HD2002 – Changes, Advances, and Good News (CAG)_n

Reported by: Lisa J. Bain

Abstract:

Work continues to accelerate in Huntington’s disease research. In August 2002, 350 researchers supported by the Hereditary Disease Foundation met in Cambridge, Massachusetts to discuss progress towards finding a cure for the devastating disease. Participants represented academic labs as well as government and industry, with significant collaborations among these three research arms. Research findings were presented in several areas: Drug screening with in vitro and cellular models of HD, cellular pathogenesis, biochemical and behavioral markers in mice and humans, outcome measures in drug trials, transcriptional dysregulation, organismal pathogenesis (murder/suicide, regional specificity, somatic expansion, and genetic modifiers), and new therapeutic approaches. While presentations covered a wide variety of topics, the focus on finding effective interventions pervaded the entire conference. 

In August, 2002, 350 scientists packed the Grand Ballroom at the Royal Sonesta Hotel in Cambridge, Massachusetts to hear not only about progress in Huntington’s disease (HD) research, but about where the field is headed. Allan Tobin opened the meeting with a brief retrospective of how far HD research has come in the past 30 years. Last year, he said, a Pub Med search for articles about HD yielded 350 hits, compared to only 50 in 1971. Moreover, HD research has always been “just in front of the wave crest,” he said; going ahead and searching for the gene without waiting for the revolution in human genetics that would be brought about by the mapping of the human genome. The gene was identified in 1993 and now, ten years later, work continues to accelerate.

            But now there’s a different slant to the challenge before us, said Tobin. “The question for the next three days is how this community can not only stay in front of the wave of contemporary science but also create our own wave; a wave with sufficient force to knock down this disease.” This wave, he said, will carry the HD research community forward towards a cure. “So in addition to the outstanding work we will hear that connects huntingtin action to the latest discoveries in cell and molecular biology; in addition to the creating and creative testing of mouse models of HD; in addition to the unraveling of cellular and neural pathology, we basic scientists need to focus still more strongly on interventions.”

            The acceleration of HD research has been accompanied by an ever-increasing number of scientists attending this yearly meeting of HDF-funded investigators, not only from academic labs but from industry and government as well. Cooperation among these three research arms has become more essential as the wave of HD research swells toward clinical trials. Organizationally, this necessitated a new format to the meeting.  Fifty-eight presentations were grouped into six “Datablitzes,” sandwiched in between panel discussions on a related topic. Investigators presented a total of 157 oral presentations and posters on their research. Abstracts from all of these presentations are available on the HDF website (http://www.hdfoundation.org.)

Drug screening with in vitro and cellular models of HD

            With Tobin’s challenge to focus on interventions in mind, speakers in the first datablitz presented data from in vitro, cellular, slice culture, and organismal (yeast and C. elegans) models of HD. With protein aggregation still considered one of the primary steps in HD pathogenesis, several screens looked for inhibitors of huntingtin protein or  polyglutamine aggregation. Other assays looked at cell death, growth inhibition in yeast, and restoration of mechanosensory function in C. elegans (see Table 1). Many of these assays were developed as part of the Neurodegeneration Drug Screening Consortium (NDSC), a novel program co-sponsored by the National Institute of Neurological Disorders and Stroke (NINDS) and HDF, along with The Huntington’s Disease Society of America (HDSA) and The Amyotrophic Lateral Sclerosis Association (ALSA). This project was ground breaking in that never before had an agency of the NIH joined with voluntary agencies to expedite the process of drug discovery.

            The assays described were especially exciting in that several of the investigators described collaborations with others in this panel. For example, Bob Hughes identified a porphyrin compound that appeared to suppress toxicity in his yeast-based assay. He sent the compound to Erik Schweitzer, who demonstrated that the compound also suppressed polyglutamine toxicity in cell culture. Hughes then sent the compound to Ron Wetzel, who showed that it inhibited polyglutamine aggregation in vitro. He is now collaborating with Sally Ditzler to test the compound into R6/2 mice.

            Another interesting collaboration was described by Laure Jamot, whose company, TROPHOS, has developed a high-throughput screening instrument to be used on striatal neuron cultures. They are testing a library of 40,000 compounds, and will collaborate with three labs to test lead molecules. A second primary screen will be carried out by Christian Neri in C. elegans. Then, lead molecules will be sent to Frederic Sandou for a secondary screen. Molecules that emerge from this screen will be sent to Nicole Deglon for an in vivo screen in rats.

