Mapping the genetic evolution and diversity of pancreatic cancer

29 Oct 2010

by ecancer reporter Clare Sansom

Cancer of the pancreas has one of the worst prognoses of all the major cancer types, with a five-year survival rate after diagnosis of about 3-4% even in the best-performing countries. This has scarcely changed over decades, largely because this cancer is generally only diagnosed once it has spread.


Now, however, two separate studies published back-to-back in Nature have revealed events in the genetic evolution of metastatic pancreatic cancer. In one, researchers from the Cancer Genome Project at the Wellcome Trust Sanger Centre in the UK, working with colleagues from Cambridge University, UK and Johns Hopkins University, Baltimore, USA, exploited new developments in DNA sequencing technology to map and annotate somatically acquired genetic rearrangements in thirteen cases of pancreatic adenocarcinoma. The samples were obtained from the primary tumour in three cases and one or more metastases in the other ten; about 65 million paired reads were sequenced and the DNA fragments aligned with the reference genome.


A total of 381 somatic and 177 germline DNA rearrangements were identified from the samples obtained and classified into seven types. There were significant differences in rearrangement type and chromosomal location between patients, and even between different metastases in the same patient; intra-chromosomal rearrangements predominated over inter-chromosomal ones in almost all cases. The rearrangements differed significantly from those in previous studies of breast tumours.  The pattern of genetic instability obtained from combining data from all tumours indicated that the tumour cells had cell cycle control abnormalities largely at the point of transition from the G1 to the S phase, and also defects in telomere function.


In ten cases, samples including rearrangements were genotyped in multiple metastatic lesions from the same patient. Finding the same rearrangement in all metastases from the same patient implied that that rearrangement has occurred early in cancer development. This was seen to occur significantly often with one type of rearrangement: a fold-back inversion, which is often associated with telomere loss. Some rearrangements were found that were common to groups of metastases but not the primary tumour, indicating that genetic diversity continues to evolve clonally. Some rearrangements were found only in metastases in a particular organ such as the lung or the abdomen, indicating that metastases within different organs can arise from separate clones.



The second study, by a large group of researchers mainly based at Johns Hopkins Medical Institutions, Baltimore, USA, metastatic lesions were obtained at autopsy from seven of the 24 pancreatic cancer patients whose tumour genomes had been sequenced in an earlier study* that had identified 426 mutations, mainly single base substitutions, in a total of 388 different genes. The genomes of the metastatic lesions were sequenced, and mutations classified into two groups. The majority of mutations – a mean of 63% per patient – were so-called “founder” mutations which occurred both in the primary tumour sample and in all metastatic lesions from the same patient. All other mutations arose in one or more metastatic samples only and were characterised as “progressor” mutations. Lesions were then classified into parental clones (a clone of the original tumour and thus containing only founder mutations) and sub-clones containing both founder and progressor mutations. The parental clones were found to contain mutations in genes that are known to drive pancreatic tumour development such as KRAS and TP53 and a higher percentage of rearrangements and deleterious, homozygous mutations than the sub-clones.   


Mapping the genetic evolution of each patient’s carcinoma from the original tumour to distant metastases revealed that clonal expansions in metastases were marked by large numbers of progressor mutations. Primary tumours from two of the patients were then divided into sections, and the DNA from each section analysed separately. It became clear that cells with the same pattern of mutations as each of the sub-clones were present in the original tumour. The primary carcinoma could therefore be characterised as a mixture of sub-clones, each of which expanded to form metastases. Mathematical modelling of tumour evolution, based on the genetic data and on measured proliferation rates, gave an estimate of over ten years from the initiating mutation in a single cell to the birth of the cell that gave rise to the primary (non-metastatic) tumour, another six years to the development of the first metastasis, and a further 2-3 years to the death of the patient (generally from multiple metastases). This indicates that there is a long “window of opportunity” from pancreatic cancer initiation to metastasis where prompt intervention might save many patients.



These studies and others are beginning to indicate a pattern of mutation and genetic instability that is characteristic of pancreatic cancer. Dissecting the genes involved should aid the identification of novel biomarkers and targets that could, further, lead to more successful treatments for both the early and the late stages of this devastating disease.





Campbell, P.J., Yachida, S., Mudie, L.J. and 22 others (2010). The patterns and dynamics of genomic instability in metastatic pancreatic cancer. Nature 467, 1109-1113. doi:10.1038/nature09460

Yachida, S., Jones, S., Bozic, I. and 11 others (2010). Distant metastasis occurs late during the genetic evolution of pancreatic cancer. Nature 467. 1114-1117. doi:10.1038/nature09515

*Jones, S. et al. Core signaling pathways in human pancreatic cancers revealed by global genomic analyses. Science 321, 1801–1806 (2008).