Huntington’s Disease: From Molecule to Miracle: Searching for the Cure

Albert Parvin Foundation Workshop at the Home of Jennifer Jones Simon
January 11-12, 2003
Los Angeles, California

Prepared by Marina Chicurel, Ph.D.

Abstract

            One of the most striking features of Huntington’s disease is its multiple faces, its varied effects on behavior, anatomy, biochemistry, and electrophysiology. Although the disease has its roots in a single, well-defined dominant mutation –a CAG expansion in the huntingtin gene-- its consequences have proved complex and multi-faceted.

            Participants at the Albert Parvin workshop suggested that HD is a multi-hit disease, in which an accumulation of cellular and molecular dysfunctions ultimately leads to the lack of motor control, cognitive impairment, and death. They emphasized that elucidating the timing of these dysfunctions, particularly identifying the earliest alterations, will be critical, not only for understanding the disease, but for developing therapies. They also highlighted the importance of resolving the question of whether the primary source of pathology lies in the direct effects of mutant huntingtin on striatal cells or on huntingtin’s effects on other cells which interact with them. In this context, participants discussed huntingtin’s disruptive effects on cellular transport, neurotransmitter receptor signaling, and the functionality of cortico-striatal connections.

            Several methods for screening candidate drugs were also discussed. The use of biolistics applied to rodent brain slices, organotypic mouse slice cultures, and Drosophila aggregation assays, for example, were described. An increased emphasis on assays that closely reflect the disease process in humans was proposed, as well as the establishment of criteria to select therapeutic candidates for clinical trials. In addition, participants reviewed the status of therapeutic candidates, including regulators of chaperone protein function, minocycline, and co-enzyme Q10. Of particular interest was the discussion of histone deacetylase inhibitors, which have recently provided encouraging results in a mouse model of disease.

The far-reaching consequences of HD

            The generous participation of three women from a family suffering from HD greatly helped workshop attendees grasp the far-reaching and complex consequences of HD. By sharing their experiences coping with the myriad ways in which their lives and those of the people they love have been disrupted, the women highlighted the multi-faceted nature of the damage caused by HD. 

             Describing the primary victims of the disease in their family, the women recounted their experiences witnessing the loss of motor and cognitive abilities in their loved ones and the heartbreaking end-stages of the disease, in which the victims have difficulty even eating. Inextricably linked to this suffering, was the pain of  associated  family members. One woman, whose mother is bed-ridden because of HD, for example, noted that her father is completely dedicated to caring for his wife, a once intelligent, talented and independent woman. Indeed, he has been unable to get a knee replacement, which he seriously needs, because of the all-absorbing nature of his role as caregiver.

            The women also noted the extremely difficult considerations they have had to face in deciding whether to be tested for the HD mutation. One of the women said she did not want to know her personal fate, but had wanted to make sure that her children wouldn’t have the disease. After aborting two fetuses because they tested positive for markers associated with her diseased father’s chromosome, however, she decided she would go through her subsequent pregnancies without testing. Now she lives with the fear that she and her two small children may be carriers of the mutation. As noted by Nancy Wexler, a major difficulty for families afflicted with HD is balancing their desire to protect their children with their aversion to learning their own fate.

One woman described how she had always been fearful of suffering her mother’s fate and had even been preocuppied with finding traits in herself that resembled her mother’s, even if they were unrelated to the disease. Yet she had avoided thinking about the possibility of getting tested. It wasn’t until a counselor suggested the idea to her, while she was visiting her mother in the hospital, that she decided to get tested. She tested positive and has begun experiencing the early symptoms of the disease. In addition to the fears she harbors for her own health, she is now intensely worried about her husband, who will probably have to become her caregiver, as well as her young children, who may suffer from her failing health and may be carriers of the mutation.

The family has also had to face tough economic issues, such as keeping their jobs and trying to secure suitable medical and life insurance. The woman who recently tested positive, for example, has not yet notified her employer for fear of losing her job. And the whole family is uncertain about how to obtain the insurance they need. As noted by Alice Wexler, HD families confront a highly complex web of social, psychological, and financial challenges.

The women’s personal and highly articulate presentation of their family’s experiences made a strong impression on workshop participants. Shaken by the terrible and wide-ranging effects that HD has wreaked on this family, participants were inspired to help develop strategies to understand the biology of HD and accelerate the search for a cure.

Understanding HD

            Just as HD has far-reaching consequences within families, mutated huntingtin has a wide spectrum of effects on brain cells and their functions. Indeed, one of the major challenges for understanding the disease is distinguishing primary from secondary and compensatory effects.

Timing

As noted by Michael Levine, understanding the progression of the disease, the timing of the multiple events and how they are causally related to each other, is critical. Consistent with this proposal, Martha Constantine-Paton suggested that HD may be a multi-hit disease in which the brain can compensate for one or a few early injuries, but as the dysfunctions accumulate, a critical threshold is reached and the system as a whole starts to malfunction. At such a point, behavioral deficits become apparent, even though the pathology has been unfolding for a long time before. Understanding the disease and designing effective therapies, noted Levine, will probably depend greatly on identifying the early events that drive disease progression.

Underscoring the importance of timing, Levine described the changing behavior of striatal cells in R6/2 brain slices. At approximately 5 weeks of age, he reported, the striatum drastically reduces its responses to cortical stimulation, as if the two structures had been disconnected. Immunohistochemical studies revealed a corresponding decrease in PSD-95, synaptophysin, and other synaptic proteins at cortico-striatal synapses. Potentially contributing to or triggering this disconnection, Levine observed inward currents five to six times larger than normal driven by cortical inputs shortly before the reduced responses.

Thus, the behavior of striatal cells changes dramatically during disease progression and may help explain apparently contradictory results.  For example, a study by Brundin and colleagues showed that R6/1 mice are strongly protected from acute striatal excitotoxic lesions induced by quinolinic acid –a finding that seemed to contradict early observations from Levine’s group showing NMDA hypersensitivity. As noted by Levine, however, the difference in timing between the two studies probably explains the paradox. Indeed, subsequent studies have shown that quinolinate-induced excitotoxicity is in fact enhanced in HD mice at early times.

To chart the development of striatal dysfunction and identify early changes that might initiate the disruption, Levine is now studying 15-21 day-old mice. Even at these very early ages, he finds abnormally high calcium conductances, as well as NMDA hyperexcitability. Furthermore, in dissociated cells, he has observed a decrease in the NMDA receptor Mg⁺⁺ block at various time points. He is now setting up experiments to assess the contributions of different glutamate receptors, including different NMDA receptor subtypes, to the observed responses.

            The disconnection described by Levine led to a discussion regarding its potential mechanism. Ethan Signer pointed out that the synaptic severing must be reversible because inducible models of HD, such as the Yamamoto mouse, have shown that HD pathology abates when the mutant gene is switched off. Constantine-Paton noted that if the disconnection was functional, rather than anatomical, it could be easily reversed. In addition, Marie-Francoise Chesselet pointed out that dendritic spines are dynamic entities that could reform easily, and Charles Wilson added that the retraction of dendrites by motor neurons as a result of axotomy was reversible, at least to a certain degree.

            An important consideration, noted by Constantine-Paton, is that the early effects of mutant huntingtin are probably occurring in the context of normal developmental processes. Between 14 and 20 days after birth, for example, there is a radical shift in the scaffolding that holds NMDA receptors at the synapse. Of particular interest, is the observation by Constantine-Paton’s group that the synaptic recruitment of scaffolding protein PSD-95, which interacts with wildtype huntingtin, appears to be highly dependent on electrical activity. Thus, disruptions in electrical activity and/or in huntingtin’s interaction with PSD-95 are likely to have significant consequences on synaptic function. Scaffolding proteins are key determinants of the types of glutamate receptors and the associated signaling molecules that populate synapses. A scaffolding protein that is abundant in fetal synapses, SAP-102, for example, preferentially recruits NR2B NMDA receptors, whereas PSD-95 recruits NR2A NMDA receptors.

To investigate these processes more thoroughly and understand how they may be altered in HD it will be important to study proteins at the synapse, rather than merely examine levels of expression in whole cells, noted Constantine-Paton. Indeed, to advance their developmental studies of the retina, her group is beginning to conduct immunoprecipitations using synaptoneurosomes, subcellular fractions enriched in synaptic components.

The importance of developmental context was also underscored by observations of the formation of aggregates and neurodegeneration in Drosophila expressing mutated huntingtin exon 1. Leslie Thompson, who is using Drosophila as a model system for screening therapeutic candidates, noted that the rhabdomeres of the fly eye that differentiate first develop aggregates and die earlier than those that differentiate later. Transcription factors associated with differentiation might be directly or indirectly responsible for the vulnerability. She also noted that whereas undifferentiated neuroblasts develop huge cytoplasmic inclusions, cells that are terminally differentiated (in a brain region called the mushroom bodies) tend to develop nuclear inclusions. Constantine-Paton suggested looking at the pathology of neurons in mammals that differentiate late, such as those in the olfactory bulb. However, Chesselet pointed out that the olfactory bulb is usually loaded with aggregates that appear early on in the disease process. Nevertheless, it might be worthwhile to specifically examine immature, dividing cells, using BrdU labelling, for example.

An important clue regarding the early events in HD pathology was presented by Peter Reinhart. Using a biolistics approach to deposit DNA-coated particles into neurons in living brain slices, Reinhart has transfected striatal and cortical neurons with a variety of huntingtin constructs bearing 23-148 poly-glutamine repeats (see Screening Assays section for more information). Among his various observations, Reinhart noticed that cells transfected with long poly-glutamine constructs become difficult to patch very shortly after transfection. Several participants encouraged Reinhart to follow up on this observation. Reinhart agreed, noting that understanding such early events and their progression will probably be key to understanding HD.

Striatal vulnerability

            Another major challenge in the study of HD is understanding the apparent vulnerability of striatal cells, particularly medium spiny striatal cells, in the disease process. As summarized by Chesselet, although many neuronal types are affected in HD --particularly if mutated huntingtin is over-expressed or carries a very long stretch of glutamines-- striatal cells show some of the earliest and most dramatic signs of degeneration. Indeed, nearby  cells, including striatal interneurons and striatal cell targets, such as neurons of the globus pallidus, appear to be mostly spared, while the axons and nuclei of spiny cells are loaded with aggregates.

            As noted by James Surmeier, one of the distinguishing characteristics of spiny cells is their expression of dopamine D2 receptors. Strongly linked to Go and working through phospholipase C, D2 receptors release calcium from intracellular stores. And as noted by Levine, they influence glutamate stimulation by decreasing its release, as well as decreasing NMDA sensitivity. Interestingly, immunohistochemical studies have revealed that D2 receptors are decreased in HD.

            Based on these considerations, Chesselet, Surmeier, and Levine stressed the importance of examining dopamine transmission in HD. Levine noted that William Yang has generated transgenic animals that express D2 or D1 receptors fused to GFP. He suggested crossing these mice with mouse models of HD. Levine also noted that his lab is currently studying the effects of dopamine on cortico-striatal responses using brain slices.

            Another distinguishing characteristic of medium spiny cells briefly discussed at the workshop was the cells’ NMDA signaling pathway. Unlike most mature neurons, medium spiny cells express high levels of NMDA receptor NR2B subunits relative to other NR2 subunits. In addition, Marina Chicurel pointed out that medium spiny cells are particularly enriched in STEP, a tyrosine phosphatase that regulates the duration of ERK signaling mediated by NMDA stimulation. as recently reported by Paul Lombroso and colleagues. The duration of ERK signaling critically determines a cell’s response to stimulation: transient activation, mediated by NMDA receptors, results in specific transcriptional changes that differ from those induced by sustained activation, as mediated by KCl-induced membrane depolarization and calcium influx. Thus, the enrichment of STEP in medium spiny cells may be an important determinant of the cells’ responses to both normal and altered NMDA stimulation.

            It is important to note, however, that the significance of medium spiny cells’ apparent vulnerability is still subject to debate. Gillian Bates was concerned that too much attention has been focused on medium spiny cells. She noted that cells die in many areas of HD brains, and that HD may be better thought of as a global disease of the brain, rather than a specific disease of the striatum. In addition, although degeneration of striatal cells is clearly a hallmark of HD, some of the early evidence for specific vulnerability has proved misleading. For example, results showing decreased levels of enkephalin in the striatum were initially interpreted as evidence of the selective death of enkephalinergic neurons. Yet, as noted by Chesselet, subsequent studies have shown that the decrease actually represents a decline in enkephalin expression, not a loss of specific cells.

Murder or suicide

            Another important point of discussion was whether the damage seen in the striatum is mediated by mutated huntingtin’s direct effects on striatal cells, or alternatively, by the protein’s effects on other cells that interact with the striatum --in other words, is striatal cell death due to ‘murder’ or ‘suicide.’ Participants agreed with Ann Graybiel that this was a key question to resolve, both for understanding the disease and for developing therapies.

As noted by Carl Johnson, there are compelling data supporting the possibility of murder. Experiments using chimeric animals whose cortices predominantly express mutant huntingtin and whose striata predominantly express wildtype huntingtin, as well as animals with the opposite configuration, suggest that cortical genotype is the critical determinant of HD pathology. On the other hand, Kurt Fishbeck noted that many studies of the toxicity of huntingtin and its aggregates suggest suicide. Susan Lindquist added that indeed it is clear that mutated huntingtin is toxic, but the question is whether this toxicity is directly affecting striatal cells or if its main effects are on other cells whose malfunction causes striatal pathology. Another possibility is that both murder and suicide contribute to striatal damage. As described by Chesselet, the phenomenon might be viewed as ‘assisted suicide,’ a concept that several participants found attractive.

