It's been just three years since the Allen Institute released its mouse brain atlas — a free, publicly available online tool that allows users to zoom in and out of a 3D brain, search by gene, and see expression data — but in that short time, this revolutionary look at the inner workings of a brain has dramatically changed the way neuroscientists do business.
Take Greg Foltz, for instance. A neurosurgeon at the Swedish Neuroscience Institute, Foltz studies glioblastoma mutliforme; for him, the glut of new data on this particular kind of tumor from the Cancer Genome Atlas project has been both an amazing resource and a cause of frustration. The "tremendous amount of sequencing data available across hundreds of patients" — while a significant first step, Foltz says — doesn't address a key issue in heterogeneous glioblastoma tumors: location. In these types of tumors, knowing whether a gene is expressed at the core of the tumor or the leading edge of the tumor can be critical to deciding how to treat a patient.
Because of that, Foltz uses the sequencing data and compares it to, of all things, the Allen Institute's brain atlas for mouse. He and his team have made extensive use of the atlas and believe that mouse brain is a fair approximation for what goes on in human brain, he says. Foltz can take the gene expression levels seen in the glioblastoma-specific data and compare them to the mouse brain for a sense of "what genes appear to be expressed in an increased fashion or silenced," he says. "That's been really useful for us. If you think about it, we have no other way of making that comparison."
Indeed, Foltz says that even that level of information helps his team partition patients into groups based on likely treatment course and predicted outcome. Of course, what he really wants is to have a brain atlas for human — and one that would reflect the changes seen in glioblastoma patients.
Foltz is getting his wish. Thanks to a three-year, $2.1 million grant from the Ben and Catherine Ivy Foundation, Foltz will be teaming up with the Allen Institute on the Ivy Glioblastoma Atlas Project. The atlas will be based on tumor samples from 32 patients and will start with a focus on 300 genes selected "because they've already been known to have increased expression in glioblastoma," Foltz says. He also plans for a comparison to non-tumor brain, using around 20 samples from operations where patients with other neurological conditions, such as epilepsy, have had part of their brain removed. When the atlas is complete, it will include clinical information on the patients and the tool will be made freely available. "It is going to be absolutely fascinating to see," Foltz says.
The human angle
Over at the Allen Institute, meanwhile, the glioblastoma project might be considered small compared to what these researchers have set their minds to: developing a human brain atlas.
This is no pie-in-the-sky goal. After more than a year of planning, institute scientists are gearing up to begin the human brain project, which is expected to cost around $55 million, include two phases — the first a macro-level look and the second focusing more on the cellular level — and be completed in 2012, says institute COO Elaine Jones. The institute is currently trying to raise $10 million to support the initiative from a variety of sources, including government agencies, private sources, and philanthropies, she adds.
As most people are reluctant to give up their brains, the study will be performed on cadavers — anywhere from four to 10, with a mix of males and females, says Elaine Shen, manager of the human brain atlas project. She notes that the end result will be multimodal, including magnetic resonance and classic histology information in addition to the micro-array analysis and in situ hybridization data. "All of the gene expression data is going to be mapped back into 3D space," Shen says. With all of that information in one place, she adds, the hope is that the new brain atlas will be far more useful to clinical researchers.
The human brain is approximately 2,000 times the size of mouse brain, Shen says, which in itself has created major challenges for a pipeline designed around brain slices that could easily fit on a typical laboratory slide. Other hurdles have come from using fresh frozen tissue: "So many of the stains have been optimized for fixed tissue," Shen says. The brain sectioning process is typically easier in fixed tissue as well. As the institute has geared up for this project during the past year and a half or so, scientists there have become experts in adapting protocols to work with fresh frozen tissue. The team also had to rebuild the informatics side of the pipeline to accommodate all the new clinical data as well as the microarray data. "The scale's so much larger," Shen says. "A project of this kind of scale and scope has not really been tried."
In the first phase of the atlas project, researchers will collect about 1,000 samples per brain and perform microarray analysis on them as a way to "systematically survey through the brain," Shen says. The second phase will include the ISH part of the study and will "very likely concentrate on specific genes and specific structures."
Shen notes that a key phase in planning for this project involved reaching out to the neuroscience community. The goal was "to understand from them what they thought, one, would be needed [by scientists in the field], and, two, the challenges and expertise that would be required," she says.
The choice of human brain for the institute's next major project was no accident. According to Jones, the institute plans in three- to five-year phases, based in large part on the recommendations of its scientific advisory board. "We are looking at what's the next big, high-impact project that we should be working on that will have a major effect on neuroscience … that nobody else will touch," Jones says.
Sidebar: MicroRNA Expression Sheds Light on Schizophrenia
A team of scientists at the Schizophrenia Research Institute in Australia have used studies of microRNA expression to determine that gene silencing appears to play a key role in patients with schizophrenia.
Murray Cairns, senior research fellow at the institute, worked on the project involving a cohort of postmortem samples from schizophrenia patients and controls. "After doing a lot of gene expression profiling, we noticed that there was a general trend — more down-regulated … genes than upregulated. We thought maybe this was gene silencing." That finding led to the current study, recently published in Molecular Psychiatry, in which "microRNA expression seemed to be globally elevated in the schizophrenia cases compared with the controls," Cairns says.
To follow up on that, Cairns and his colleagues looked at the primary and precursor microRNAs of some of the differentially expressed microRNAs, finding that "the mature and precursor forms were ... upregulated in schizophrenia but the primary transcripts didn't seem to be altered," he says. That suggested some kind of post-transcriptional alteration, which the team is currently investigating.
The researchers are using microarrays and next-gen sequencing in their latest studies. "We're looking at some other brain regions and we're also trying to do some laser capture and identify the actual cell types," Cairns says. They're adding some functional genomics approaches as well to get a better sense of the pathogenesis of the disease.