The brain is, arguably, the most complex organ in the human body. Scientists who stuck to the so-called "low-hanging fruit" in their genome-related research have avoided tackling brain studies at all costs.
The only explanation, then, is that most scientists have in fact not pursued the low-hanging fruit. After cancer research, neuroscience represents what could well be the next biggest slice of the pie in the world of genomics, proteomics, and bioinformatics experts. Two NIH institutes focus exclusively on this area — the National Institute of Neurological Disorders and Stroke and the National Institute of Mental Health, which together issued more than $2.3 billion in grant funding in FY 2005 — while several others, such as NHGRI, the National Institute on Aging, and the National Eye Institute, to name a few, are members of the neuroscience initiative at NIH.
According to NINDS, there are more than 600 disorders known to affect the nervous system; NIMH, meanwhile, points out that in the US alone more than 26 percent of the adult population can be diagnosed with a mental disorder. No wonder, then, that scientists around the world and in all disciplines that comprise systems biology have leaped into brain research.
From RNAi to sequencing to gene expression, the technologies widely used in the systems biology field are being brought to bear in neuroscience. In this article, Genome Technology provides a sampling of innovative research using a number of these technologies in pursuit of a better understanding of the brain.
UCSC team identifies new gene, links it to human cortex development
Katie Pollard knew she had asked a clever question, but even in the best-case scenario she envisioned could the results have been this noteworthy.
Pollard was a postdoc in David Haussler's University of California, Santa Cruz, lab working in — what else, when Haussler is involved? — comparative genomics. Three years ago, she saw an opportunity to get involved in the chimp genome project, where a lot of attention was going toward protein-coding genes. "There aren't that many differences between the proteins between chimps and humans," she says. "So I wanted to look at the whole genome, including everything that people might think was junk DNA." She wrote a computer program to march through the chimp and human genomes with a fairly simple command: "Find pieces of DNA where a chimp looks more like a rodent than it does like a human." That is, she took the intersection of DNA regions that were under heavy negative selection before the human-chimp split and regions that are most divergent between human and chimp.
Pollard's hunch, as well as Haussler's longtime stumping for the importance of so-called junk DNA, paid off. Of the top 49 regions that the query produced, Haussler says, "only two landed in protein-coding regions."
The first region on the list proved to be an extreme case — and a brand-new gene. What Pollard discovered "is apparently an RNA gene," says Haussler, who along with his team named it HAR1, short for Human Accelerated Region. The 118-base gene had been so highly conserved that a chimp and a chicken differ by just two small changes. Compare that 300-million-year reign of stability to its fate in humans, where in the mere 5 million years since the human-chimp split there have been 18 changes. "That's a very significant speedup," Haussler says.
Early investigation into HAR1 indicated that it might play a role in neurological development. "That really motivated us," says Pollard. Haussler's team worked with Pierre Vanderhaeghen, a neuroscientist at the University of Brussels, to find that the gene expressed in a special group of neurons during weeks seven through 19 in human embryonic development. That happens to be the same timeframe, and through the same neurons, that the cerebral cortex is being developed, Haussler notes. "Its expression pattern is remarkably similar to that of reelin [a fundamental protein involved in developing the cortex] in many ways," he says. "That's a very strong suggestion that it might be involved in this process."
Currently, the Haussler lab is hard at work studying the biochemistry of HAR1 "to determine what it interacts with in the cell," he adds. An ongoing collaboration with Vanderhaeghen aims to pursue the gene using functional studies in a mouse model. Meanwhile, Pollard has relocated to UC Davis, where she is an assistant professor of statistics. She has no plans to abandon HAR1 either, but says her focus will remain on the computational side. She notes that HAR1 was just the first region on the hit list generated by her project; she has a host of other regions to track down, and many of those appear to be located near other genes implicated in neurodevelopment. "There's a lot more there besides just HAR1," she says.
Genome for a song: To know how the caged zebra finch sings
Bird brains are considerably more complex than the grade school put-down would have you believe. Songbirds, one of the few non-human animals to exhibit vocal learning, provide a particularly rich model for scientists interested in a range of neurobiological phenomena — from molecular events underpinning plasticity of the central nervous system to big questions about development and the very nature of learning and memory.
David Clayton has his eye on all of these topics as he forges ahead in a collaborative project to sequence the genome of the male zebra finch — a sex-specific subset of an avian species known to generate new neurons throughout adulthood.
