Research

We have chosen to model malignant progression in patients and genetically engineered laboratory mice, focusing upon pancreatic cancer and more recently melanoma as prototypical diseases with poor prognoses despite the identification of the predominant oncogene in each disease. Our approach entails the characterisation of molecular and cellular events following driver oncogene expression, the identification of cell intrinsic and extrinsic pathways in tumour evolution, and the evaluation of therapeutic and diagnostic strategies in these model systems. These preclinical results have prompted the initiation of a number of clinical trials that our laboratory directs.

Cellular consequences of oncogene expression

The expression of oncogenes may be productive, negative or neutral to cellular physiology, thereby illuminating the tropism of oncogenes in human cancer. Recently, oncogenes have been shown to cause a proliferative arrest in primary human and murine preneoplasms in vivo. This event, termed oncogene-induced senescence (OIS), has been described as an important barrier to the development of malignancy. We have sought a detailed characterisation of the immediate and prolonged effects of conditional endogenous Kras and Braf oncogenes in primary cells, as they are associated with most cases of ductal pancreatic cancer and cutaneous melanoma, respectively. We find that the endogenous KrasG12D allele promotes partial transformation of primary fibroblasts and epithelial cells in culture, and stimulates proliferation in relevant tissues (pancreas and lung) in vivo. There exist a low percentage of OIS cells in pancreatic preneoplasms, which may instead reflect terminally differentiated cells that occur during metaplasia. Instead, we find that a more formidable barrier to tumour initiation may be the negative influence of neighbouring wild-type epithelial cells that suppress the proliferative and transformed properties of oncogene expressing cells (Figure 1). We are currently characterising the molecular details of this interaction. Whether oncogenic Kras has deleterious or neutral effects in tissues that are not normally transformed is also currently under evaluation.

Figure 1
The effect of normal cells on pancreatic preneoplasm development. (A) and (B) KrasG12D was expressed in a complete manner in LSL-KrasG12D; p48-cre mice and a mosaic manner in LSL-KrasG12D; pdx-cre mice as shown by gfp staining when crossed to a cre reporter strain. (C) The presence of normal cells results in delayed preneoplasm development with time. (D) A cartoon illustrating where neighbour suppression is postulated to operate.

Cell autonomous and non-cell autonomous events in carcinogenesis

Both cell autonomous and non-cell autonomous events shape the evolution of carcinogenesis through internal signalling events and homotypic and heterotypic cellular interactions. Whereas such events are difficult to characterise using human specimens, they can be readily pursued with our murine cancer models through a variety of biochemical, cellular and genetic approaches.

We have undertaken both unbiased and targeted approaches to identify important mediators of KrasG12D transformation. Indeed, a global proteomic and transcriptional evaluation of primary cells following oncogenic Kras expression has revealed alterations in Reactive Oxygen Species (ROS) metabolism as an important facet of cellular immortalisation in primary cells. Although ROS have previously been proposed to promote cellular transformation, we instead find that endogenous ROS is deleterious to cellular immortalisation and that KrasG12D expressing cells enact a cellular programme to rapidly neutralise ROS and other cellular toxins. The regulation of altered ROS metabolism in KrasG12D expressing cells is due to signalling by the MAP kinase cascade, and thus represents a possible therapeutic target in early neoplasms. Other common oncogenes such as BrafV600E and c-Myc also activate the ROS metabolism pathway to lower ROS levels, suggesting that this is a general pathway of relevance during the initiation of cancer. In addition to the MAPK pathway, we are seeking alternative methods to inhibit this ROS metabolism pathway as a means to prevent the onset of cancer or treat established cancers.

Our targeted approaches include the conditional ablation of several pathways that may cooperate with oncogenic Kras to promote tumourigenesis. For example, to determine whether specific branches of the Raf signalling pathway are required to sustain tumourigenesis, we have concomitantly deleted CRAF or BRAF while expressing KrasG12D in lung epithelial cells in vivo. We find that CRAF, but not BRAF, is required to promote early lung cancer. Current work is focused upon elucidating the exact molecular pathways downstream of CRAF, besides MAPK, that are required to promote KrasG12D oncogenesis.

We also use both classical and forward genetic approaches to uncover novel genes and pathways that facilitate pancreatic cancer initiation and metastasis. This entails high throughput sequencing of spontaneous and transposon-stimulated tumours, and the validation of potential candidates in cell culture and in vivo. We recently identified a major new pathway important in pancreatic cancer genesis, and are currently establishing its clinical relevance.

Cell biological approaches afford the opportunity to explore the non-cell autonomous interactions in developing tumours and thereby identify features that may both suppress and promote tumourigenesis. Our initial efforts have focused on the identification of the cell types that comprise the immune, stromal and vascular constituents of the tumour microenvironment. We recently reported the presence of immature myeloid cells in pancreatic preneoplasms and invasive tumours. Such cells might stimulate tumourigenesis as they have been implicated as potential suppressors of the acquired immune response. The role of inflammation in the genesis of pancreatic cancer is also being investigated. To do this, we are analysing the importance of the immune system and stromal cells in the generation of the desmoplastic stroma common in preinvasive and advanced pancreatic cancer, and we are investigating cell autonomous paracrine and autocrine factors that could stimulate the microenvironment. Work done in collaboration with the laboratory of Douglas Fearon (Department of Medicine) has revealed that the cancer fibroblasts in our pancreatic tumours are immune suppressive. We are currently evaluating this finding.

Pancreatic cancer medicine advances in the preclinical and clinical settings

We have established with our mouse model of pancreatic ductal adenocarcinoma that the intratumoural vasculature in such tumours is sparse and highly compressed, correlating with poor tissue perfusion and therapeutic delivery. These observations have been extended to human tumours, and we have developed various strategies to correct this defect by targeting the tumour stroma. We have initiated several clinical trials to confirm our findings, including a pre-surgical trial that uses a Hedgehog pathway inhibitor for two weeks prior to surgery. Likewise, our work has directly stimulated the opening of three other clinical trials that our laboratory either clinically directs or supports.