Table 1 – Datablitz 1 – Drug Screens

The drug pipeline

            Following the drug screening datablitz, a panel consisting of Carl Johnson, Jill Heemskerk, Brian Pollok, Paul Ko Ferrigno, and Brent Stockwell led a discussion with the audience about the drug pipeline.  Heemskerk gave an overview of the NDSC program[i]. A masked set of 1,040 compounds was tested in a total of 37 different assays. The compounds included FDA-approved agents as well as some natural compounds and other substances with known activities, including the ability to cross the blood brain barrier. No compounds worked in all assays, and there was less overlap than expected from assay to assay, said Heemskerk.  Stockwell has overseen the development of a database to analyze the results from these assays. The database includes chemical structures, known properties, and retest dose response curves. It allows investigators to look across assays and compounds for structure/activity relationships among compounds that were active in multiple assays. The database, with a usable web interface, will be available to investigators shortly and then to the public after this has been published. “It’s an interesting rich dataset to mine,” said Stockwell.

            The next step will be prioritization of these compounds to determine which should go on to be tested in mice. “We’re going to have many more drugs to test than we possibly can do,” said Allan Tobin. “It will be important for us to be rational in selecting targets that we can assess effectively.”

Erik Schweitzer, one of the NDSC collaborators suggested that it may be wrong to assume that validation in multiple assays is a necessary criterion. The assays are testing different aspects of neurodegeneration, and until we know exactly how huntingtin causes disease, we will not know which assays are testing critical steps in the pathway. “If we set as a criterion that a compound must be effective in three different assays, we may be eliminating the ones that are effective in the disease,” he said.

One of the factors that will be important to optimize with any therapeutic agent is delivery not only into the brain but into the nucleus if studies support the need for nuclear action. The nuclear question highlights the importance of understanding the biology of the disease, said Anne Messer.  Drug delivery into the brain may be accomplished either by chemical modification or by mechanical means such as intrathecal injection. Messer noted that other investigators have created encapsulated factories that pump out proteins into the brain.

Organizationally, there are other additional hurdles ahead. Kurt Fischbeck noted that major pharmaceutical companies have extensive experience in drug development, and wondered whether this group of primarily academic researchers would be able to pull off a similar feat. Brian Pollok said that while pharmaceutical companies estimate it costs about $800 million to bring a drug to market, that estimate is not applicable to this situation. His thumbnail estimate: $30 to $40 million. Heemskerk added that NINDS has already mounted a successful drug development program in anti-convulsants.

At the end of this session, Bob Baughman, associate director of technology development at NINDS announced a new translational research program with multiple funding streams. Its goal is to fill the zone between basic research and mechanistic studies and clinical trials. This is where most potential treatments fail or never get off the ground, he said. Standards will be very high, said Baughman. Investigators must have something that looks promising to get into the program. Furthermore, projects will be designed using a milestone-oriented model rather than a set of specific aims. Milestones will be well defined ahead of time; and support will be terminated if they are not achieved. NINDS will establish cooperative agreements with investigators and stay involved through the research process. There will also be an exploratory/development program to help researchers get started and mentored career development awards for PhDs. Possible projects include high throughput screens using a well-defined assay, secondary screens, new uses for existing drugs, gene therapy, stem cell therapy, and development of devices for drug delivery. The awards will be available in all disease areas and to non-profit, academic, and for-profit companies.

Cellular pathogenesis 

            The second session focused on the mechanisms of cellular pathogenesis .Are aggregates the primary pathogenic feature or merely contributory? If aggregates are pathogenic, are nuclear entry and nuclear inclusion formation required? What proteolytic events result in fragments that can enter the nucleus, and is proteolysis a good target for intervention?  Might conformational changes in the protein be pathogenic regardless of aggregation? Most of the presentations included discussion of the therapeutic implications of the research.

            Structure. Andrew Herr analyzed the structure of soluble huntingtin exon 1 fusion proteins with both non-expanded and expanded polyglutamine tracts. His data suggests that there is no global conformational change above 36 glutamines (the toxicity threshold) and supports a “linear lattice” model for polyglutamine. This new structural paradigm, he said, has implications for both disease origin and drug design. 

