Lorne special: Sleuthing oncogenes

This feature appeared in the January/February 2010 issue of Australian Life Scientist. To subscribe to the magazine, go here.

The origin and nature of oncogenes was an enduring mystery of 20th century biology. It was known that some viruses contained oncogenes, and these viruses could cause mutations in cellular proto-oncogenes, leading to the development of cancerous growths.

However, it was a great surprise when Harold Varmus and Michael Bishop showed in 1975 that viral oncogenes were not actually true virus genes. Instead, they were cellular genes that the virus had ‘acquired’ during replication. These proto-oncogenes are pervasive in the animal kingdom and play a crucial role in normal cell growth and division as well as cancer.

Varmus and Bishop were awarded the Nobel Prize in Physiology or Medicine in 1989 for their discovery, which has since transformed cancer research, and set off a search for other oncogenes that continues apace today.

At the Diamantina Institute for Cancer, Immunology and Metabolic Medicine in Brisbane, Professor Tom Gonda is leading an ambitious project to screen the 23,000-odd protein-coding genes in the human genome for these Jekyll-and-Hyde genes.

Gonda, who heads the institute’s Molecular Oncogenesis Laboratory, is a veteran of the hunt for oncogenes. His key collaborator in this endeavour is Associate Professor Brian Gabrielli, also of the Diamantina Institute, who heads the Cell Cycle laboratory.

Proto-oncogenes are a diverse class of genes that normally keep a tight rein on growth and division in healthy cells. But when proto-oncogenes transform into their malignant alter-egos through mutation or dysregulation, the unshackled cells can lapse into uncontrolled growth and division.

Gonda’s team has been running pilot experiments with its new, custom-designed, high-throughput facility, which is designed to screen genes for their potential to induce abnormal cell growth when overexpressed.

Gonda is an invited speaker at this year’s Lorne Genome Conference, where he will describe the screening methodology, how it was developed, how the data are analysed, and will also present some results from the pilot screening program.

Gene hunting

When the facility is fully operational, the Diamantina Institute will offer a service to gene hunters throughout Australasia on a collaborative, cost-recovery basis. “What we’ve developed is a high-throughput, functional genomics platform that involves over-expressing candidate genes in human or other mammalian cell lines,” says Gonda.

“Other research groups are using short interfering RNAs (siRNAs) in lentiviruses to look for new oncogenes. We’re taking a very different approach. We are making a library containing every available human gene, by cloning all the corresponding cDNAs into lentiviruses. We have about 2000 genes cloned into viruses already, and we’ve got about 15,000 genes to do. It’s a long, slow process,” he says.

---PB---

“We screen the infected cell lines, one gene per well, using a 96-well format, and look for any changes in cell phenotype. We’ve designed the system to be as flexible as possible, rather than optimise it for a single experiment. The experiments we do depend on what we’re looking for, or what a collaborating researcher is looking for. Because we’re a cancer research group, we’re interested in cancer-related functions.

“We’ve already set up all the liquid-handling processes, the robotics and the data-analysis for our 2000 cloned genes. We’ve made a selection of genes from the library, and we’ve run them through the screening process a couple of times and gone through the analysis procedures, looking for genes that accelerate cells through the growth cycle. It’s quite challenging work.”

The primary aim of the pilot experiment is to provide proof of principle, rather than to discover new oncogenes. “It’s all-new technology that can’t be bought off the shelf. We have to prove that it works,” says Gonda. “Having said that, some very interesting things have already come out of our early experiments. As you might expect, we have identified a number of kinases and phosphatases that have not previously been associated with oncogenesis.

“Some were already known to be involved in the cell cycle, but not during S-phase, which is what we screen for.” During S-phase, or synthesis phase, the cell replicates its chromosomes as a prelude to dividing. “We’ve identified several G-protein signalling molecules that alter cell growth, which wasn’t entirely unsuspected, but they’re not ones that have been implicated as major players in cancer.