Clayton, principal investigator of the collaborative Songbird Neurogenomics Consortium, traces his interest in songbird biology back to his PhD days in James Darnell's lab at Rockefeller University. Clayton was already interested in "applying all of this technical stuff — what has now become genomics — to studies of the brain." Darnell advised starting off with a simpler system, so Clayton spent the bulk of his time studying mouse liver gene expression and cell differentiation.
Then came a seminar on the lateralization of song learning in the canary brain by Fernando Nottebohm, Rockefeller's pioneer of songbird neurobiology. Clayton and Darnell emerged from the talk thrilled with the prospect of examining gene expression in what was presumably a model system for within-subject analyses. Clayton started work on his "Saturday night project" investigating songbird gene expression while finishing up the mouse liver project. He was hooked. After graduation, he established his own small lab at Rockefeller to continue working on the marriage of molecular biology with the songbird model.
Now a professor of cell and developmental biology at the University of Illinois, Clayton says the focus of songbird neurobiology for the past 15 years has been the search for genes differentially regulated in various contexts. In the pre-genomic era, two genes received most of the attention — the songbird homolog of the alpha-synuclein gene and FoxP2, both of which have implications for human learning and language effects.
In the process of doing correlational experiments targeting those two genes, Clayton says it became clear that the genes themselves were parts of much larger, more complex networks. "The Songbird Neurogenomics Initiative was born as recognition that all people working with songbirds could benefit by having access to more high-throughput technologies," he says.
For the first step to this end, Clayton's group whipped up 1,000 microarrays and issued a call for experimental proposals from the songbird research community a couple of years ago. Interested researchers had only to provide the RNA and tissue; Clayton's team offered to do the rest for free. The resulting data, for which proposers have first publication rights, will go into databases from which the community can try to extract a superstory. "We expect by the end of this calendar year to have collected all those data from the 1,000 arrays," Clayton says.
One goal of the microarray project was to justify the complete sequencing of the zebra finch genome. It turns out that NHGRI's interest coupled with the final days of the favorable federal funding climate prompted Clayton to pull together a white paper last July. The proposal was accepted, and the songbird sequencing consortium was born. The whole shotgun sequencing and assembly of the zebra finch genome — itself estimated to be around 1.7 billion base pairs — is taking place at Washington University's genome sequencing center, also responsible for the recently completed chicken genome.
"Once we get the whole zebra finch genome, we'll have really high-quality information that annotates chicken, and chicken that annotates zebra finch," Clayton says. "All of this gives a better platform for considering the human genome and relevance relationships to other branches of the evolutionary tree."
RNA expression profiling
Swiss team aims for personalized approach to MS treatment
There's no shortage of data on the prognostic and developmental factors of multiple sclerosis. But when it comes to accurately predicting what course the disease will take in a particular individual, and what that individual's treatment response might be, clinicians are still very much in the dark.
"To [make a prognosis for] this disease is difficult because it is so heterogeneous," says Raija Lindberg, a leading researcher at the Department of Research and Neurology in Basel, Switzerland. "All the patients respond differently to different treatments."
Lindberg and her colleagues are attempting to eliminate this uncertainty with RNA expression profiling of brain tissue and peripheral blood samples from MS patients. By identifying altered physiological pathways at different stages of the disease's development, they hope to derive new approaches for individualized treatment. "We are doing expression profiling at different levels," says Lindberg. "Earlier, we studied the brain, but after that, we expanded our expression profiling on the peripheral blood and we tried to find prognostic and diagnostic markers," she says. "Next, we are planning to combine these two compartments and find a link to what is happening in the brain and in the peripheral blood, so we are focusing on the cerebral spinal fluid, which is kind of in between the two," she says.
Recently, Lindberg contributed to a two-year study aimed at developing a personalized approach to the treatment of early-relapsing MS. BEST PGx, which began in 2004, is geared toward determining which genes respond the best to interferon beta 1b, the most widely used medication for the disease. This study may also reveal new data regarding the underlying variations in immune system pathology as it relates to the disease.
A major challenge with researching MS is the same as most other neurodegenerative research: access to samples is limited to postmortem brain tissue. "It is very seldom that a biopsy is taken from MS patients," Lindberg says. "The only [brain] tissue materials we have are from autopsies, and for RNA expression, that's always a little complicated and not so reliable."
Lindberg believes that in the long run, a multi-tiered approach is the best plan of attack. "In my opinion, you have to look at all these different levels: pharmacogenomics, but also genomics, proteomics, and metabolomics," she says. "We are also expanding our expression profiling to the protein level, and we have done metabolomic studies. … That is the way to go.