Steven Finkbeiner previously developed a monoclonal antibody (3B5H10) that recognizes the disease-associated conformation of polyglutamine. Using a robotic microscope designed to monitor htt expression in neurons, estimate the amount of the disease-associated conformation, track inclusion body formation, and observe neurodegeneration, he concluded that the amount of htt conformation recognized by the 3B5H10 antibody is a better predictor of neurodegeneration than is the formation of inclusion bodies.

In the late-breaking news session on Sunday afternoon, Michael Sauder presented data in which he analyzed the sequence of htt in terms of probable structure, and compared this with sequence data available in other species. He zeroed in on repeating motifs found in htt and other proteins called HEAT repeats (a protein motif found in huntingdon ,elongation factor 3 (EP3) regulatory A subunit of protein phosphatase 2A,and TORI).  In htt, there are at least 4 separate HEAT-containing domains containing as many as 66 HEAT repeats. One possibility his lab is investigating is whether polyglutamine stretches may interrupt HEAT repeats.

Aggregation. Michael Sherman’s data, collected in yeast mutants, suggest that deficiencies in chaperones and deletion of the RNQ1 prion gene both suppressed aggregation and toxicity.  Enhanced toxicity appeared in mutants with suppressed endocytosis. He concluded that inhibition of endocytosis may be the primary cause of polyQ toxicity. Andreas Wyttenbach, working in a cell model of HD, found that the heat shock protein HSP27 reduces polyQ toxicity but does not reduce aggregates.

            David Rubinsztein studied the role of autophagy in the degradation of aggregates. He found that rapamycin, which stimulates autophagy and enhances the clearance of aggregate-prone proteins, reduced the appearance of aggregates and cell death associated with polyglutamine expansions, suggesting that this drug may have therapeutic potential. 

            Nuclear localization and proteolysis. Caroline Benn’s data suggested that nuclear localization contributes to pathogenesis. Cynthia McMurray analyzed single primary striatal neurons through the cell death process and concluded that the initial phase of toxicity occurs in the cytoplasm with nuclear entry following.  Nuclear accumulation of protein is not required but can accelerate cell death.  “We believe that once the mutation occurs, proteolysis is inhibited. It happens more slowly in the cell. So the real issue is that the mutation alters turnover. This allows accumulation and eventual aggregation of the protein, and inhibits proteolysis,” she said.  N-terminal fragments may be generated late in the process. Therapeutically, it may be better to look at some of these earlier events.

            Ray Truant’s lab discovered a functional nuclear export signal (NES) in the C-terminus of htt; yet found that this NES is absent in toxic N-terminal fragments. Using time-lapse live-cell video microscopy in a clonal striatal cell line, Truant showed that htt shuttles between the nucleus and cytoplasm of striatal cells. His research suggests that this shuttling may be facilitated by the association of htt with a protein called HYP-C, which contains two nuclear localization signals (NLS). 

            Chris Ross’s work also supports the idea that the toxic fragment of htt lacks a functioning NES but preserves NLS function; and that toxicity is enhanced by nuclear localization. To study this further, his group has developed an inducible transgenic model of HD expressing full-length htt with either 23 or 148 glutamines and an N-terminal myc tag. The 148Q mouse shows a robust behavioral phenotype, early death, and characteristic nuclear pathology including intranuclear inclusions but no substantial neurodegeneration. These data indicate that proteolytic cleavage and nuclear localization are key pathogenic events and suggest proteolytic cleavage as a therapeutic target.

             Sandrine Humbert’s work suggested that Insulin Growth Factor (IGF-1) inhibits both nuclear inclusions and neuronal death mediated by htt with an expanded polyglutamine stretch. In studying the mechanism of this neuroprotection, Humbert identified a particular site on htt that is phosphorylated by the serine threonine kinase Akt/PB, which is activated by IGF-1. He further showed that Akt is altered in HD patients.

Fitting the pieces together

            In the panel discussion following the datablitz, David Housman asked, “Can all of these pieces can actually fit together? And if they contradict each other, how do we make sense of it all?”