            Using the terms ‘murder’ and ‘suicide’, however, may be somewhat misleading. As pointed out by Wilson, the main problem in HD is probably not cell death per se, but neuronal dysfunction. As the studies of Levine, Reinhart, and others show, striatal electrophysiology in HD is abnormal. And the results of this abnormality may wreak havoc on the normal communication of the striatum with other brain areas which is required for mediating complex behaviors, including cognition. As emphasized by Wilson, more harm may be done by neurons that are sending incorrect or garbled messages, than by dead neurons that are simply silent. Key to resolving this issue will be to learn more about the language of the healthy striatum –how these neurons encode information through temporal patterns of activity. Electrophysiological studies in monkeys and imaging studies in humans are beginning to yield some clues, but our understanding of the spiny neuron as a computational machine is still in its very early stages.

One of the therapeutic implications of this situation, noted Wilson, is that it may be difficult to develop effective cell replacement strategies. In addition, until a better understanding of the function of the normal striatum emerges, it may be therapeutically more appropriate to silence dysfunctional neurons rather than trying to keep them alive and active.

Supporting the idea of improving the understanding of the basic physiology of HD’s key players, Surmeier emphasized the need for electrophysiologically characterizing normal and HD cortices, and said his group may start examining cortical slices. In addition, Chesselet noted that the striatal-pallidal pathway should not be neglected and recommended analyzing it in healthy and HD brains.

Cellular mechanisms of toxicity

            Regardless of whether mutated huntingtin mediates its toxic effects directly or indirectly, participants agreed that understanding its molecular effects within cells will be key to understanding the disease. A potentially important source of toxicity was presented by Larry Goldstein, who described experiments indicating that poly-glutamine can interfere with cellular transport. Using the UAS/GAL4 system to drive the expression of poly-glutamine constructs in Drosophila embryos, Goldstein observed a depletion of motor proteins, including kinesin, dynein, and to some extent dynactin. Although expression levels remained close to normal, the soluble pools of the proteins were dramatically decreased as assessed by Western blots. In contrast, the soluble levels of other proteins, such as actin, tubulin, and various synaptic proteins, remained mostly unchanged.

            These results support previous data suggesting that mutated huntingtin might disrupt the cell’s transport machinery. For example, Signer noted that the two-yeast hybrid system had revealed that HAP-1, a huntingtin associated protein, binds to the p150(Glued) subunit of dynactin. Interestingly, a missense mutation in p150(Glued) which interferes with its binding to microtubules, noted Fishbeck, results in a phenotype that is similar to that of another polyglutamine disorder, spinal bulbar muscular atrophy (SBMA). In SBMA, the androgen receptor, instead of huntingtin, carries an expanded stretch of glutamines. In addition, Chesselet pointed out that abnormalities in the expression of a protein involved in veiscle trafficking, complexin II, have been reported in mouse models of HD.

            To extend his studies, Goldstein is setting up experiments in mice. Chesselet recommended using knock-in HD mice with long glutamine repeats because of their abundant axonal aggregates. Lindquist added that using an inducible mouse model might be useful to assess the reversibility of the phenomenon. So far, Goldstein has been unable to obtain inducible mice, but Levine noted that he should be able to obtain the Tet-inducible Yamamoto mouse fairly easily since it is now available at UC Irvine, and may soon be available at UCLA as well.

            Participants also discussed other potentially important contributors to mutated huntingtin’s toxicity. Tobin, for example, noted that Eric Schweitzer in his lab has obtained evidence suggesting that glial cells may cope with mutated huntingtin better than neuronal cells because of their proteasomes. Schweitzer transfected an astroglia-like cell line, neuronal-like PC12 cells, and immortalized neurons with GFP-tagged huntingtin constructs regulated by an ecdysone-responsive element. He observed that whereas the neuronal cells died hours after the induction of mutated huntingtin expression, the astroglial cells accumulated cytoplasmic aggregates but remained alive. When exposed to proteasome poisons, however, the astroglial cells died too. Reinhart noted that, using his biolistics transfection system, he also noticed a robustness in glial cells’ response to mutated huntingtin. When he did co-transfections with a fluorescent tag driven by the GFAP promoter, he observed that GFAP-positive cells appeared much healthier compared to transfected neurons. One possibility, suggested by Fishbeck, is that the aggregates in astroglial cells are aggresomes, structures that, rather than being toxic, represent a protective cellular response (see next section). To extend these findings, Tobin is interested in comparing proteasome activities in glial cells and neurons. In addition, Anne Young suggested performing co-culture experiments and Signer proposed generating heterokaryons.

            The importance of investigating huntingtin’s effects on other cellular pathways was also discussed. Surmeier and Graybiel emphasized growth factor signaling, and Goldstein noted his interest in examining how HD affects JNK signaling.  David Housman asked about the potential importance of huntingtin’s interactions with lipids. Early experiments in lipid bilayers had suggested that huntingtin might form channels across membranes. Johnson noted, however, that recent experiments revealed no change in the channel activity of Xenopus oocytes injected with mutated huntingtin mRNA. Nevertheless, huntingtin may have as yet unrecognized effects on membranes through interactions with lipids. Isaiah Arkin is now examining these interactions, conducting biophysical and biochemical analyses. In addition to these direct molecular interactions, huntingtin may affect cells’ lipids through transcription. As described by Thompson, recent microarray studies have revealed changes in the expression of genes involved in lipid metabolism in mouse models of HD.

            Putting forth a particularly original proposal, Wilson noted that, depending on the electrical conductivity of huntingtin aggregates, they may act as electrical resistors, slowing down conduction along neurites. Goldstein was intrigued by this possibility and said he was interested in testing it experimentally.

A multiplicity of aggregates

A major focus of the discussion on huntingtin aggregates, however, centered around their heterogeneity. The nature and effects of huntingtin aggregates have been subject to intense debate, but the emerging consensus is that not all aggregates are the same, and that they can differ radically in their degree of toxicity. As noted by Jeffrey Kelly, unlike most proteins which form homogenous aggregates when overexpressed, huntingtin aggregates are very heterogenous –a characteristic that probably contributes to their varied behaviors. Indeed, in some cases, huntingtin aggregates may not be toxic at all. As described by Fishbeck, several disease-causing proteins can form aggregates which are closely associated with the microtubule organizing center, called aggresomes, that appear to be protective, actually reducing the proteins’ toxicity. In addition, as described by Lindquist, low concentrations of soluble species of poly-glutamine proteins can be highly toxic. Thus, huntingtin toxicity is a complex phenomenon that cannot be simply equated with generic aggregate formation.

