Reshuffling Cancer’s Deck with Tumor Genomes

flickr photo of "Bootstrap DNA" by mira66

flickr photo of Charles Jencks' "Bootstrap DNA" by mira66

If you want to describe cancer in the fewest words possible, try this: When Genetics Go Wrong. All cancers can be traced back to genetic mutations, either in genes meant to help the cell “self-destruct” when damaged or in genes that promote the replication of cells. Tracking down the genetic origin of a particular cancer – as Janet Rowley did in the ’60s and ’70s with chronic myelogenous leukemia – can point the way to new, effective treatments. But so far, these have been isolated victories chipping away at the complex and diverse world of cancer, where countless other genetic varieties reside.

Various projects are starting up to address try and map cancer’s genetic landscape, from the work published last week in Nature describing the genomes of two cancer cell lines to the National Cancer Institute’s Cancer Genome Atlas. These efforts seek to find new ways of diagnosing, treating and understanding cancer by focusing on the genome not of patients but of the tumor itself, looking for the unique genetic “errors” that caused good cells to go bad. To be most effective, such projects will require close cooperation between researchers at all levels of science, from the laboratory to the clinic to computational analysis centers.

Fortunately, all of those elements already exist at the University of Chicago Medical Center, where a unique twist on cancer genetic research has already been underway for the last year. The Chicago Cancer Genome Project, a collaboration that is bringing together experts from virtually every corner of the Medical Center, is an ambitious effort with a local focus, a merger of biology, computational science and medicine that could reshuffle the how cancer is categorized and how new treatments are discovered and tested.

“It’s a long road from having a piece of genome sequence to actually implementing that into patient care, but the path is at least clear enough that we can shine a light down it and see the end,” said Kevin White, professor of human genetics and ecology & evolution. “If we know enough about the genetic basis of the tumors we should be able to at least connect the dots and draw a path toward clinical trials, and if those work out, toward actually getting things into the clinic.”

In the first year of the project, White’s group has sequenced more than 100 tumors, predominantly from patients with breast or head & neck cancer. Rather than sequencing the entire genome of a tumor cell, White says their current strategy is to decode the “transcriptome,” only sequencing the genes that are actually expressed, or turned on. Combined with the rapidly decreasing cost of genetic sequencing – currently, each tumor only costs about $2,000 to sequence, White said – the project hopes to gather sequences from at least 1,000 tumors in the next 3 years.

Those tumors will come from the biopsies of (consenting) patients seeing cancer experts at the Medical Center, such as Kevin Roggin, a surgeon who treats pancreatic cancer, and Tanguy Seiwert, an expert on head & neck cancers. As sequences are gathered, researchers will compare them in a number of different ways:  whether a tumor responded to treatment or not, what kinds of treatment worked or didn’t work, whether it became metastatic and spread to other parts of the body. Such analysis could create new sub-categories of tumors within organs such as the pancreas or the lungs, classified according to what mutation is at the root of the tumor.

In types of cancer that have already been split into clinical categories – such as the triple-negative vs. hormone-receptor-positive tumors seen in breast cancer – genetic information could explain why different tumors have different characteristics. Eventually, White said, sequencing each individual tumor could point directly to an individualized treatment plan that is most effective for treating a person’s cancer while minimizing side effects of the treatment.

“It’s probably infinitely divisible at some point so that every cancer is its own story,” White said. “The first question is how can we lump things into useful categories.”

Those categories may include a new way of classifying cancer by each tumor’s genetic cause rather than its tissue of origin. For example, White said, an analysis may find that certain head & neck cancers are more like estrogen receptor-positive breast cancers than other head & neck cancers, suggesting that drugs developed to treat certain breast cancers may be effective in a previously unrecognized group of patients with cancer in a different part of the body.

“There are likely some common mutations that make some cancers between tissues more like one another than cancers within tissues,” White said.

All of these findings would be expected to change the way clinical trials are run and analyzed, allowing experimental treatments to be targeted towards sub-groups of populations where they are expected to be more effective. Some current treatments, White said, may have failed clinical trials because patients with cancer in a particular tissue were grouped together, without regard for genetic differences that could make those cancers differently responsive to the treatment.

But while the Chicago Cancer Genome Project will use only tumor samples from the Medical Center, White emphasized that its results will not be isolated from national research efforts to create a massive genetic database for cancer. Part of the computing challenge, which is being addressed with Robert Grossman from the University of Illinois at Chicago, is to prepare the computational systems to integrate the data from the 1,000 Chicago tumors with the torrent of data expected to come from the National Cancer Institute and other projects.

“The big picture is, you’ve got all this data, and then you’ve got the local investigations going on, so how can you use both the world’s data and our local data combined to enhance what we’re doing here at the University of Chicago,” White said. In the meantime, “we can take a smaller number of genomes, but in a more focused way so that we ask questions that are of clinical relevance.”

About Rob Mitchum (526 Articles)

Rob Mitchum is communications manager at the Computation Institute, a joint initiative between The University of Chicago and Argonne National Laboratory.

%d bloggers like this: