Two research teams have independently developed translocation sequencing strategies that they used to find sites prone to double-strand DNA breaks and translocations in primary mouse B cell genomes.
Both approaches were described in the most recent issue of Cell. In one paper, a National Institutes of Health-led team used a method called translocation-capture sequencing, or TC-seq, to catalog chromosomal rearrangement patterns in activated mouse B lymphocyte cells. Meanwhile, researchers from Children's Hospital Boston, Harvard Medical School, Massachusetts General Hospital, and elsewhere tracked translocation in the same cell type using a similar method that they call high-throughput genome-wide translocation sequencing, or HTGTS.
The targeted, deep sequencing strategies both rely on creating a DNA break in a known site in the genome and using it as bait to capture pieces of DNA that have naturally broken off in other parts of the genome.
By looking at translocations and rearrangements in primary cells as they happened, the teams were able to get a snapshot of the "early, unselected translocatome," University of Massachusetts Medical School researchers Rachel Patton McCord and Job Dekker commented in an accompanying article in the same issue of Cell.
"With distinct, yet similar experimental protocols, these two studies paint a picture of some of the underlying mechanisms that predispose certain regions to translocations in B cells," Patton McCord and Dekker noted.
In the past, researchers have used whole-genome sequencing to look at chromosomal rearrangements in tumors. For example, in a study published in Nature in 2009, researchers at the Wellcome Trust Sanger Institute used paired-end sequencing to assess chromosomal rearrangements in two dozen breast cancer samples (IS 1/5/2010).
But while such studies offer valuable insights into the genes affected by translocations in cancer, they can only identify the chromosomal rearrangements that have survived selective pressures in the cell and gone on to contribute to cancer, explained Rafael Casellas, acting chief of genomics and immunity for the National Institute of Arthritis and Musculoskeletal and Skin Diseases.
"There have to be other events that accumulate for the cell to retain a translocation that can transform it," Casellas, co-corresponding author on the TC-seq paper, told In Sequence. "Our guess is that there are many cells in our body that have translocations, but because they don't deregulate an oncogene or don't mess up the cell too much, then the cell will survive with it."
Another consideration for those trying to determine sites in the genome that lend themselves to involvement in chromosomal rearrangements is the frequency with which such changes arise in cells.
Tumor samples are comprised of many cells containing the same chromosomal rearrangements, which makes it possible to detect these changes by whole-genome sequencing.
In primary cells, though, translocations are only found in a fraction of cells and the sequence involved can vary from one cell to the next, Monica Gostissa, co-corresponding author on the HTGTS study, told IS.
That means that the type of whole-genome sequencing approaches used to study tumors will miss many translocation events in normal, primary cells, explained Gostissa, a researcher in the lab of co-senior author Frederick Alt at Children's Hospital Boston and Harvard Medical School.
In the mouse B cell lymphocytes, for instance, translocations turn up in between one in every 500 and one in every 1,000 B cells, Gostissa noted.
Engineered B Cells
B cells are adept at rearranging their genetic patterns to achieve the flexibility needed to produce antibodies against a wide range of potential pathogens and other immunogens in the environment. An enzyme called activation-induced deaminase, or AID, is expressed in B cells and damages the genome at specific sites, creating translocation hotspots in the immune cells.
"The immune system has to be prepared for all kinds of antigens emerging from the environment, which is to some degree completely unpredictable," said Markus Muschen, a pediatrics, biochemistry, and molecular biology researcher with Children's Hospital Los Angeles and the University of Southern California Norris Comprehensive Cancer Center. Muschen works on B cell development and leukemia and was not involved in either of the new studies.
"That is the reason why the immune system, not only in humans and other mammals, but also in other organisms, has maintained a certain level of genetic variability, including DNA recombination, but also a process called hypermutation, which is catalyzed by the AID enzyme," he added.
While AID activity appears to help B cells respond to a wide range of immune insults, it also increases the risk of deleterious translocations that can lead to B cell leukemia or B cell lymphoma, Casellas explained, though the cells have safeguard systems to help protect them against unwelcome or damaging mutations.
To get a better look at the range of possible translocations in B cells and the mechanisms contributing to them, each team engineered mouse B lymphocyte cells with restriction sites recognized by the enzyme I-SceI at either the immunoglobulin heavy chain or the c-myc locus.
Sequences from the IgH locus, which codes for the antibody heavy chain in B cells, and the c-myc locus, which codes for an oncogene, are often found within translocations in mouse and human B cell tumors, Gostissa explained.
"We focused on those two loci to start with," she said. "But in principle the approach is feasible for any locus that you may want to test."
By introducing the I-SceI restriction site, normally found in yeast but not mammalian cells, the investigators created a system in which double-strand DNA breaks could be reliably introduced at either IgH or Myc loci, creating bait for breaks from other parts of the genome.
"One of those pieces of DNA, once in a while, will have a translocation between the I-SceI restricted site and another piece of the genome, which is unknown," Casellas noted.
After setting up this system and allowing such interactions to occur, researchers isolated and sheared genomic DNA into fragments a few hundred base pairs long. Among them: DNA fragments comprised of bait sequence fused to a translocated sequence from an unknown site in the genome.
[ pagebreak ]
To identify this sequence, Casellas and his NIH team added linker DNA to their DNA fragments and did PCR amplification using primers that recognize sequence in the bait DNA and the linker DNA.
Gostissa and her colleagues came up with two PCR methods for their HTGTS experiments, one that used linkers and another relying on circularization PCR, which involves circularizing the DNA fragments prior to amplification.