            Among the various factors that can affect aggregate toxicity are the activities of other cellular proteins. Describing her findings in yeast, for example, Lindquist noted that the toxicity of mutated huntingtin can be dramatically modified by RNQ, a glutamine-rich prion protein. (A growing number of prions have been found in yeast which act as genetic elements that regulate different aspects of cell metabolism. Their propagation is often regulated by chaperones.) When RNQ is in its prion state, huntingtin forms aggregates that are much more toxic than those formed when RNQ is in a non-prion state. Although the two types of aggregates do not seem to differ greatly in their solubility as assessed by filter-trap assays, they are clearly distinguishable under the microscope: the more toxic ones tend to be amorphous, while the less toxic ones have sharp borders. How RNQ affects huntingtin’s toxicity is still unclear, but one possibility is that RNQ interacts directly with huntingtin to alter aggregate formation. Alternatively, RNQ may bind to chaperones or other proteins, which in turn affect aggregate formation. Citing another example in which the morphology of aggregates seems to correlate with toxicity, Kelly noted that a recent study of Alzheimer’s disease revealed that toxic aggregates were usually spherical.

            The subcellular environment in which aggregates reside seems to be another key factor determining toxicity. As noted by Fishbeck, the toxicity of androgen receptors carrying expanded poly-glutamine repeats depends on the presence of androgens which, when bound to the receptors, allow them to move into the nucleus. And as described for several polyglutamine proteins, aggregates can be particularly toxic in the nucleus because they can disrupt transcription by sequestering transcription factors. Bates added that experiments in her lab using exon 1 bearing either nuclear localization signals (NLS) or nuclear export signals (NES) suggest that NLSs speed up disease onset and cause nuclear pathology. Although still preliminary, her results also suggest that NESs result in a delayed onset of disease, consistent with studies in cell culture indicating NESs decrease cell toxicity.

Tobin suggested examining this issue more closely. Could a better understanding of how proteins are kept out of the nucleus help develop screens to identify new therapeutic candidates? Is the androgen receptor a particularly good model system to work with? Lindquist considered that if specific import proteins for poly-glutamine proteins were identified, searching for modifiers of nuclear transport might be a worthwhile avenue to pursue. However, she also noted that the mechanisms that restrict proteins to the cytoplasm, often involving large numbers of chaperone and co-chaperone proteins, have proven complicated, diverse, and difficult to study. 

Another factor to be considered is that the toxicity of huntingtin aggregates does not seem to be uniquely confined to the nucleus. For example, working with Drosophila, Goldstein has observed cell death associated with huntingtin constructs carrying either NLSs or NESs. Whereas he has observed amyloid protein aggregates moving rapidly within cells, poly-glutamine aggregates seem to be mostly stationary, potentially causing roadblocks in the cellular transport system.  In agreement with Constantine-Paton’s proposal of HD as a multi-hit disease, Goldstein noted that mutated huntingtin may disrupt cell function both in neurites and in the nucleus. Lindquist added that, in yeast, there are fewer nuclear poly-glutamine proteins potentially susceptible to sequestration by huntingtin aggregates than in higher organisms, which may explain the predominantly cytoplasmic toxicity observed in yeast.

Understanding and tracking aggregate maturation is also likely to be key for understanding aggregate toxicity. A novel technique presented by Alex Osmand to identify sites that foster aggregate growth, or foci, in patient brain slices promised to help address this issue. Working with Ron Wetzel who has studied the dynamics of poly-glutamine aggregation in vitro, Osmand developed a method to stain aggregate foci in vivo. Using short chains of biotin-labeled poly-glutamine bearing a few lysine residues to keep them soluble, he has been able to identify sites that recruit poly-glutamine in patient brains. So far, he has found that the large pyramidal neurons in layers 3 and 5 of the cortex have numerous foci which do not overlap with the typical aggregates detected by immunohistochemistry. Indeed, the high numbers of aggregates he found in several neuropil regions correlated with low numbers of foci. He also noted that patches of cells with high densities of foci are often surrounded by areas that are completely devoid of staining. There is an enormous variability in staining within and between brain regions from a single patient, as well as between matching brain regions from different patients. One of the most consistently and dramatically stained areas, however, is the primary motor cortex.

At a cellular level, most neurons sport dozens to hundreds of foci when they are positively stained –a pattern that seems to be independent of neuronal size. Supporting the possibility that these foci represent sites of early aggregate formation, Osmand has observed that PC12 cells carrying a construct coding for a chain of 65 glutamines have many foci shortly after transfection which gradually disappear over time, as aggregates increase. As suggested by Johnson, it might be interesting to study the behaviors of constructs of different lengths: one might expect that the longer the poly-glutamine chain, the faster the rate of aggregate growth, and consequently, the shorter the window in which foci are visible. Housman added that the transition from foci to mature aggregates may correlate with SDS-resistance.

Participants discussed several possibilities for additional future experiments. Osmand noted that he is now setting up control experiments using labeled poly-proline chains. He also pointed out that he is planning to examine the striatum more extensively and test the technique on fresh, rather than fixed, tissue. Tobin noted that frozen brain slices were available from the UCLA brain bank. Testing the technique on other systems was also discussed. Greg Lemke suggested examining other poly-glutamine diseases in humans, while Housman noted that the Drosophila model might be of interest because the neurons appear to undergo waves of aggregation. Others suggested looking at mouse models of HD. Osmand is particularly eager to obtain early pre-symptomatic patient brains to further explore the dynamics of aggregate formation in humans. Nancy Wexler said that fixed and frozen pre-symptomatic brains are available from the Venezuelan cohort.

Yet another approach to investigate aggregate formation in vivo was described by Thompson. Interested in examining whether aggregate formation is a dominant phenomenon, she is planning to perform cell fusion experiments between PC12 cell lines that do and don’t form aggregates. Lindquist noted that current data in yeast suggest that aggregation is indeed dominant.

As discussed in this section, many parameters affect aggregate formation and toxicity, making these phenomena difficult to study. Of particular importance, then, is to ensure that experimental parameters are as consistent and as representative of natural conditions as possible. Underscoring this point, Lindquist noted that the design of huntingtin constructs was critical –she has found that the commonly used FLAG tag, for example, can significantly affect the toxicity and morphology of aggregates.

New tools and model systems to study HD

            In addition to analyzing key unresolved questions in the study of HD, participants discussed some of the new tools that promise to expand their current capabilities. For example, Johnson noted the generation of new variants of mouse models of HD. The Jackson Laboratories have agreed to put the R6/2 and Detloff mouse models into two different congenic backgrounds. They are now selecting mice to minimize strain-specific phenotypes and plan to include the strain sequenced by Celera to facilitate the search for modifier genes. Johnson also noted that Chris Ross has developed a new inducible mouse model that carries full-length mutated huntingtin coupled to a prion promoter.

            Participants were enthusiastic about the new additions, but some were concerned about their availability, as well as the availability of older mouse models. Goldstein advocated for a centralized repository for all mouse models of HD, and Lindquist suggested setting it up through the Jackson Labs. Bates noted that the Jackson Labs have done a good job of distributing R6/2 mice and have also done well at investigating the phenotypic drift that appears to have occurred between Jackson’s batch of R6/2 mice and those kept in the Bates lab. However, Goldstein noted that, in his experience, Jackson Labs have been expensive and very slow at sending out animals. Levine added that housing mice in universities was problematic because of space constraints and because distribution could be difficult due to legal issues, especially with mice that are not in the public domain. To circumvent these problems, Goldstein and Levine suggested hiring an independent contractor. Taconic was mentioned as an option, although it is expensive.