Vernon Ingram asked why only striatal neurons are affected when htt is universally expressed. Ron Wetzel pointed out that the data are viewed through different lenses depending on the person viewing them. From a cell biologist’s perspective, unique properties of cells and the relationship of striatal cells with the cortex may help explain why striatal neurons are more vulnerable. Biochemists look at aggregation, the level of expression of htt, subcellular levels of proteases, chaperones and transcription factors. Meanwhile, electrophysiologists look to physiological changes in the system rather than in single cells.

Peter Lansbury expressed concern about the focus on aggregation, and suggested that investigators look upstream at events that precede aggregation, such as regulation and expression of htt, post-translational modifications of the protein, and proteolysis. “Proteolysis can lead us down many paths,” said Marian DiFiglia. She urged investigators to think more broadly about pathways in the cell, nucleus and cytoplasm.

Elena Cattaneo added that attention should be paid to evidence that wild type htt is protecting the cell. She suggested that normal htt may be inhibiting the transcription of the BDNF (brain-derived neurotrophic factor) gene. Therapeutically, that might mean that one would have to block the mutant htt and restore normal htt activitiy. Marian DiFiglia added that many people are looking at normal htt and that many interacting proteins have been identified, including proteins with trophic properties and those involved in membrane trafficking and endocytosis.   

Jim Olson asked about the need to investigate the role of full-length vs. truncated protein. Peggy Shelbourne reported that mice that express the truncated protein progress at the same rate as those that express full-length protein. However, the distribution of aggregates differs. Mice with the full-length protein have aggregates mainly in the striatum, with a few in the cortex and cerebellum. In mice with truncated protein, aggregates are distributed widely throughout the brain.  Olson commented that these in vivo data appear difficult to reconcile with in vitro data.

Mouse screening -- recommendations

            Day two opened with a session on biochemical and behavioral assessment in mouse models of HD. Steve Dunnett, Jenny Morton, Mike Levine, and Gillian Bates reported on a recent HDF workshop held in Cardiff, Wales[ii], where a series of recommendations for development of screening trials were formulated. The purpose of these recommendations is to standardize data collection so that comparisons can be made among different investigators, mouse lines, and compounds.

            Mice. The group recommended that each screening trial use two mouse lines – one with a rapidly developing phenotype, such as R6/2 or N171; and a second line with full length genomic htt, such as a knock-in or YAC transgenic. With any line used, the genotype should always be known. New mouse mutants should be fully characterized by backcrossing onto a number of inbred strains and then defining the phenotype in each line. Phenotypic characteristics attributable to the mutation would be those seen in all lines. When people publish work they should also report the repeat size.

            Housing.  Basic housing parameters should be consistent across labs, including access to food and water, single vs. group housing, and environmental enrichment. In addition, behavior can be strongly affected by circadian rhythm, so testing should be done consistently during the circadian dark phase when mice are most active.

            Pathology. Key aspects of neuropathology and the age of the mouse should be reported when doing behavioral testing. Histology tests should include: Nissl stain, Ubiquitin stain (DAKO antibody), htt antibody (EM48 or MW7), and possibly GFAP. Coronal sections done at the level of the mid-striatum would also be useful.

            Motor assessments.  Preferences are converging on the rotarod using an accelerating rather than constant protocol.

            General neurology.  SHIRPA is widely used for neurologic assessments; however there are problems with reliability, power, and sensitivity. More automated tests would be useful.

            Cognitive assessments.  Steve Dunnett noted that there is insufficient experience to make solid recommendations for cognitive tests. Assessments of memory/learning vs. prefrontal/executive functions are needed. Maze tasks and operant chambers have been used but require more critical evaluation. 

            Design issues (experimental).  Mouse environment should be rigorously standardized as noted above. Gender also needs to be considered. If both genders are tested, it may be difficult to randomize between litters. In addition, single housing is expensive and stresses the mice. For sufficient power, group size will need to be calculated for each assay, such that it will give an 80% chance of detecting a 25% improvement. Timing of drug administration also should be standardized.

            Design issues (pharmacokinetic). These parameters should be investigated: blood-brain barrier permeability, bio-availability, solubility, stability, half-life, clearance, effective dose, and LD50. There was a suggestion that a contract research organization (CRO) be enlisted to conduct pharmacokinetic studies.

            Other issues were raised in the discussion following this presentation.

