Tracking evolution in prostate cancer with cfDNA profiling
- Gerhardt Attard, London, United Kingdom, University College London Cancer Institute
Tweaking DNA repair pathways to improve cancer immune surveillance
- Alberto Bardelli, Candiolo, Italy, Istituto di Candiolo - FPO - IRCCS
Clonal evolution of metastatic disease
- Samra Turajlic, London, United Kingdom, Royal Marsden Hospital NHS Foundation Trust
3O - Temporal dissection of altered pathways during the evolution of cancer
- Johanne Ahrenfeldt, Aarhus, Denmark, Institute of Clinical Medicine, Aarhus University Hospital Skejby
Abstract
Background
Cancer arise as normal cells acquire insults to the genome through both mutations and gross copy number alterations. Our current understanding of the molecular makeup of cancer is based on a snap-shot, a biopsy taken at a single time point. Large-scale studies have shown an abundance of genomic alterations, but few are shared across samples which makes interpretation of effect difficult, particularly as different processes may be dominant at different time points in the development of cancer. Here, we have performed pathway analysis on timed mutational driver events to investigate if specific pathways are preferentially affected at different time points during the life-history of cancer.
Methods
We analyzed mutation data and copy number information from bladder cancer acquired from The Cancer Genome Atlas, to time the mutations as early and late relative to copy number events affecting the same genomic location (McGranahan et al., 2015). The mutations were annotated to genes and potential driver mutations were annotated (Jamal-Hanjani et al., 2017). Timed driver events across each cohort were combined into early and late, and subjected to pathway analysis using Reactome (Fabregat et al., 2017).
Results
Focusing on bladder cancer, 28% of the timed mutations are early and 72% are late. The most frequent early driver mutations are TP53 (12%), MUC4 (8%), and PIK3CA (5%), while in late they are PABPC1 (21%), KMT2C (16%), HLA-A (9%). Pathway analysis on early drivers showed how events that affects cell cycle and proliferation dominates, while late drivers are particularly enriched in pathways associated with the immune system.
Conclusions
Our results demonstrate a clear time-separated preference for specific events. These indicates that selection in early cancer development may be primarily focused on genomic alterations that increase proliferation. Conversely, late cancer development enriches in events targeting pathways associated with the immune response, consistent with increased immune pressure as the cancer develops. These results show how timing of events may provide novel insights into how cancer develops, and may help determine the evolutionary developmental stage of cancer with potential implications for improved prognostics and differential therapy.
Legal entity responsible for the study
The authors.
Funding
The Lundbeck Foundation.
Disclosure
All authors have declared no conflicts of interest.
4O - Tumour cells mirror embryonic developmental programs to acquire invasive and metastatic capabilities
- Sadegh Saghafinia, Lausanne, Switzerland, EPFL - Swiss Institute for Experimental Cancer Research (ISREC)
Abstract
Background
Pancreatic Neuroendocrine Tumours (PanNETs) are the second most common form of pancreatic cancer. We previously identified two principal subtypes of PanNET: insulinomas (islet tumors; IT) and metastasis-like primaries (MLP), corresponding to the low- and high-grade classification of human PanNETs. This study describes the mechanisms by which PanNETs progress from IT to more aggressive MLP tumors and eventually metastasize.
Methods
We profiled single cell transcriptomes, bulk mRNA and miRNA transcriptomes and the proteomes of primary and metastasis specimens from the genetically engineered mouse model of PanNETs (Rip1 Tag2).
Results
We found that the IT tumours maintained the expression of mature ß cell markers. In contrast, the MLP tumours expressed pancreatic progenitor markers. By integrating the data on ß-cell differentiation from pancreatic progenitors to mature ß cells, we demonstrated that the tumour progression from IT to MLP follows the reverse embryonic and postnatal developmental path. Furthermore, we identified Hmgb3 and miR-181cd cluster as the MLP master regulators. Over-expression of the miR-181cd cluster in IT-like cancer cell-lines resulted in the acquisition of the MLP gene signature and MLP morphologic phenotypes, in addition to activation Hmgb3 expression. This suggested a central role for Hmgb3 in initiating the MLP program. Furthermore, inhibiting the expression of Hmgb3 in MLP-like cancer cell-lines resulted in a significant growth decrease, demonstrating the importance of Hmgb3 in the maintenance of MLP-like cell state in vitro. Using transcriptomic data from human patients, we evaluated the relevance of the MLP program in human tumours. We established that aggressive human PanNET tumours also follow the same reverse developmental trajectory of dedifferentiation. Notably, patients with high MLP genes expression had a worse prognosis.
Conclusions
These data demonstrate dedifferentiation as a mechanism by which malignant neuroendocrine cancer cells acquire progenitor-like features, enabling them to become more aggressive and metastatic. In addition, miR-181cd cluster and Hmgb3 act as the core regulators in the initiation of dedifferentiation and maintenance of progenitor-like features in tumour cells.
Legal entity responsible for the study
CMSO Group at ISREC, EPFL.
Funding
Swiss National Science Foundation.
Disclosure
All authors have declared no conflicts of interest.