Once the sequences were amplified by PCR, each team sequenced their libraries using high-throughput sequencing.
For the TC-seq study, Casellas and his colleagues relied on deep, paired-end sequencing with the Illumina GAII, using bioinformatics to identify the sequences involved in the translocation.
Gostissa and her collaborators, meanwhile, did their HTGTS study using the Roche 454 GS FLX Titanium, which generates longer reads that are better suited for sequencing across translocation breakpoints.
"The reason why we chose 454 is that we were interested in getting the junction of the translocation, so the junction between the bait sequence and the unknown sequence," Gostissa said.
Each platform has its advantages and disadvantages, Casellas explained, since the Illumina short-read system generates deep sequence relatively quickly, while the 454 approach provides longer reads but somewhat lower resolution.
There are price differences as well. Casellas noted that he and his colleagues typically ran one sample on each lane of a seven-lane slide, at a cost of around $4,000 per slide.
On the 454 platform, Gostissa estimated that her team spent around $1,500 to sequence each region on an eight-region plate. Each region allowed them to sequence one library containing roughly 5,000 to 8,000 translocations, she said.
From the more than 180,000 rearrangements that they detected in 400 million mouse B lymphocyte cells using the TC-seq approach, Casellas and colleagues found several clues about the sequences that tend to translocate in the lymphocytes.
The rearrangements detected often included sequences from actively transcribed genes, they reported, especially sequences from the transcription start sites of these genes.
And while the bait sites interacted with sequences from across the genome, the Myc or IgH sites were somewhat more likely to attract translocations from their respective chromosome sequences than from other chromosomes.
Gostissa and her colleagues found a similar propensity for intrachromosomal rearrangements and for translocations involving the transcription start sites of genes in transcribed regions of the genome from their HTGTS-based analyses of almost 150,000 translocation junctions.
As expected, both teams found translocation hotspots at sites in the genome where the AID enzyme is known to interact, but there were also non-AID-interacting sites where breaks were more common than expected.
In both studies, researchers also found some sites in the mouse genome outside of their engineered I-SceI restriction sites at the IgH or c-myc loci that seem to be cut by the I-SceI enzyme.
"We engineered the cells to have the I-SceI site, but when we did the experiment, we learned that in the mouse genome there are actually a bunch of the cryptic I-SceI sites," Gostissa said. "They are a little bit divergent from the canonical consensus, but they are still cut."
That may be useful down the road, she noted, since it raises the possibility of doing translocation sequencing studies that exploit these cryptic sites to generate bait sequences, rather than relying on restriction sites that have to be engineered into the mouse genome.
Translation to Human Cells?
But not everyone is convinced that such mouse studies will be directly applicable to understanding human B cells.
While he called the translocation sequencing approach technically beautiful, Muschen questioned whether the rearrangement patterns and mechanisms identified in the mouse B cell model will translate to human cells, since mouse B cells appear to have a bias toward rearrangements involving Myc, which does not seem to exist in human B cells.
"I'm not convinced that this model is immediately applicable as a model to reveal mechanisms of chromosomal translocations in humans, which I think is the ultimate objective here," Muschen said.
"As a proof of principle, to show that there are indirect ways of mapping genetic lesions in a genome-wide manner is fascinating to me," he said. "The only limitation I would see here is that there is the huge bias in the mouse in favor of Myc."
For her part, though, Gostissa said that it's too early to know whether there really is a bias toward Myc translocations in mice, since researchers' understanding of mouse translocations stems from studies of tumor rather than primary cells.
"There are translocations that involve Myc in human cells as well," Gostissa said. "So I wouldn't confidently say that mouse B cells are more prone to translocated Myc than human cells."
"And we always should keep in mind that whatever translocation we see in a tumor is a product of a lot of different factors," she added. "One is the intrinsic ability of a locus to translocate and there's also the selective pressure that the translocation is subjected to."
To understand the relationship between translocations in primary cells and those that persist in or cause tumors, Gostissa argued that more studies of primary cells, including human primary cells and cell lines, are needed.
She and her colleagues are currently working on strategies for transitioning their HTGTS method into human cells. The general system will likely be the same, she said, but will require additional steps to design good primers and to select an appropriate bait.
Rather than engineering in an I-SceI restriction site, for instance, the researchers plan to focus on finding existing sites in the human genome that could serve as feasible bait sequences.
Alternatively, Gostissa noted that some companies are developing endonucleases that target specific loci in the human genome — enzymes that could theoretically help her team create double-strand DNA breaks at a desired location in the genome of human cells, which do not tolerate artificial restriction sites.
It may also be feasible to do similar analyses in other cell types. Gostissa said that she and her colleagues are interested in comparing the type and frequency of translocations in B cells with those in other primary cell types, such as fibroblasts.
Although they are primarily focusing on B cells, Casellas said he and his colleagues are also interested in trying their hand at translocation sequencing in fibroblasts and other cell types.
That team is now looking in more detail at what Casellas called "fragile sites" that they detected in the mouse B lymphocyte genome in the hopes of determining whether the propensity for DNA breaks and translocations at these sites is due to intrinsic sequence properties and/or other factors.
They have started using the Illumina HiSeq 2000, Casellas said, which generates roughly ten times as much sequence as the GAII platform they used for the current study. Given the boost in sequence output, he said, one lane of HiSeq data will probably generate too much sequence data. Consequently, the team plans to barcode their samples so they can run three or four in each lane for future translocation sequencing studies.
Have topics you'd like to see covered in In Sequence? Contact the editor at anderson [at] genomeweb [.] com.