• Statistical consultation to help analyze these data would be useful
• A position paper published in a significant journal could help disseminate and gain compliance with these recommendations
• Automation should be encouraged

Jill Heemskerk said that NINDS will be putting out an announcement of funding available for drug screening. She questioned whether data collected so far in different labs should be used for prioritization of compounds, given all the potential sources of variability.  Jenny Morton said that the group discussed this and decided that variability among labs is useful and would be helped if a positive control were available. Mike Levine added that variation among mouse models can be advantageous. A drug with the same effect in multiple models may suggest a higher probability of success in a future clinical trial.

            Susan Browne said her lab has presented data on two different drugs, one that extended survival and ameliorated weight loss and another than extended survival but had no effect on weight loss. The important factor, she said, is that her lab adheres to strict standards and rules of randomization. “We’re looking for robust effects. If effects are going to be masked by having two mice from one litter in one group and three mice in another group, that’s not very robust.”

            Dan Goldowitz added, “Standardization is great, but the literature tells us different.” He referred to a recent study[iii] where a standardized set of experiments were done exactly the same in three different parts of the country, yielding different results. He suggested taking all the results from different labs and then doing multivariant analysis to make the differences worthwhile and not disconcerting.

Biochemical and hehavioral markers in mice and humans

            Following up on the overriding theme of developing treatments for HD, Anne Young opened the next datablitz by reminding the audience of the importance of finding biomarkers of the disease in humans and animals: In humans, to show a modest improvement (20-40%) in symptomatology, 800 people would need to be enrolled. Presentations in this datablitz included both CNS and extraneural markers that may be more amenable to monitoring clinically. Potential biomarkers discussed in the subsequent presentations included:

• Retinal degeneration and dysfunction in R6 mice. May be useful for testing therapeutic strategies in living animals. (Dominique Hemlinger)

• Transcriptional dysregulation in muscle of R6/2 mice. Useful for assessing transcriptional changes in an easily accessible tissue with well characterized fundamental biology. (Andrew Strand)

• Cardiac dysfunction in R6/2 mice. Non-invasive echocardiography detected cardiac defects in HD mice at 12 weeks.  (Michael Mihm)

• Cortical degeneration assessed using novel imaging tools that measure cortical thickness in human HD patients. Detected progressive cortical thinning beginning presymptomatically. Sensitive and progressive surrogate marker may be useful in therapeutic trials of neuroprotective agents. (Diana Rosas)

• In situ polyglutamine accumulation in humans. Detected aggregation foci and polyglutamine recruitment in presymptomatic HD brain. May be useful for testing efficacy of drugs in modulating aggregation.  (Alex Osmond)

• Patch clamp on whole cells for electrophysiological markers in R6/2 mice. Saw significant decline in glutaminergic transmission in cortico-striatal pathway as well as “bizarre large events” indicative of epileptiform activity. May be useful to look at modifications by therapeutic agents. (Carlos Cepeda)

• In situ hybridization for mRNA expression in R6/2 mice. Showed down-regulation of proteins that modulate neurotransmission, particularly complexin II.   (Whitney Freeman)

Behavioral and neuropathological phenotypes of new models were also discussed:

• Knock-in mice with full-length htt. (Marie-Francoise Chesselet)

• YAC 128 mice with full-length htt. (Elizabeth Slow)

• Acute model of HD using a lentiviral vector to deliver the transgene to multiple species, including non-human primates. (Nicole Deglon)

Drug trials: outcome measures in mouse and human

In the panel discussion following the biomarkers datablitz, Ira Shoulson reminded the audience of the distinction between biomarkers and surrogate endpoints. Surrogate endpoints take the place of a primary indicator of disease. For example, blood pressure is a surrogate endpoint for heart disease. Treatments developed to lower blood pressure actually are therapeutic for heart disease. “We’re nowhere near this [for Huntington’s disease,]” he said. Biomarkers, however, while they may not be primary indicators of disease, give internal validity to clinical endpoints and are important for drug screening.

Extraneural markers. Extraneural markers such as retinal degeneration, cardiac dysfunction and transcriptional changes in skeletal muscle could be particularly intriguing. Clinical studies in humans should be done to see if similar changes are seen. Several clinicians noted, however, that they rarely see cardiac failure in humans with HD. Blair Leavitt added that in the CARE-HD study[iv], ECGs were routinely done and no consistent defects were reported. Nonetheless, Jenny Morton noted that cardiac dysfunction in mice could be a good marker for measuring drug effects.