HUPO Brain Proteome Project emerges from pilot phase
The Human Proteome Organization's Brain Proteome Project pilot studies are complete, and the results may already be in a nearby copy of Proteomics. But this is just a first step for the project, which seeks to accomplish even more in the year to come.
The project's central goal is to fully characterize both mouse and human brain proteomes to get a handle on the polymorphisms and modifications in normal brains, as well as disease-related proteins associated with neurodegenerative disease and aging. Led by Herbert Meyer at the Medical Proteom-Center in Bochum, Germany, the BPP is comprised of an international set of participants from North America, Europe, and Asia.
So far, project researchers have used differential protein expression profiling, as well as a range of biomarker discovery and validation techniques, to pin down the travels of all proteins in the brain. In one pilot study, collaborators compiled quantitative analyses of normal mouse brain throughout several stages of development.
This includes proteome-transcriptome comparisons performed via mRNA profiling, as well as quality assessments of gel- and non-gel-based results. Human samples, both post-mortem and surgically resected disease tissue, have been put through similar analyses.
Proteomics is distinguished by the vast amount of data that its researchers generate, and this is especially so for those investigating a structure as complex as the brain. "If you generate data based on different instrumentation and different techniques, it's not a surprise to get a lot of different results," says Herbert Thiele, director of bioinformatics at Bruker Daltonics and a researcher with the BPP.
The HUPO BPP bioinformatics subcommittee also realized this early on, and thus adopted the use of ProteinScape, Bruker's centralized proteomics project management system for data warehousing and management. ProteinScape serves as a common software platform for the entire BPP community, thereby obviating the source of typical translation problems for data generated in different labs, Thiele says.
The platform can process data to be compatible with HUPO's PSI mzData standard, which enables all projects to be archived according to one common standard. Locally derived results can thus be transferred to a central server managed at the Medical Proteom-Center's data collection center, Thiele says. Once there, the processing of data is accomplished with the help of Bruker's ProteinExtractor, which Thiele says can generate lists of true, non-redundant proteins based on MS/MS searches.
NIMH targets $3 million for microRNA research into mental disorders
With a $3 million pledge from the NIH to fund research investigating the role of microRNAs in psychiatric disorders, RNAi technology is officially making its mark in neuroscience.
"We're always thinking about what's the next hot area that offers hope to elucidate the mechanisms that underlie mental disorders," says Thomas Lehner, acting director of the Office of Human Genetics & Genomic Resources at the National Institute of Mental Health. Lehner points to several papers that helped seal the deal for the funding, including one that found a link between Fragile X mental retardation and the microRNA pathway. It is believed that microRNAs, although about only 20 nucleotides in length, regulate roughly 20 percent of our genes and are responsible for regulating the expression of thousands of target mRNAs. By integrating sequence-specific modulators of post-transcriptional gene expression into a theoretical framework of disease pathophysiology, researchers will attempt to shed light on the biological mechanisms involved in mental disorders such as autism and schizophrenia.
The NIMH, which expects to award up to seven grants next year, has had no shortage of proposals. "The response has been almost overwhelming," says Lehner, who's quick to point out that in such a new area of mental research, there really are no rules. "The field is wide open. We understand so little and there's so much to be done, particularly in gene expression in the brain around regulation, using non-coding RNAs," he says. "This is all very new."
The impetus for the funding arose from an increasing frustration with the limitations of standard studies to understand the genetics of mental disorders. "We've been trying so long with conventional studies to understand the genetics of mental disorders," he says. "I think we've made great strides, but there are still issues of not being able to replicate, and the identified genes probably only explain a small portion of the variation."
Lehner is cautiously optimistic about the potential of microRNAs to unlock the mysteries of mental pathologies. "I don't believe [microRNAs] will be the answer ... but I think they'll provide an important additional piece that we need in our ultimate understanding of the pathophysiology of mental disorders," he says.
Neurosurgery research sheds light on malignant astrocytoma
With an eye always on the clinic, no doubt because his lab is literally next door to the bedside, Gregory Riggins' investigations are well placed to find the most promising molecular targets for the highest grade of gliomas — glioblastoma multiforme.
Glioblastoma isn't only the most malignant type of astrocytoma, it's also the most common. The highly invasive tumor condemns patients to an average survival time of a year or less. Clearly, there is a need for more effective therapeutics. Riggins, director of neurosurgery research at Johns Hopkins School of Medicine, is working to harness genomic technologies to do just that.
