9 May 2003. On May 4-5, Harvard Medical School briefed reporters on new directions in biomedical research. The meeting featured diverse presentations ranging from basic research on RNA splicing and clinical trials of proteasome inhibitors to the evolutionary role of genes involved in intelligence and mental retardation, and more. Here are selected tidbits of interest to the neurodegeneration community.
Bernardo Sabatini of Harvard Medical School outlined the development of new technologies to study dendritic spines, the tiny bulges on a neuron’s dendrites that form the cell’s input regions as axons from other neurons synapse onto them. These spines are the site of synaptic plasticity, and are also likely sites of early dysfunction in Alzheimer’s, but they are difficult to study in the densely packed environment of the brain. Their diminutive volume of one femtoliter per spine, and their constant dynamic change and turnover within minutes (see ARF related news story) don’t make the task any easier. Sabatini is currently devising methods to study individual spines not in dissociated neurons, which have lost the connections and neighbor relationships of the live tissue, but in cultured slices that maintain synapses with connecting neurons at least in close and medium range for a few days.
To do this, Sabatini is combining two-photon microscopy, a method now widely used in neuroscience, including in AD research, with whole-cell recording, a method in which the scientist pokes a small hole into a cultured neuron with a glass pipette to establish a physiological continuum that allows one to introduce chemicals into the neuron and record the electrophysiological response. Previously, Sabatini has used whole-cell recording to study the neuron’s response to changes in calcium dynamics in dendritic spines (see ref. Sabatini et al, 2002).
Doing this in tissue slices, however, is difficult and tedious, Sabatini said, because finding the axon that forms synapses onto the dendritic spine from which one is recording is like looking for a needle in a haystack. To overcome this problem, Sabatini has adapted photoactivation technology, or uncaging, to study individual spines. This technique relies on an inactive derivative of glutamate that contains a large photosensitive moiety covalently attached to the glutamate backbone. Called MNI-glutamate, this derivative can be cleaved by photons to release the neurotransmitter. Sabatini bathes tissue slices in a solution containing the derivative, and can then activate spines in highly localized areas with focused light from a laser. Sabatini calls this device the dual laser scanning microscope, and he uses it to mimic synaptic transmission around the recording electrode. He found that it is stable, produces reproducible EPSPs, and should be useful in studying spatial relationships among synapses as, for example, in elucidating if those spines that are close to each other are more likely to "talk to each other."
Sabatini also described a method of culturing brain slices and generating local transgenic patches by shooting in gold particles covered with a transgene of choice with the help of a helium-beam gene gun. This allows one to study the effect of a given gene in a small area of adult brain, a more specific assessment of gene function than standard transgenic or knockout procedures, which generate more wide-ranging changes.
Among the topics in AD research that could be tackled with these methods are the role of presenilin, and its mutated forms, in perturbing calcium flows in the neuron, and the consequences of that in individual dendritic spines, Sabatini said.
Chris Walsh of Beth Israel Deaconess Medical Center described recent work on genes that not only control the formation of the cerebral cortex, but are also involved in mental retardation and microcephaly. Most of these genes are yet to be discovered, but Walsh hypothesizes that they may be the ones on which evolution acted to produce larger brains. Recent examples include the gene locus MCPH5, identified in a Pakistani kindred (see Roberts et al., 2002) and the gene ASPM (Bond et al., 2002). Expressed in dividing stem cells in the brain’s ventricular zone, ASPM is intriguing in that the roundworm, fruit fly, mouse, and human versions differ mainly by the increasing number of a region, aptly dubbed IQ domain, from two repeats in C. elegans to 71 in humans, Walsh said. The ASPM protein is suspected to control the mitotic spindle during neurogenesis.
Also affecting brain size is the protein β-catenin, which numerous groups have implicated in Alzheimer’s. For example, Eddie Koo’s lab at University of California, San Diego has linked presenilin to β-catenin phosphorylation and cell proliferation (see Kang et al., 2002). This points to a speculative role of presenilin in neurogenesis, where presenilin gain-of-function mutations might undercut neurogenesis via effects on β-catenin, and insufficient presenilin would lead to excessive proliferation.
Walsh told reporters that during development, β-catenin influences the switch between asymmetric cell division at the ventricular zone, in which one daughter cell goes on to differentiate, generating fewer cells but more mature neurons, and symmetric cell division, in which both daughter cells continue to divide, yielding higher cell number but fewer mature neurons. Overexpressing β-catenin creates mice with a large cerebral cortex that is folded in like the human cortex. (The cortex in wild-type mice is smooth.) Yet, while these mice have many more dividing cells, they do not have more differentiated neurons than do normal mice, Walsh said, and they die soon after birth.
Steven Gygi of HMS described proteomics studies using tandem mass spectrometry for fast protein sequencing. This technology greatly increases the capacity to study the large protein complexes that carry out most biological processes. Tandem mass spec also enables the large-scale comparison of protein expression in normal vs. diseased cells, and it helps with identification of posttranslational modifications on proteins. These research areas deserve much attention in the study of neurodegeneration, as well.
Gygi’s research focuses on ubiquitin, the most conserved protein known. To date, most scientists associate ubiquitin with protein degradation. This tightly regulated process is at the center of normal cell physiology, but also has been implicated in cancer and most major neurodegenerative diseases. Ubiquitin chains form on amino acid 48 of ubiquitin and are attached to a target protein to tag it for destruction in the proteasome, a protein-grinding machine inside the cell. This attachment is where the specificity arises; there are 600 different E3 ubiquitin ligases known to date, said Gygi. (Parkin, for example, is one (see Shimura et al., 2001). Yet, protein degradation is no longer the only basic cellular process in which ubiquitin figures prominently, Gygi said. Proteomics studies have shown that chains can also form on amino acid K 63 of ubiquitin, and these chains play a still-mysterious role in DNA repair, possibly in recruiting DNA repair enzymes, Gygi said. Yet, other types of ubiquitin chains can form, as well, he added, raising the possibility that degradation is but one of the hats this protein likes to wear.