Proteins that are exported into the blood for analysis would also be valuable markers of disease. Array analysis might help identify some of these proteins.  Darren Monckton suggested collecting a specific subset of tissues from HD patients for analysis as more biomarkers are identified. Jim Olson said that rectal ganglia can be easily obtained and may prove useful.

CNS markers. While Gill Bates said she was encouraged by the focus on the action of the mutation outside of the CNS, she also drew attention to the new techniques Alex Osmond presented to detect aggregation foci in the human brain. We need to use the most sensitive methods to try and determine when we can first detect aggregated forms of the protein in models and in patients’ brains, she said.  “What we really want to know is, is it a therapeutic target?” 

Al LaSpada agreed that the goal is to detect the earliest events in pathogenesis. He suggested using Steven Finkbeiner’s antibody in the YAC transgenic mice to determine if there is an abnormal form of htt in the tissues destined to degenerate. That might help determine whether proteolytic cleavage is creating an abnormal form of the protein or if the abnormal protein is predisposed to proteolytic cleavage, he said. And that might clarify whether the goal should be to target the ability of the peptide to aggregate or to inhibit a proteolytic cleavage event. 

            Jenny Morton also suggested further exploration of the changes in electrophysiological activity in animal models. Currently, there is no way to measure these changes in mice, but if the technology could be developed, electrophysiological changes might make good markers.

            Combined therapy.  Several people commented on the need to test multiple treatments used simultaneously. Ira Shoulson noted that this approach originated in cancer therapy and has since become a mainstay of AIDS therapy. More recently, it has been used clinically in other areas including neurology. Michael Hayden commented that very robust endpoints will be needed if multiple interventions are to be tested.

Transcriptional dysregulation

Dysregulation of gene transcription, thought to be an early event of HD pathogenesis, was the topic discussed in the fourth datablitz. Many of the papers presented arose from the work of the Hereditary Disease Array Group, which was formed under the leadership of Jim Olson with support from the HDF[v].

Transcriptional changes

            In the discussion following the transcriptional dysregulation datablitz, Kurt Fischbeck questioned whether transcriptional dysregulation is an early step in HD pathogenesis. While there is evidence to support this idea, Fischbeck said the verdict is still out. Larry Marsh added that it is still not clear whether transcriptional dysregulation is a primary causative step in pathogenesis or if downstream effects are more important.

            If transcriptional dysregulation is an early critical event in pathogenesis, asked Fischbeck, which factors are involved? Some of the transcriptional factors that have been implicated in polyglutamine toxicity, such as CBP (CREB binding protein), Sp1, TFIID, BDNF, and HAP may be therapeutic targets. No one particular target stood out, which may be the result of different models having different perturbations, said Jang-Ho Cha. “We need to start thinking about which are the real transcription factors that are critical,” he said. “It may not be a general transcription factor expressed throughout the nervous system.”

Moreover, results from Elena Cattaneo and other groups suggest that wild type htt may be a transcription factor. Transcriptional dysregulation may be due to loss of expression of wild type htt rather than a primary effect of mutant htt, said Cattaneo.  Kimberly Kegel, however, said that her work suggests that wild type htt suppresses transcription.  In the late breaking news session on Sunday, Edoardo Marcora presented data supporting a model in which wild type htt functions as a scaffold for neuronal transcription factors such as NeuroD.  NeuroD plays an important role in neuronal differentiation and survival as well as regulation of BDNF gene expression, said Marcora. 

Jim Olson suggested that gene expression changes might arise from the response of the animal to disease rather than to primary transcriptional effects. Is it possible that some drugs, like histone deacetylase (HDAc) inhibitors, are acting via genes that are permissive for neuroprotection rather than by transcriptional dysregulation?  Adaptive changes could also explain why there is so much variability among different mice. Olson also raised the issue of homeostasis. If transcription factors are being sequestered into aggregates, the response of the organism would be to produce more of these factors, presumably leading to gigantic aggregates.

Victoria Richon, who was involved in the early work on development of HDAC inhibitors for cancer treatment, noted that histone deacetylation is a misnomer, because many proteins other than histones are also deacetylated. In addition, histones are modified by not just by acetylation, but by methylation and phosphorylation as well.