Riggins leads a lab with a mission to identify and evaluate molecular targets for new brain tumor therapies. By bringing the relatively new techniques of cancer genomics to bear on small molecule cancer screens, he uses technology both classic and cutting edge to uncover the inhibitors of pathways that are found to be mutated or activated in brain tumors. One such method that Riggins has used in his hunts is SAGE, the serial analysis of gene expression, which yields profiles of gene expression based on the number of transcripts present in tumor samples. Such profiles can be searched for candidate prognostic markers or antigens, he says.
"I think the fundamental question that cancer genomics can answer now with present technology is, ‘What is the entire mutational spectrum within brain tumors — or really any cancer?'" Riggins says. It's not a trivial question. Identifying acquired or inherited mutations can lead the way to novel (and relatively accessible) molecular targets, which Riggins and his collaborators aim to defuse via whatever pathway is most effective. Moreover, targeted therapies for cancer are few and far between — especially for brain cancer.
Last year, however, a multi-institutional team led by Riggins and Bob Strausberg of the Venter Institute reported their success in generating the first comprehensive sequence analysis of the receptor tyrosine kinase gene family in glioblastoma tumors. This family of genes was found to play a key role in signaling between brain cancer cells and their microenvironment, which means that they might serve as attractive molecular targets for therapeutic interventions.
Currently, Riggins and his collaborators are continuing to build on the kinase sequencing project with a new NIH-funded glioblastoma project. The team is set to perform high-throughput mutational analyses of gene families, and specifically those characterized by products that can be targeted with pharmaceuticals. This complements the genome-wide screening for amplifications or deletions in glioblastoma genomes, which Riggins says may highlight "additional genes with amplifications that may lead us to activated pathways."
DiaGenic has its eye on clinical validation for Alzheimer's test
Currently, diagnosing a patient with Alzheimer's involves mostly guesswork and inferences drawn from cognitive evaluations and memory tests. Clinicians can only know for sure if they were on the right track after postmortem brain tissue samples have been analyzed. But a new blood-based gene-expression profile test from DiaGenic aims to change all that.
"With Alzheimer's, there's really no good diagnostic tools available and [it's] a disease that will more or less explode due to the aging population," says Dag Christian Christiansen, DiaGenic's director of marketing. "And now, there's more and more documentation that, with early diagnosis and early intervention, you [can actually] delay the decline of cognitive functions."
The company's new diagnostic test shows that the gene expression signature can accurately distinguish between healthy individuals and those suffering with Alzheimer's, as well as individuals with Parkinson's disease. DiaGenic's study used 330 blood samples, including 125 taken from recently diagnosed patients, to create an expression model for the disease. An independent test set validated an accuracy of 87 percent.
This could be a major advance in the fight against a disease that affects roughly 4.5 million Americans, a number expected to grow to as many as 16 million by 2050, according to the NIH. The test, which will initially be marketed as a research-only assay, will be ready before the end of the year with clinical validations slated for early 2007. DiaGenic is also developing a gene expression test to identify the disease at its onset.
"Gene expression technology has the potential to establish a diagnosis at a very, very early stage, perhaps before other clinical signs are clear for the physicians," says Christiansen. He predicts that five to 10 years down the line, an increasing number of healthcare providers will have access to similar blood-based diagnostics. "We are convinced that there will be lots of these tests available, first in the hospitals, and then later, as a point-of-care or physician office test."
Although the test does not yet point to specific treatments, the technology does hold that promise for other neurodegenerative disorders. "I clearly see that this technology has that potential," Christiansen says. "There are publications indicating that this technology can be used to identify [MS patients] who will benefit from beta interferon treatment versus those who will not."
Harnessing regenerative medicine for CNS disease
When disease or injury strikes the central nervous system, therapeutic options are limited at best. Multipotent neural stem cells thus hold great promise for researchers looking to repair damaged tissue or replace dwindling cell populations. Mahendra Rao is one such scientist whose career has perpetually been at the forefront of investigations concerning the biology and regenerative potential of stem cells in the CNS.
Rao took on the post of vice president of research, stem cells, and regenerative medicine at Invitrogen earlier this year. Prior to this position, he served as the section chief for stem cells at the US National Institute on Aging, where his work on neural stem and progenitor cells was firmly rooted in the idea that the same signaling mechanisms regulating the nervous system undergo alteration during aging and age-related disease.
In the last few years, knowledge of stem cell biology has grown rapidly in the wake of large-scale genomic research. The availability of diverse array platforms has resulted in gene expression studies finding markers unique to neural stem cells, or to cells in a particular disease state. Extremely powerful investigations are now possible because of the existence of a well-annotated mouse genome, Rao says, in addition to the ability to create transgenics to test specific hypotheses in live systems. "We can now look at genomics biology in situ, which we simply couldn't do before," he says.