            Johnson, however, didn’t consider there was a great need for a central repository. He noted that most labs are willing to supply their mice to others, and advised participants to call him if they experienced difficulties. He also said that if there was a consensus about a particular mouse that was difficult to obtain, the foundation could probably facilitate its distribution through the Jackson Labs.

            Participants also briefly discussed the use of new tools, such as proteomics, to study HD. Constantine-Paton noted that, in addition to the published and ongoing microarray experiments to assess changes in gene expression caused by HD, it might be useful to directly monitor protein changes since they do not always mirror changes in mRNA levels. In particular, she proposed, proteomics could be used to monitor the effects of candidate drugs on HD models. She noted that Applied Biosystems appears to be offering proteomic analysis services starting with as little as 100 mg of tissue. Others agreed with the potential value of the technique, but noted that its applications should be carefully considered. Lindquist noted that distinguishing relevant protein changes from the less important ones could be difficult. Thompson added that timing would be key, since both compensatory and primary changes would be expected to occur. And in the case of candidate drugs that affect transcription, such as histone deacetylase inhibitors, assessing expression changes would be a more direct assay, noted Tobin.

Screening assays

            A major tool in the search for treatments for HD are screening assays. Two major types were discussed at the workshop: screens to identify genetic suppressors or modifiers of HD, and screens to identify candidate therapeutic compounds. As noted by Kelly, the study of several diseases, such as amyloid diseases and cancer, have demonstrated the potential  benefits of identifying trans-suppressors, genes that when mutated, suppress some or all of the pathological effects of the disease-causing mutations. And, as noted by Tobin, suppressor screens have already benefited HD greatly, revealing, for example, the major effects chaperone proteins can have on HD toxicity.

One of the screening systems that has proved particularly useful because of its well-characterized genetics and potential for conducting high throughput screens has been yeast. Describing work in her lab, Lindquist reported they are screening a wide variety of mutants, including deletion and insertion mutants, for suppressors of HD toxicity. 

Another potential source of HD modifiers is being examined by Andrew Dillin using C. elegans. Dillin is interested in the genetics of aging, particularly in non-dividing cells. C. elegans is a particularly good model system for this research not only because of its well-characterized genetics and ease of manipulation, but because all of its cells are non-dividing. Previous genetic screens in C. elegans have revealed pathways that, when mutated, increase lifespan. For example, when certain components of the insulin signaling pathway are inactivated, C. elegans lives longer, a finding that has also been confirmed in mice. Dillin is now interested in performing genome-wide screens to find other modifiers of longevity which, he thinks, may also have effects on HD onset and/or progression.

            Using double-stranded RNA to inactivate individual genes, Dillin has blocked the activity of every gene in the C. elegans genome. So far, he has found that the largest class of genes that increase lifespan when inactivated are nuclear-encoded mitochondrial genes. Interestingly, these inactivations lengthen lifespan only if they occur during larval development. Similarly, mitochondrial inhibitors increase lifespan only when administered during the larval stage. As described by Dillin, it is possible that decreased mitochondrial function is compensated for in development resulting in a lower metabolic rate which, in turn, prolongs life. Dillin has also observed that caloric restriction has an additive effect on the inactivation of these genes. He is now monitoring reactive oxygen species to ascertain whether these metabolic by-products are being generated at a reduced rate.

An important next step will be to examine whether reducing mitochondrial function in neurons delays HD onset. To accomplish this, Dillin is planning to collaborate with researchers working with mouse models of HD. In addition, Dillin is searching for genes that reduce lifespan when mutated. He has found that 15-20% of all genes fall into this category and, of these, 60-80 do not appear to have drastic, widespread toxic effects because they don’t have a clear deleterious phenotype early in life. He suspects that some of these mutated genes are causing early neurodegeneration and, thus, may increase susceptibility to HD.

            Within the second category of screening assays --those designed to identify exogenous therapeutic compounds-- participants discussed approaches based on various model systems, ranging from cells in culture to brain slices, and from Drosophila to mice. Thompson and Tobin described using PC12 cells carrying GFP-tagged, inducible huntingtin constructs. Tobin’s approach is a medium throughput screen in which he monitors the inhibition of cell death. In the absence of exogenous compounds, the cells die approximately 24 hours after induction of the expression of mutated huntingtin.

Thompson, on the other hand, is using the cells to monitor aggregation. She has selected a line in which 60 to 80% of the cells develop aggregates. Conducting a pilot study to monitor the effects of compounds known to affect aggregation, she found that whereas cystamine and Congo red suppressed aggregate formation in a dose-dependent manner, minocycline had no effect. As pointed out by some participants, including Thompson, aggregation can be a difficult parameter to use as an indicator of drug action, however. It is hard to quantify, timing can be critical and, as discussed above, aggregates are heterogenous in their appearance and their effects.

            Therefore, assays that monitor neurodegeneration and cell death complement aggregation assays well. Indeed, the Drosophila toxicity assay developed by Thompson and Housman has lent support to Thompson’s aggregation results, yielding consistent results regarding the effects of cystamine, Congo red, and minocycline. Thompson explained that their fly system relies on a UAS/GAL4 system to drive the expression of poly-glutamine or exon 1 constructs. The expression of constructs with high poly-glutamine repeats results in early death dependent on poly-glutamine length, the formation of aggregates in the developing eye and brain, and neurodegeneration --which is also poly-glutamine length dependent and can be monitored in the eye. Thompson also noted the presence of movement disorders, such as the transgenic flies’ difficulties in climbing the walls of a tube. A particularly useful readout, however, is the degeneration that occurs in the eye because it is relatively easy to assess. Reinhart cautioned, however, that the eye may not be entirely representative of other neural tissues because it is very easily disrupted.

            Thompson is now using the fly assay to assess hits from chemical compounds obtained by other laboratories. In addition, she is continuing to investigate the suppression of HD pathology induced by blocking histone deacetylases, which she originally observed in this system.

            Mammalian screening systems were also discussed at the workshop. As previously mentioned, Reinhart has developed a slice model of HD using biolistics. Transfection and inclusion formation can be easily monitored because the huntingtin constructs, as well as the beads, carry fluorescent tags. In addition, expression levels can be readily titrated because they correlate well with the amount of DNA loaded onto the beads. Also, the preparation is relatively long-lived. Reinhart can maintain the slices with neurons firing normally for 5-6 days. The system is also simpler and easier to manipulate than other mammalian models, yet closer to the human disease than yeast or cell culture systems. Furthermore, it allows the study of the early effects of mutated huntingtin. The efficiency of transfection is less than 1% (providing 300-400 transfected cells per slice), but this ensures that individually transfected cells contain a single bead --allowing for expression titration. It also ensures that transfected cells are far enough apart from each other to allow clear visualization of their somas and neurites.

