Mutations in a cell’s DNA — due to damage, viral attack or even spontaneous changes — can produce dramatic changes in a cell. Some mutations have little or no effect, while others change the functionality of the cell or induce cell death. In some cases, however, genetic mutations can cause the cell to multiply out of control to become a tumor.
The fundamental link between genes and cancer is one of the most important research targets in clinical science, and unveiling the genetic mechanism of how a normal healthy cell can abruptly evolve into a cancer is critical in order to develop novel therapies. Without the advent of new treatments, the International Cancer Genome Consortium predicts that the number of people dying from cancer worldwide will more than double by 2050 to 17.5 million each year. Large-scale efforts are now underway around the world to sequence thousands of cancer genomes and identify genetic variations that could be linked to disease. But amid prolonged global economic turmoil, big science like large-scale cancer sequencing is under considerable budgetary pressure, and geneticists are asking the question of whether such massive projects will ever provide meaningful data of clinical relevance, despite their ambitious scope.
Could smaller-scale, alternative approaches be more effective? “A major struggle is how to fund these large-scale international efforts, which are clearly valuable and important, and at the same time not diminish the funding for investigator-initiated research. How do we balance it?” asks Neal Copeland, executive director of the A*STAR Institute of Molecular and Cell Biology. “There is no easy solution to this, and scientists are still debating the issue.”
Strategic shift
Since the Human Genome Project identified some 20,000 protein-coding genes in humans, geneticists have competed to find out how variations in those genes are associated with disease. However, as the research has progressed, it has become apparent that genetic functions are far more complex than initially thought. The angles for research have therefore broadened to cover everything from genetic variations and single nucleotide polymorphisms (SNPs) — changes in the activation of specific genes — to the function of RNA and other targets beyond protein-coding genes. Even networks of genes and molecules have been considered. And although technological advances have dramatically reduced the time and cost of genome sequencing and SNP typing, which has led to the popularization of approaches such as genome-wide association studies, scientists continue to struggle to obtain a clear explanation of the link between genetic function and cancer. That frustration has led to ever-larger projects involving as many samples as possible.
Edison Liu, executive director of the GIS
The A*STAR Genome Institute of Singapore (GIS) has shown leadership in cancer genomics in supporting a number of large-scale international sequencing projects, predominantly in Asia, by contributing its geneticists and sequencers to the cause. One such collaboration is the Pan-Asian SNP Initiative, which aims to characterize genetic diversity in Asia. Singapore itself, through institutes like the GIS, is too small to support its own large-scale sequencing projects and cannot compete on scale with countries like China and the USA, where hundreds of the latest sequencing instruments are being made available for sequencing projects. The GIS, too, has only a handful of sequencers, but its size has not impeded its determination to establish a strong international presence. “We’ve decided not to compete on size,” says Edison Liu, executive director of the GIS. Instead, Liu says, the GIS is strengthening its integration between biology and genomics, with emphasis on transcriptional regulation, the process by which DNA codes are copied into RNA molecules. Researchers at the GIS are constructing comprehensive maps of novel transcripts in human cancers and clinically relevant cancer pathways. The transcripts associated with cancer survival are then assessed for their ability to change cellular behavior. In this manner, Liu’s group has discovered a number of clinically important genes that directly affect cancer cell biology.
Pathway-based approaches
Jianjun Liu, senior group leader and associate director of the GIS, says his recent study1 on breast cancer is a good example of how the data obtained from large-scale projects has helped to evolve his own investigations into cancer pathways. Jianjun Liu’s research focuses on estrogen exposure, currently the most important risk factor for breast cancer, as well as the nuclear estrogen receptor, which is activated by binding to estrogen and consequently acts as a DNA-binding transcriptional factor and gene expression regulator.
Jianjun Liu, senior group leader and associate director of the GIS
The production of the hormone estrogen is regulated by a network of enzymes encoded by various genes. Jianjun Liu believes that mutations within these enzymes could modify estrogen exposure, possibly leading to hormone-related diseases like breast and endometrial (uterine lining) cancers. Previous research, however, has been unable to identify a consistent association between mutations in the metabolic pathway and cancer. Believing that the lack of progress is due to the presence of multiple weak genetic variations with complex interactions, Jianjun Liu in collaboration with researchers from around the world set out to analyze the results with a more sophisticated statistical approach.