Pathogenesis

            Sunday began with a session aimed at finding answers to some of the vexing questions about HD pathogenesis: Is cell death the result of murder or suicide? Why is degeneration limited to certain regions of the brain? Could regional specificity be explained by somatic expansion of the polyglutamine repeat stretch? And, are there genetic modifiers that might explain the variations among people with similarly sized mutations?

Organismal Pathogenesis

 * Late breaking news presented Sunday afternoon

            In the discussion following these presentations Darren Monckton argued that somatic expansion is the primary pathogenic event in HD pathogenesis. “Pure simple logic,” he said, suggests that increasing the repeat length influences severity. And that, he continued, means that polyglutamine repeat is the only validated target for therapy although one of many potential targets. He added that drug trials should look at effects on DNA instability.        

            However, Robert Horvitz said that while he found the somatic expansion hypothesis “incredibly intriguing, and potentially a very important observation,” it could also be an epiphenomenon; or the expansion could be downstream of and not related to pathology. In between these two extreme views, he continued, is the idea that there is some other initiating event, followed by somatic expansion in certain cells causing an exacerbation of the phenotype. 

            Blair Leavitt reminded the group that the htt mutation expands both intergenerationally and somatically, and that regions that are more selectively vulnerable seem to have the most instability.  Despite an incomplete understanding of the pathogenic mechanisms, most evidence suggests that at a certain size, the protein becomes toxic. This may be an epiphenomenon, for example a response to oxidative stress, DNA damage, or DNA repair, he said.

            Marie Francoise Chesselet added that the somatic expansion hypothesis could explain the observation in post mortem human brains. “The process is so stochastic,” she said. “You see degenerated neurons right next to healthy neurons. This is so striking.” Moreover, she said, you can’t dissociate somatic expansion from the idea of regional specificity.  Stochastic damage goes against the idea of murder within the striatum, but murder coming from another region would make sense.

            Downstream effects and genetic modifiers. Allan Tobin suggested that downstream effects may be what bothers people more than the primary event. Marcy MacDonald said that evidence from human studies suggests that different factors affect disease initiation and progression. Downstream events relate more to disease progression, and these are not grossly affected by CAG repeat.

            Blair Leavitt drew attention to other polyglutamine diseases. “I’m always struck by the dichotomy between how incredibly similar some of the aspects of polyglutamine disorders are and then the relative differences.”  In the juvenile form of all of these disorders, regional specificity disappears and the phenotypes blur. That may result from pure polyglutamine toxicity. But in adults, each disorder has a distinctive pattern, which is likely to depend on protein context.

            In the late breaking session on Sunday, Russ Margolis introduced another complication. He has identified a large pedigree in which affected persons present very much like juvenile HD but have a mutation with a CAG/CTG expansion on 16q14.3. Pathology is indistinguishable from HD except that there are no intranuclear inclusions. More work will be needed to sort out the mechanisms of pathogenesis for this disease and relate these mechanisms to HD.  

New therapeutic approaches

            “This has been an extraordinary meeting,” said Ethan Signer, introducing the session on new therapeutic approaches. “Even five, four, three years ago, one could not have imagined 250 academic scientists talking about drug discovery. This is something new, and I think it’s very good.”

            Therapeutic approaches discussed included trials of single agents in mouse models; a trial testing the combined effects of two compounds; two efforts to interfere with huntingtin at the RNA level; and one cell replacement approach.

            Drug treatment. Ascorbate, the deprotonated form of vitamin C, showed promise as a therapeutic agent in work presented by George Rebec. His group demonstrated that ascorbate levels are reduced in the striatum of the R6/2 mouse. Moreover, regular injections of ascorbate reversed the motor signs of HD such as repetitive grooming and stereotypical motor patterns without affecting overall motor activity.   

            Michael Christie tested the HDAC inhibitor phenylbutyrate (PB) in R6/2 mice. PB is currently available for clinical use, has known bioavailability in the brain, and has been shown to be safe and well-tolerated in clinical studies for other conditions. Christie reported that PB increased survival, ameliorated weigh loss, and improved rotarod performance in transgenic mice.

             Combinations of drugs have become the norm in treating cancer and AIDS, among other conditions, said Robert Ferrante. He tested the effectiveness of combining two agents that target different mechanisms of cell death: mithramycin A, which inhibits gene expression, and cystamine, a transglutaminase inhibitor that reduces aggregate formation. Mithramycin alone extended survival of R6/2 mice b 29.1%. Cystamine given along extended survival by 19.5% and reduced aggregates in the striatum and cortex. Given together, these two agents extended survival by 40.1%.