Peter Lansbury of Brigham and Women’s Hospital reinforced how difficult it will be to exploit for actual drug discovery the drug targets that genomics and proteomics research are generating. He said that the difference between a potential target arising from basic research and a confirmed target is huge; it presents a hurdle most academic labs are unprepared to clear. In their drive to minimize the risk and cost of development, most pharmaceutical companies require validation of a target in the human. Once screening against a target has begun, going the distance from a compound that hits the target to a refined compound that does not do much else (i.e., a drug that is safe), can take five to 10 years. This unglamorous process scuttles most initially promising compounds.
Some universities have set up drug discovery laboratories that can assume the scientific and some of the economic risk of early drug discovery. They work to partially validate targets and generate early drug leads that a pharmaceutical company may then take over. This is key, particularly for orphan diseases that affect fewer than 200,000 patients, which includes Huntington’s disease, ALS, and other neurodegenerative diseases, though not AD and PD.
Robin Reed of Harvard Medical School summarized recent advances in understanding RNA splicing. This process clips introns out of the raw pre-mRNA that peels off the DNA polymerase during transcription and stitches the exons together into a continuous, mature mRNA that is then fed into ribosomes for translation. Like most biological processes of interest, this task falls to a large protein machine, the spliceosome, which recognizes specific sequences at the beginning and end of each intron. Most of these sequences are not yet identified, however, and the field is just beginning to understand how important it will be to do so, Reed said. Why is that? It turns out that 15 percent of genetic diseases are likely due to splicing mutations. These include breast cancer, spinal muscular atrophy, the neurodegenerative disease familial dysautonomia, muscular dystrophy, and many others. In the BRCA1 gene, for example, a single base change that perturbs splicing is known to wipe out an entire exon, Reed said.
In basic research, recent advances in mRNA splicing are beginning to open a window to new concepts and uncharted areas of investigation, Reed added. For example, alternative splicing can generate astonishing mRNA diversity, particularly in the brain. The neurexin gene can be turned into 2,346 different mature mRNAs, the gene for Down Syndrome Cell Adhesion Molecule (dscam) can yield more than 30,000 mRNAs. Are all of these expressed? What is the function, if any, of this variation?
In characterizing the spliceosome machine that generates all these variants, Reed’s and Gygi’s labs identified 146 distinct protein components in it, 58 of which were newly identified (see Zhou et al, 2002). The scientists published their data on these proteins with live links to the GenBank database, and researchers from around the world annotated many of the unknown sequences, tying them to Wilm’s tumor, DiGeorge’s syndrome, and other diseases, Reed said. Coordinating the splicing process are about 30 coupling proteins that tether individual spliceosomes together, raising the image of intracellular factories with veritable protein expression assembly lines, Reed said.
Alfred Goldberg of Harvard Medical School, Julian Adams of Millennium Pharmaceuticals in Cambridge, Massachusetts, Ken Anderson of Boston’s Dana-Farber Cancer Center, and Peter Howley of HMS presented a line of research into protein degradation reaching from basic studies all the way to human therapy. Currently, three
proteasome inhibitors are under consideration for final approval by
the FDA for multiple myeloma and are in phase II trials for many
other forms of cancer, Goldberg said. The inhibitor VelcadeTM has shown good results in about 200 patients with relapsing multiple myeloma, Anderson reported, and the researchers are hoping for equally good results from trials for prostate and breast cancer.
Interestingly, neurodegeneration research tends to cut the opposite way. The precise pathways are unclear, but the preponderance of evidence from studies on ubiquitination and proteasomal degradation in neurodegenerative diseases indicates that problems in these conditions arise from insufficient degradation, not from too much proteasome activity. The consequences appear to be that proteins that are ubiquitinated, misfolded, or otherwise unneeded accumulate to toxic levels, or that stressed neurons die in other ways (see, for example, ARF related news story; McNaught et al, 2003; Taylor et al, 2003;Sherman & Goldberg, 2001). Basic research on the mechanisms of synaptic plasticity is also beginning to involve the proteasome in activity-dependent turnover of proteins in the postsynaptic dendritic spines (see Ehlers, 2003; Ehlers section of ARF related news story). Adams said that both Velcade and related compounds in earlier phases of development are being carefully tested for crossing of the blood-brain barrier (they generally do not enter the brain) and for neurologic side effects.
Yang Shi of Harvard Medical School, Gary Ruvkun of Massachusetts General Hospital, and Judy Lieberman of Children’s Hospital spoke on the basic biology and therapeutic potential of RNAi. Shi collaborates on RNAi projects with Zuoshang Xu, who recently led an ARF live discussion about RNAi in ALS and other neurodegenerative diseases. Shi also develops additional constructs to test RNAi delivery approaches into the spinal cord as a form of gene therapy. Shi, Lieberman, and Ruvkun agreed that targeted delivery of the RNAi poses the knottiest problem for therapy development. On the plus side, RNAi is several orders of magnitude more specific for each gene it knocks down than is antisense RNA, Lieberman said. Partly for this reason, RNAi has largely supplanted antisense projects in industry, Ruvkun added. Antisense RNA generated excitement and a flurry of pharmaceutical and biotech activity 10 years ago, but that has not yielded a product to date, though some candidates are in clinical trials. "RNAi is like antisense-except it works," Ruvkun quipped.-Gabrielle Strobel.