Jianjun Liu started by identifying the single SNPs with strong disease associations using more than 4,000 cancer samples, then began investigating multiple-SNP associations within the metabolic pathway by evaluating the cumulative effects of multiple variants. Taking three sets of related tumor samples, he selected 239 SNPs out of 35 genes in the estrogen metabolism pathway, and found that the data could be separated into three sub-pathways: one involved in the synthesis of androgen, a male sex hormone that is converted to estrogen in women; another involved in the conversion of androgen to estrogen; and another responsible for removing estrogen. This approach revealed a strong association between the androgen–estrogen conversion pathway and breast and endometrial cancers. “Pathway-based approaches are just beginning to be applied in association analysis, and can be a powerful method when analysis is guided by well-defined biological information,” says Jianjun Liu.
Mouse models for human pre-clinical testing
Nancy Jenkins and Neal Copeland from the A*STAR Institute for Molecular and Cell Biology are developing mouse models for human cancers
Meanwhile, Copeland and Nancy Jenkins, who jointly run a laboratory for cancer genetics at the A*STAR Institute of Molecular and Cell Biology, have been working in cancer research using mice for over 30 years. Although they are not participating in large-scale human cancer genome sequencing, they have watched developments in the field closely since these projects started on a smaller scale. “We are hoping some of the cancer genes that are coming out of the sequencing studies can be validated using mice,” says Jenkins. The two researchers are aiming to create a mouse disease model that can be applied in the pre-clinical testing of potential human therapies. Their tool is a transposon, a sequence of DNA that can jump to different positions in the genome where it can activate or deactivate specific genes. The transposon they use, called Sleeping Beauty, can be inserted into cancer-causing genes like a tag, allowing clinicians to identify target cancer-causing genes much more quickly and cheaply than other methods, with broad implications for reducing the cost and time required for clinical trials. So far, the team has created models for colon, liver, blood and many other cancers in mice2. “What we want to do is to compare the list of mouse cancer genes with the mutations identified by human cancer sequencing,” says Copeland.
Recently, the team of Copeland and Jenkins looked into colorectal cancers, and investigated which mutations are ‘drivers’, associated with tumor formation and progression, and which are extrinsic ‘passengers’. The team compared their list of Sleeping Beauty-tagged mouse genes with human genes listed in the Catalog of Somatic Mutations in Cancer database. They then analyzed more than 16,000 transposon insertions and identified 74 candidate genes with human ‘homologs’, 15 of which are likely driver mutations in humans. The screen also identified 17 candidate genes that had not been previously implicated in the disease.
Collaborative research
Many scientists have realized that large-scale data provide a lot of useful information, but “in order to fully utilize it, very big samples are needed, and nobody could achieve that working alone,” Jianjiu Liu says. Compared with the situation 5–10 years ago, when massive international collaborations were rare, more and more individual scientists are collaborating with other laboratories, and sometimes the authorship of a single paper can swell to about 300 people. “You can see that the scale of collaboration is really big.”
Others, such as Copeland and Jenkins, are keen to bring their scientific findings to the bedside in collaboration with the pharmaceutical industry. Collaboration is important because in contrast to conventional drug therapy, which might involve one or two drugs with general effectiveness, gene-based therapy is relatively complex, involving multiple drugs for multiple genetic targets. Furthermore, as individuals can respond differently to the same therapy despite having the same type of disease, the therapies need to be personalized. Copeland says pharmaceutical companies are increasing their investment on collaborations with basic scientists, and he hopes that they will be able to fund large-scale international projects in the future. “We would need to partner with pharma companies in order to validate our lists of mice genes in humans and in human cancers,” says Jenkins. Copeland adds: “They need to know which are the driver mutations in cancer, and they need to know which mutations are the best targets for developing therapies. That is expensive and requires worldwide effort. Neither academics nor the pharma industry can do it alone.”
The A*STAR-affiliated researchers mentioned in this feature are from the Genome Institute of Singapore and the Institute of Molecular and Cell Biology.