            In addition to conducting electrophysiological recordings, Reinhart is using the system to monitor cell death, inclusion formation, and neurite degeneration, which he has shown are dependent on huntingtin and poly-glutamine length in his system. The fastest output measure with the lowest background, is cell death. But the other indicators provide more specific information about the disease process. To track neurite alterations, Reinhart has developed a semi-automated system which captures fluorescent images and then performs a shell analysis, counting the number of neurite crossings in a series of  concentric circles drawn around the cell body. The analysis allows the quantification of neurite thinning and distinguishes between proximal and distal degeneration. Fully automating the system, however, has proved difficult. Kelly suggested looking into imaging software developed at Xerox Park.

Using this HD model, the now-defunct biotechnology company Cogent Neuroscience was developing medium throughput screens of a selected library of small molecules. Based on several parameters, including chemical identity or similarity to compounds known to affect HD or neurodegeneration, as well as more general considerations regarding ‘druggability,’ a group of chemists at Cogent assembled a collection of compounds to test on Reinhart’s slices, which could be processed at a rate of up to 5000 slices a day. Proving the system’s potential value, a blind screen picked up two transglutaminase inhibitors, compounds that have been previously identified as therapeutic candidates for HD.

            Now that Reinhart is working strictly as an academic, however, the direction of the project has changed somewhat. In agreement with Lindquist’s and Lemke’s suggestions, Reinhart plans to extend his studies to include slices from transgenic mice, carrying GFP tags or knocked-out genes. As noted by Constantine-Paton, the biolistics approach focuses on the suicide model of HD, while transgenics could help Reinhart examine the murder model as well. Whereas Reinhart was using both rats and mice before, he now plans to shift his focus to mice because of their genetic potential, as pointed out by Lemke. In addition, Signer suggested using the biolistics approach to do co-transfections of huntingtin constructs and suppressors of HD toxicity. Reinhart noted that he is interested in this possibility and has begun exploring it with chaperones. Yet another suggestion, put forth by Constantine-Paton, was to do biolistic transfections in the cortices of living animals.

Participants also discussed Reinhart’s future selection of compounds to screen. Reinhart stressed the importance of using ‘smart’ chemistry to narrow drug screens, not only in his own work, but in general. He noted that the company library he was using capitalized on this concept.   He also acknowledged, however, that the library was somewhat conservative, constrained by industrial financial considerations. Kelly noted that a worthwhile addition might be mechanism-based, enzyme-specific inhibitors, which have demonstrated great potential for research.

            Another mammalian screening system, being developed by Bates, involves the use of organotypic hippocampal slices from HD mice. The robustness and layered architecture of these slices, which can be kept alive for 5 weeks, make them particularly useful for testing the effects of drug candidates on aggregate formation. Bates and her team rely on immunofluorescence to stain the aggregates and use confocal microscopy to examine 2 mm optical sections of tissue. Staining is done in parallel across slices to enhance consistency, while aggregate scoring is performed atuomatically, using software that measures aggregate intensity, area, and number. As noted by Bates, the system is not well-suited for primary screens because it is labor intensive, but is proving very useful for obtaining information relevant to pre-clinical trials, such as dosage effects.

Bates has used this system to test the effects of several candidate therapeutic compounds, including benzothiazols, tetracyclines, a transglutaminase inhibitor (cystamine), and a histone deacetylase inhibitor (suberoylanilide hydroxamic acid, SAHA). She has found that neither SAHA nor cystamine affect aggregate formation. On the other hand, all tetracyclines tested (tetracycline, doxicycline, and minocycline) decreased aggregate formation when used at high doses. As discussed in more detail below, however, the doses required for inducing these effects, particularly with minocycline, could not be reached in vivo because of their toxicity.

Bates is now interested in extending her experiments using doxicycline, as well as testing other tetracyclines. To potentially improve the assay’s readout, Lindquist suggested using mice carrying GFP-labelled huntingtin to monitor aggregation, instead of immunfluorescence. This could help prevent problems that can arise when using antibodies to monitor aggregates, such as antigen masking. Bates replied that she has tried setting up this system, but has not yet achieved high enough expression levels.

Participants were encouraged by the many advances in the development of screening assays and by the diversity of systems which promise to complement each other. Johnson noted, however, that additional assays, that more closely reflect what occurs in the diseased human brain, are needed. He pointed out that this was an area with particular growth potential and encouraged participants to move in this direction.

In addition to using systematic screens to identify new therapeutic candidates, Young suggested examining the medical records of HD patients in search of hidden associations. For example, these records could reveal correlations between HD progression and diet or the use of specific dietary supplements or medications. Housman cautioned, however, that studies in which a large number of variables are tested simultaneously yield a high number of false positives --extremely large numbers of patients and sophisticated statistical analyses are needed to ensure the significance of any particular correlation. In addition, medical records are often limited in their scope and reliability.

Candidate therapies

            One of the most exciting candidate therapies discussed at the workshop was the histone deacetylase inhibitor SAHA. As previously mentioned, Thompson and her colleagues showed that inhibition of histone deacetylases (HDACs) in Drosophila dramatically decreases cell death and neurodegeneration caused by mutated huntingtin. Now Bates has found that SAHA greatly improves the motor performance of R6/2 mice, moving this candidate an important step towards testing in clinical trials. Describing her recent experiments, Bates noted that her team originally had trouble administering SAHA because it was too viscous to inject. But by mixing it with cyclodextrins and placing it in the animals’ drinking water, they were able to administer it effectively. They showed that the compound could cross the blood-brain barrier and affect histone acetylation in the brain. Conducting Rotarod tests on R6/2 mice treated with SAHA, Bates observed dramatic improvements in the animals’ motor performance. After only three weeks of continuosly receiving 0.67 g/l of SAHA, 8-week-old mice were performing significantly better than placebo-treated controls. And at 12 weeks, they were performing as well as the placebo-treated controls performed at 8 weeks. Thus, SAHA delayed Rotarod performance decline by 50%. The results are even more striking, considering that the mice were exposed to an enriched environment, which alone improves Rotarod performance.

            As expected, SAHA appeared to have no effects on aggregate formation based on a non-quantitative examination of brain tissue. It did not cause observable changes in gross morphology either. However, a qualitative examination revealed some regression of the loss of Nissl staining associated with HD progression. 

            Encouraged by these findings, participants discussed future directions. Bates noted that she has begun collaborating with Jim Olsen to examine the effects of SAHA on gene expression using microarrays. Lindquist added that it may be useful to establish whether SAHA’s effects are mediated exclusively through histones, or if the acetylation of other proteins, such as p53, contributes to the observed effects.

Kelly emphasized exploring the therapeutic implications in greater depth. For example, it will be important to identify methods for sustained dosing in humans. Bates agreed and added she is planning to test newer, more potent inhibitors. Thompson suggested testing the effects of HDAC inhibitor cocktails –by using inhibitors with additive effects, it may be possible to reduce the individual doses of each drug, thus reducing overall toxicity.