“We’ve also found a few very unusual things, including cell-adhesion and cytoskeletal-associated proteins, which may be involved in oncogenesis. That was unexpected,” says Gonda. “But we’ve reconfirmed the results of the original assays; they all scored highly for their ability accelerate cells into S-phase.

Gonda speculates that these proteins are involved in metastasis. “While, our tests aren’t designed to pick that up, interested researchers could pursue these ‘hits’ through more conventional experimental approaches,” he says.

“In all, we’ve had about 40 hits. For some, we have no idea why they’re on the list. It’s very intriguing. We’re getting plenty of interesting hits, but we may not chase them all up. Not every gene that comes out of the screen will be an oncogene, but they may be growth-promoting genes.”

---PB---

Beyond oncogenes

The assay looks for changes in cell morphology, growth rate as measured by rate of DNA synthesis and distribution through the cell cycle, by measuring DNA content. The analysis is performed using high-resolution, automated fluorescence microscopy.

“There are few tricks involved, like looking at changes in the DNA content of cells to detect growth. One of the biggest challenges is to go through the data and determine how to identify genuine hits,” says Gonda. “Automated image analysis requires a statistical approach to distinguish the real effects of overexpressed genes against a background of random variation. The Institute’s senior bioinformatician, Dr Paul Leo, has helped us enormously with this work.

“We use the lentiviruses themselves to solve the problem of creating controls. Not all of the target cells are transduced by the gene being tested, so we link the gene to a green fluorescent marker that automatically identifies transduced cells. We’ve found that the best way of identifying the hits is to look for phenotypic changes in the fluorescent green cells, and compare them with the untransformed cells. All these steps, including cell transfection with the lentiviruses, are performed by liquid-handling robots.”

Gonda says the 18,000-odd genetic elements in the Institute’s gene library probably represent 15,000 to 16,000 unique genes, or around 75 per cent of the protein-coding genes in the human genome.

“At this point, we’re only interested in protein-coding genes. We can over-express, or under-express, in the case of genes like protein kinases. We aren’t doing it yet, but we have the capability to test the effects of overexpressing – or ‘knocking down’ – micro-RNAs, which can cause phenotypic changes in cells by dysregulating the synthesis of proteins from their corresponding coding genes.”

Gonda says the platform is not only useful for identifying oncogenes. With appropriate screening techniques, it could also be used to identify potential tumour-suppressor genes, or genes that confer drug resistance in cancerous cells.

“Beyond that, it has the potential to identify any dominant-acting gene where over-expression confers a detectable phenotypic change. So, while my main interest is in cancer research, and identifying cancer-associated genes, the system’s capabilities go well beyond this field. We’ve already had a lot of interest from researchers studying other types of disease, such as metabolic disorders.

---PB---

“We have many ideas for new screens, but for the time being, we have to prioritise. We will be focusing on helping researchers who supported the project in our original grant applications, but we will also give priority to other projects based on interest and feasibility. We are looking at cherry-picking sets of kinases and transcription factors, but some of our gene sets may come out of genetics, rather than microarrays.

“We plan to offer our services to a very broad clientele. Our model will not be simply fee-for-service, because we hope to establish collaborations with some of the researchers who bring their problems to us. The screens will only work if we can work with them to develop successful robotic assays – you can’t walk in straight from the laboratory bench and expect to achieve the same result with robotics.

Gonda says the technology has the potential to unmask anonymous, normal genes that act as accomplices to well-known oncogenes in the progression to cancerous growth. “It will depend on how we do the screens, and I’m not convinced that we know all the important genes any more. A lot of genes keep popping up when we perform functional screens on genes with multiple cellular roles, so they appear to be oncogenic only in certain contexts.

“I’m certain we will continue to turn up surprises. There are lots of things we can try that would complement findings from more conventional gene-knockdown or loss-of-function experiments. There are lots of different combinations of mutant genes and multiple pathways that can lead to cancerous growth. Our screening technology can accelerate progress in mapping them.”

This feature appeared in the January/February 2010 issue of Australian Life Scientist. To subscribe to the magazine, go here.