced PDA was observed in a clinical trial, but these studies now support further research to assess the therapeutic value of EGFR inhibition in more defined patient subgroups, such as those that retain p53 function, or for prevention in those at a high risk of PDA.