Compounds that target chaperone proteins are also emerging as potentially promising therapeutic candidates. Indeed, as reported by Bates, several labs, including hers, are currently performing crosses between HD mouse models and transgenic mice overexpressing chaperone proteins to test this idea. These proteins are particularly interesting targets because they may affect the disease process in a number of ways. On the one hand, they may affect mutated huntingtin folding and consequently have effects on aggregate formation and protein clearance. In addition, as previously mentioned, they play a role in keeping proteins out of the nucleus, which may be critical for huntingtin toxicity. Also, as noted by Lindquist, increasing chaperone levels in the cell may prevent toxicity by sequestering mutated huntingtin, reducing the protein’s free cytoplasmic levels. Another interesting facet about chaperone modulation, as noted by Fishbeck, is that it may prove useful for the treatment of several poly-glutamine diseases. Lindquist agreed but cautioned that different diseases often involve different interactions with chaperone proteins, and added that the normal function of chaperones is very complex. For example, Hsp90 has approximately 200 different cellular targets, and Hsp90’s interactions with different sets of them depends on the levels of Hsp90 expression. Another therapeutic consideration is the ability of chaperone regulators to cross the blood-brain barrier. Lindquist noted that efforts are underway to develop such compounds. 

One compound that regulates the heat shock response and may have therapeutic effects in poly-glutamine diseases is the herbal supplement celastrol. As described by Fishbeck, celastrol reduced toxicity and cell death in a model of SBMA used in a large screening effort led by the NINDS. Fishbeck is now pursuing the study of this compound, which has yet to be tested in mice.

Another regulator of the heat shock response with therapeutic potential is geldanamycin. Geldanamycin binds to Hsp90 which triggers increased production of Hsp70 and Hsp40 which, in turn, prevents mutant huntingtin aggregation. Several HD investigators, including Bates, are currently working with the compound. Lindquist noted that this was a powerful drug because its dose can be titrated to make its effects specific for certain targets. In addition, Signer pointed out that other inhibitors of Hsp90 are currently being studied. Neal Rosen, for example, is testing several for their therapeutic use in cancer patients.

            Participants briefly mentioned the status of other candidates that target huntingtin aggregates. Thompson and Bates have been testing poly-glutamine constructs interrupted by stretches of poly-alanines as potential suppressors of aggregation. In Drosophila, these constructs appear to have dramatic effects, reported Thompson. In mice, however, the data are still unclear. Bates is still analyzing her results, but noted that one construct was toxic and another appeared to have no effect. Others have tried using antibodies to prevent aggregation. Reinhart noted that some antibodies have provided partial suppression of aggregation, but Johnson said that Paul Patterson’s MW1 antibody actually enhanced aggregation. Aptamers have been found to reduce aggregation, but interfere with transcription.

            Targeting the electrophysiological disruptions of HD also emerged as a potential therapeutic option. Based on Levine’s findings that calcium signaling is altered in HD mice, Surmeier suggested testing compounds that inhibit L-type calcium channels, such as nifedepine which is already approved for other clinical applications. He also noted that crossing an L-type calcium channel knockout mouse with an HD mouse model, such as R6/2, might be worth pursuing. L-type calcium channels are activated in a voltage-sensitive manner by AMPA and kainate currents and can boost the NMDA response. Surmeier suggested that moderately dampening this calcium signal might help protect striatal cells from overstimulation. Levine agreed with this proposal, but added it will be important to target the appropriate time window in disease progression, since the electrophysiological alterations he observes are biphasic.

In addition, it may be useful to test the effects of regulating other components of the glutamate signaling pathway. As pointed out by Chicurel, recent data from Lynn Raymond’s group suggests that, to counter the selective degeneration of striatal cells in HD, it may be important to specifically regulate NMDA signaling, and in particular signaling mediated by NMDA NR2B receptors. Striatal cells seem to be particularly vulnerable to excitotoxic cell death mediated by NR2B receptors, but not other mediators of calcium signaling.  The calcium profile generated by L-type calcium channels, however, differs significantly from that generated by NMDA receptors. Activation of NMDA receptors causes a rapid but transient increase in ERK activity, whereas L-type channels mediate sustained ERK activation. And as mentioned previously, this difference results in distinct downstream, transcriptional effects.

Another therapeutic candidate briefly mentioned at the workshop was the use of RNAi. Fishbeck is exploring this approach to inactivate mutated alleles in poly-glutamine diseases other than HD. Although similar studies are currently underway for HD, these were not discussed in detail since they were the focus of another recent workshop (see December 2002 workshop report).

            Two therapeutic candidates discussed at the workshop are already in clinical trials: co-enzyme Q10 (CoQ10) and minocycline. CoQ10 is involved in mitochondrial metabolism and is a scavenger of free radicals. Flint Beal and colleagues showed that it improves the symptoms of R6/2 mice. Presenting an update of CoQ10’s performance in a clinical trial led by the Huntington Study Group, Young noted that the progression of HD decreased by 13% in patients treated with either CoQ10 or CoQ10 and remacemide, a glutamate antagonist. (Remacemide alone had no measurable effects). However, the decrease was not statistically significant because the study was powered to detect effects of at least 40%. Since CoQ10 appears to be completely innocuous, and twice the dose used in this study was shown to decrease the progression of Parkinson’s disease in a small trial, Young noted that a new trial is being considered using higher doses of the compound and a larger number of patients.

As mentioned previously, minocycline has been considered a potential therapeutic agent, yet recent tests of its effectiveness have yielded mixed results. Minocycline originally caught the foundation’s attention when Robert Friedlander reported that minocycline-treated mice lived significantly longer than controls.  These mice also showed improved performance on the Rotarod, with no changes in blood glucose levels or weight. Friedlander proposed that the antibiotic was acting through the inhibition of caspases. Based on these findings, minocycline was fast-tracked into a human clinical trial. Yet as pointed out by Tobin, several recent studies indicate that minocycline is probably acting through other mechanisms, such as through aggregate inhibition and, more importantly, that it may not be very effective, even in animal models of disease. Thompson, for example, was unable to detect any effects of minocycline on aggregate formation in Drosophila. More strikingly, Bates observed no behavioral improvement, nor obvious change in aggregate formation in mice, even though she estimates she achieved a concentration ten-fold higher than that reported by Friedlander. She also noted that she couldn’t increase the dose any further because of its toxicity. As proposed by several participants, delivering the drug directly to the brain through a cannula might circumvent this problem, but the discrepancy between findings still remains unsolved.

Bates suggested that Friedlander’s study might be difficult to reproduce because it was based on only 6 or 7 animals, and the placebo-treated group had a very low survival rate to start with. Tobin added that variances in R6/2 strains might have contributed to the discrepancies and Levine noticed that the mice used were of different ages. Bates used younger mice than Friedlander, and as previously mentioned, the pathology of the disease changes significantly as a function of time. In additon, the higher doses used by Bates may not necessarily be more effective. As pointed out by Chesselet, different dosages can result in the targeting of different systems. Indeed, Bates herself has found that Congo red inhibits aggregate formation at one dose, but actually enhances it at another.