But so far, the clinical realization of regenerative medicine using neural stem cells has been slow. This is partially due to the complexity of CNS biology itself, Rao says, in addition to the need for non-autologous sources of cells. "There aren't signals telling a neural stem cell what it needs to do once transplanted," he says, "even if you can get them in large enough numbers." Food and Drug Administration rules also state that the transplantation of foreign cells ought to follow the standard investigational new drug process. All of these factors have colluded to slow the march of neural stem cells in clinical settings.
Invitrogen's mandate is to speed up all of the bottlenecks in stem cell regenerative medicine, and Rao's lab focuses on ensuring that the tools and reagents are in place to do so. His research team is working to optimize processes across the gamut of stem cell lab work — from cell manufacture processes to animal testing to clinical grade tests — using Parkinson's disease and transverse myelitis as models. The work has already resulted in a product: defined media for embryonic stem cells, which Rao estimates will become available by the middle of next year
Iowa's Davidson uses RNAi in treatment of dominant brain diseases
RNAi showed up on Beverly Davidson's doorstep as the answer to a question she had asked for more than 10 years.
A biological chemist who studied at the University of Michigan, she began early in her career to work with rare neurological diseases, beginning with one called Lesch-Nyhan Syndrome, which centered around a mutated protein. "The thing that intrigued me most was the clinical manifestation of what happened when that protein was mutated," she says. At the time, she wondered whether molecular methods could be used to treat these diseases even before they were fully understood.
Now a professor at the University of Iowa, Davidson homed in on recessive genetic diseases that showed up in childhood onset, and her lab focused for a decade on "developing novel methods of delivering stuff to the brain" — particularly in rodent models. But in the back of her mind, she kept asking about dominant diseases. What was never clear to her: "How do you take out something bad instead of add something" that's missing?
Several years ago, "RNAi gave us our first clue as to the methods that we might use to tackle dominant inherited brain diseases," she says. And with that technology, the results of her 10 years of work fell into place. "It was … a marriage of molecular biology with some of the recombinant viruses that we had already made," she says. "When we put the two of those together, we had a system that we could use to rapidly test efficacy in vivo."
Davidson's 30-person lab now splits its time fairly evenly between dominant and recessive genetic brain diseases. In the next few years, she hopes to see some of the work she and her team have done "translate into human trials," she says. In the meantime, they're adding new avenues of research as they appear on her radar. MicroRNA research is one such path, she adds; her team is already looking into "the pathophysiology of how some of these mutations may be leading to the disease characteristics we see in patients."
Allen Brain Atlas completed; brain browser ready for use
There's a new browser online. It maps, it zooms — and no, it's not Google Earth, but it may help you to think of it that way.
The Allen Brain Atlas is the result of a three-year tour de force from the folks at the nonprofit Allen Institute for Brain Science located in Seattle. In that time, institute staff mapped the entire transcriptome of the mouse brain using in situ hybridization, gene expression, and imaging technology on a gene-by-gene basis, and then rendered it all in an interactive, 3D tool. (See our informatics coverage of this.)
From the start, public access to data was high on the priority list, says institute Chief Scientific Officer Allan Jones. "We're really dedicated to making the tools and the information that we create readily available to the scientific community," he says. To that end, data from the project has been released along the way, starting in December 2004 and culminating in the final push of data scheduled for the end of September this year. In fact, he notes, the Allen staff had to create this atlas from the ground up instead of adding data to an existing brain atlas in order to avoid any of the publishing and other restrictions common to most repositories of this kind.
The finished product, which Jones notes will continue to be annotated by the Allen team, comprises 21,000 genes, takes up 600 terabytes, and necessitated the use of 250,000 microscope slides each holding several 25-micron slices of mouse brain. The data can be accessed by downloading a tool called Brain Explorer, which (much like Google Earth) facilitates searching and viewing the data. Users can select genes of interest and zoom in to the cellular level to see where in the brain they're expressed — regions of the brain are color-coded for easy viewing — as well as related heat maps and primary data.
Because of the data release policy, a number of scientists were using this information long before the atlas was completed, Jones says. The site has been averaging 4 million hits per month, with about 250 scientists from academia, pharma, and biotech using it on any given work day.
But the big release won't mean a long vacation for the 85 people at the Allen institute. Jones says the team is eager to move beyond the mouse model, for one thing. "We're really keen on moving into human studies, very specifically in the neocortex of the brain. That offers a lot of unique opportunities." Diseases such as epilepsy, autism, and schizophrenia have been linked to the neocortex, Jones adds.