On your marks...

Credit: Getty

Not all cell types are amenable to transformation by the overexpression, deletion or deregulation of specific genes. Some cell types require multiple genetic and epigenetic changes before becoming tumorigenic, whereas others appear inert, irrespective of the genetic insult. Changes to the surrounding microenvironment can also influence this process, inflammation being a prime example. While searching for the cell of origin of pancreatic ductal adenocarcinoma (PDAC), Tyler Jacks and colleagues found that cellular plasticity — the ability of cells to trans-differentiate in specific conditions — can change the response of a cell to oncogenic mutations.

Activation of KRAS occurs in more than 95% of PDACs and is thought to be the initiating mutation for this disease. So, which of the many cell types in the pancreas are susceptible to KRAS mutation? The authors used several different CreER transgenic mice to specifically express an oncogenic form of KRAS (lox-stop lox (LSL)–Kras^G12D) in distinct cell types in the adult pancreas. Activation of Pdx1CreER^TM resulted in KRAS^G12D activation in sporadic islet, acinar and ductal cells, consistent with the expression of PDX1 in adult β-endocrine cells, ductal and acinar cells and potentially some stem and progenitor cells. These mice developed pancreatic intraepithelial neoplasia (PanIN) grades 1 and 2 — precursor lesions from which PDAC arises. The additional mutation of the tumour suppressor genes Trp53 or Ink4a–Arf was required to induce PDAC and metastatic disease in most of the Pdx1CreER^TM; LSL–Kras^G12D mice. By contrast, activation of Kras^G12D in predominantly acinar cells using the procarboxypeptidase (proCPA1) CreER^T2 resulted in only 2 of 20 mice developing PanIN-like grade 1 lesions. Loss of Ink4a–Arf did not lead to PDAC formation, and only one proCPA1CreER^T2; LSL–Kras^G12D mouse developed metastatic PDAC from five which lacked Trp53.

A previous mouse model that targeted Kras^G12V to pancreatic acinar cells indicated that the induction of chronic pancreatitis using caerulein was required to induce PDAC. Use of caerulein for 1 month prior to activation of proCPA1CreER^T2 resulted in grade 1 PanIN in proCPA1CreER^T2;LSL–Kras^G12D mice, and 3 of 6 proCPA1CreER^T2;LSL–Kras^G12D; Trp53^flox/flox mice developed PDAC. Interestingly, most of the cells in the PanIN lesions did not express carboxypeptidase, indicating that trans-differentiation of carboxypeptidase-expressing cells as a result of the pancreatitis could be instrumental in the development of PanIN lesions.

Expression of KRAS^G12D in insulin-positive cells using RIPCreER^TM resulted in no PanIN lesions, and this result was unchanged by the loss of either Trp53 or Ink4a–Arf. Therefore, cells expressing insulin in the pancreas seem refractory to transformation. However, grade 1 PanINs were evident in one of two RIPCreER^TM; LSL–Kras^G12D mice treated with caerulein prior to expression of KRAS^G12D, and all three RIPCreER^TM; LSL–Kras^G12D; Trp53^flox/flox mice developed metastatic PDAC when treated in this way. So did PDAC in these mice arise from insulin-positive cells? Activation of RIPCreER^T2 followed by exposure to caerulein indicated that insulin-positive cells expressing KRAS^G12D could give rise to PDAC, but that few cells in the resulting tumours produced insulin. Therefore, trans-differentiation of insulin-positive cells to a cell type more amenable to transformation by KRAS^G12D can occur, and interestingly the data in this model also imply that the tumours arose from the minority of insulin-positive cells that reside outside of islets.

There is still much to decipher from this information; in particular, how does caerulein-induced inflammation and the resultant tissue remodelling influence the capacity of KRAS to transform different cells in the pancreas? And, given the relative ease with which cells expressing PDX1 are transformed by KRAS, is trans-differentiation into a cell that expresses PDX1 an essential part of this process?