Lights, camera, action for cells
Time-lapse films reveal the functions of human genes.
Working out the functions of individual genes in human cells is now much simpler thanks to a new database of time-lapse movies showing cells in action.
Jan Ellenberg of the European Molecular Biology Laboratory (EMBL) in Heidelberg, Germany, and his colleagues across Europe introduced the freely accessible database of 190,000 videos to the scientific community along with a paper published today in Nature1.
Ellenberg and his team set out to observe what happens to cells when each of the 21,000 human protein-coding genes is disrupted. They perturbed gene expression using short interfering RNA molecules (siRNA) and then observed the effects over two days on fluorescently labelled chromosomes using time-lapse imaging. Because this generated huge amounts of data — more than 19 million cell divisions — the researchers developed computational tools to analyse all of the videos. Their technique automatically tracked the position of the nucleus in each cell and classified its appearance into 16 different categories. The method recognized with 87% accuracy changes in the nuclear shape that were related to basic functions such as cell division, proliferation, survival and migration.
"Technically this paper is really a tour de force," says Jason Swedlow, a cell biologist at the University of Dundee, UK. "The systematic way the group has gone through and knocked down genes and filmed the results is really impressive."
Ellenberg and his colleagues more than doubled the number of genes that are known to be involved in mitosis, a type of cell division. "This is a very important achievement," Swedlow says. "Even though we know the sequences of different genomes, we don't yet know the names of all the genes involved in a fundamental process like cell division."
After performing the initial genome-wide screen of 21,000 genes, the researchers identified 1,249 genes associated with changes in the appearance and spatial arrangement of chromosomes. Nearly half of these genes are involved in mitosis (see video showing normal cell division). Then they linked together snapshots of cell events across time to visualize how cells adapt to disruptions in this process. This showed that delays in mitosis result in cell death and abnormal chromosome segregation, the process by which paired chromosomes split (see video of cell division disrupted using siRNA).
Some genes were associated with problems in early cell division, and others were linked to problems in a later stage called cytokinesis, in which cells containing two nuclei divide in half. The scientists used these groupings to predict the mitotic functions of the newly discovered genes, including those involved in cytokinesis and spindle assembly — a process that helps to align the chromosomes in the middle of the cell before it divides. The authors also went on to identify 783 genes that could have a role in cell survival and 360 genes that might guide cell migration.
"It's a very comprehensive study," says Michael Boutros, a cell biologist at the German Cancer Research Center, also in Heidelberg. "It broadly reflects the different effects that genes have on the cell cycle. That's going to be a very important contribution."
The next step is for multiple research groups to validate the hundreds of candidate mitotic genes found in this screen and to figure out how they guide cell division, Ellenberg says. Because these genes are often involved in cancer, it will be useful to compare how they regulate the cell cycle in cancerous and healthy cells. "In the long run, the technology will really allow us to diagnose and treat cancer much better," he says.
Early signs indicate that the database will be popular. "Overwhelmingly the reaction is very positive," Ellenberg says. "People bombard me with e-mails and ask, 'Can I have the movies for my favourite gene?'" Now they can log on to the database and discover clusters of genes that underlie a particular response in cells, he says. "People want to have the data. They don't want to do the silencing experiments themselves anymore."
"It's a great resource," Boutros says. "People can go to the data set and discover new relationships that have not been described in the paper. To have everything available online is a big step forward."
- Neumann, B. et al. Nature 464, 721-727 (2010).