            Targeting the HD gene. Eric Kmiec approached disruption of the huntingtin gene by using synthetic oligonucleotides. In PC12 cells, Kmiec showed that synthetic oligonucleotides targeted the triplet repeat sequence of the huntingtin gene and altered it through nucleotide exchange; and disrupted the formation of aggregates.

             Wanzhao Liu used double stranded RNAs to interfere with gene expression. His team developed three small interfering RNAs (siRNA) that corresponded to segments of huntingtin exon 1 upstream of CAG repeats. In tissue culture, all three siRNAs suppressed expression of the huntingtin gene with varying efficiency. More work will be needed to see if siRNA has any effect on the normal allele.

            Edgardo Rodriguez used a gene transfer approach to try to reduce the expression of mutant huntingtin in the striatum of R6/1 mice. Recombinant adeno-associated viral vectors (rAAV) were used to deliver htt-specific ribozymes unilaterally to the striatum of adult mice.  Three to five weeks later, they demonstrated that mutant htt mRNA was reduced by 60% on the side treated with the ribozymes.

            Cell replacement. Richard Faull raised the possibility that neural stem cells may be coaxed into replacing degenerated cells in the brains of people with HD. Using an antibody directed at a marker of proliferation in cells, Faull demonstrated that cells positive for this marker as well as for an early neuronal marker are found in the subependymal layer in the brains of HD patients, suggesting that neurogenesis occurs in this part of the brain.

            In the discussion following this datablitz, Steve Hersch noted that this meeting has had the greatest proportion of presentations on therapeutics than any other HD meeting he had attended.  As the field moves closer to testing these approaches in clinical trials, he encouraged investigators to move quickly. While investigations continue to study mechanism, toxicity, etc., he said, “We have patients who don’t have anything else that are waiting for us. Things don’t need to be perfect. Clinical research is like basic research in that you make calculated risks. We need to make calculated risks with things that look reasonable and will hurt the least in the face of a disease that’s progressing and getting worse.”

            Steve Goldman said he was struck by how little attention has been given to replacement therapies including stem cell therapies. He suggested that for HD, endogenous induction may be more reasonable than transplantation. Anne Messer brought up other therapeutic approaches that had not been discussed in this session: those using small molecules combined with a targeted protein or antibody therapies. The questions these approaches raise relate to delivery. Ron Mandel said that research into vectors for gene therapy are proceeding logarithmically and vectors are now available that can transduce large gene segments. The lentivirus vector discussed earlier by Nicole Deglon, for example, works like a fancy pump to get larger molecules into targeted areas of the brain.

            Mandel added that he would like to see more research in the area of genetic modifiers, which conceivably could be delivered with a vector. Messer suggested that some production facility could take Deglon’s lentivirus and scale it up. Ethan Signer said the HighQ foundation would be interested in supporting this work.

            Ethan Signer encouraged the group to think outside the box, propose what might be considered “wild ideas,” and think about what needs to be done in the future. David Rubinsztein said he does not think that testing 100 hits in mice will be that difficult, but translating into human trials presents a far bigger problem. Biomarkers are sorely needed for human studies, said Steve Hersch. In addition, the clinical infrastructure for conducting several trials in parallel is needed.

Conclusion

            In wrapping up the meeting, Allan Tobin likened the three phases of moving from bench to bedside -- discovery, development, and delivery -- to a “fragile bridge.”  “I think we’ve got the discovery phase pretty well under control and things could not be better,” he said. Focusing on the development phase is something new for the HDF, but has received a tremendous boost from the NINDS.  NIH is now looking at the NINDS as an exemplar of how the drug pipeline can be pushed along. The milestone driven approach announced by Bob Baughman represents a “sea change” for these agencies. Tobin highlighted work in multiple areas, including pharmacology, behavioral testing, biomarkers, gene arrays, and imaging. These all give us handles for following the disease in humans and mouse models.

            “We’re at a very ripe time,” said Tobin.  “We’re able to do these experiments in a scientifically engaging way and in a way that will allow us to traverse the fragile bridge from bench to bedside.”