Seeking to improve the selection of candidate therapies that are moved into clinical trials, Tobin asked what lessons could be learned from minocycline. Should there have been more experiments performed before moving minocycline into clinical trials? How should the foundation proceed in the future? As Wexler noted, the foundation wants to encourage rigorous pre-clinical research, but at the same time, wants to accelerate the search for a cure and realizes that animal models aren’t completely predictive.

Optimizing decision-making at the clinical trial stage is critical because of the trials’ cost in patient recruitment, money, and time. As pointed out by Wexler, the statistical power of studies, such as the CoQ10 study, depends greatly on the number of patients, which are often a limiting factor.  In theory, a drug that completely cured HD would require a very small set of patients and a low level of predictability to reveal its effects.  But the number of patients rises steeply when testing drugs with more moderate effects.

Some participants advocated developing a set of standard criteria to guide the selection of compounds tested in clinical trials. Goldstein, for example, said that, although the criteria didn’t have to be formal, they should at least include reproducing the key results, under the exact same conditions, in one or two independent labs. Carl Leventhal agreed and added that setting up a reliable machinery for putting worthy candidates into clinical trials was a top priority. He emphasized the importance of clearing up experimental discrepancies by fostering maximum disclosure and communication between screeners.

Kelly noted, however, that although criteria could be useful, most candidates should be evaluated on a case-by-case basis. He commended the foundation’s efforts, noting that it was hard to predict what would happen with minocycline and, given that it is an approved drug, it was reasonable to speed up its testing in humans. Young agreed and pointed out that, based on  the new studies, the Huntington Study Group has now truncated the full set of minocycline trials.

One resource that promises to help streamline the movement of compounds into clinical trials is a pre-testing facility being set up by the High Q Foundation. It is intended as an engineering installation, equiped to test promising therapeutic candidates generated by scientific research. A representative of High Q at the workshop encouraged participants to provide suggestions to optimize its performance. In addition, Wexler urged participants to advise the Hereditary Disease Foundation on how it can improve communications between researchers and Minka von Beuzekom noted the importance of keeping funds flowing smoothly into the projects with the most potential. Furthermore, Fishbeck, who works at the NIH, noted that he was interested in the distribution of funds towards therapeutic endeavors and hoped he could also help in this area.

Perhaps as important as deciding on new research paths to pursue therapeutically, noted Tobin, is deciding when to drop research avenues that lose their promise. Developing efficacious strategies to address this issue will be particularly important as the number of candidate drugs rapidly increases.

A few final thoughts

            Despite the many challenges associated with HD, participants were optimistic about the future. Kelly and Lemke both commented on the tremendous progress they have witnessed in the past few years and praised the foundation for its efforts. Lemke considered that a key emerging concept, that will undoubtedly help guide future attacks on HD, is the disease’s multiple pathologies, occurring at multiple times. Reiterating Constantine-Paton’s proposal of HD as a multi-hit disease, he suggested that the best plan of action will probably entail fostering the exploration of a variety of approaches in parallel.

            The growing reciprocal enrichment of basic and applied science in HD was also noted as a very positive trend. Lemke noted that when he first entered the HD field, there was a void in cell and molecular knowledge. Yet HD research has advanced so much recently, that it has actually become a driving force behind some basic cellular and molecular questions. Indeed, Lindquist pointed out that the study of glutamine-rich proteins and their structure may not only yield important therapeutic insights, but advance our understanding of protein folding, prion behavior, differentiation and development. 

            Looking toward the future, participants predicted that several new therapeutic candidates will emerge within the next few years. As discussed previously, the optimization of the movement of candidate drugs into clinical trials will thus be of central importance. In addition, the chemical optimization of active compounds is also likely to be of increasing value. Kelly noted that it is probably time to begin recruiting more chemists to streamline drug development.

List of action items:

1.      Basic science -- Cell biology

a.      Assess whether the cellular transport machinery in HD mice is altered (Goldstein).

b.      Extend studies of aggregate foci (Osmand). Test technique on fresh tissue and on more tissues from pre-symptomatic patients (Osmand, Wexler). Test technique on model organisms and other poly-glutamine diseases (Lemke). Compare the detection of foci in cultured cells expressing poly-glutamine chains of different lengths (Johnson).

c.      Characterize the local distribution of receptors and scaffolding in HD cortical and striatal synapses and correlate with the normal developmental changes (Constantine-Paton).

d.      Search for correlations between neurons’ differentiation state and HD pathology in vertebrate systems (Constantine-Paton, Thompson).

e.      Investigate the differences between neuronal and glial proteasomes (Tobin). Conduct co-culture experiments (Young) and generate heterokaryons (Signer).

f.        Assess the effects of HD on JNK signaling (Goldstein)

g.      Extend studies to determine how RNQ affects aggregate toxicity (Lindquist).

h.      Investigate potential strategies to prevent nuclear transport of poly-glutamine proteins (Tobin). Identify specific chaperones involved; stabilize cytoplasmic interactions with chaperones (Lindquist). 

i.        Test whether aggregate formation is dominant using cell fusions (Thompson)

j.        Complement expression analyses with proteomics (Constantine-Paton)

k.      Continue studies on other poly-glutamine diseases (Fishbeck). Investigate effects of celastrol on SBMA, effects of ligands in mutated androgen receptors, transcriptional dysregulation.

2.      Basic science -- Electrophysiology

a.      Characterize the progression of HD-associated electrophysiological alterations of striatal cells and their dependence on different glutamate receptor subtypes (Levine).

b.      Characterize early alterations in membrane behavior in slices transfected with long poly-glutamine constructs using biolistics (Reinhart)

c.      Examine the role of dopamine receptors in HD (Levine, Surmeier, Chesselet). Study the effects of dopamine on cortico-striatal electrophysiological responses. Cross HD mouse models with transgenic mice carrying D2 or D1 receptors tagged with GFP.

d.      Characterize normal and HD cortical electrophysiology (Surmeier).

e.      Characterize normal and HD striatal-pallidal transmission (Chesselet).

f.        Test the hypothesis that HD aggregates interfere with electrical conductance in neurites (Wilson, Goldstein).

g.      Cross L-type calcium channel knockout with HD mice (Surmeier)

3.      Screening assays

a.      Continue yeast genetic screens for inhibitors of toxicity (Lindquist).

b.      Assess whether genes that affect lifespan in C. elegans are relevant to HD (Dillin). In particular, test effects of altering mitochondrial function in HD mice.

c.      Continue screens in PC12 cells for inhibitors of cell death (Tobin) and aggregation (Thompson).

d.      Continue screens in Drosophila for inhibitors of cell death and neurodegeneration (Thompson, Housman).

e.      Extend studies with biolistics technique (Reinhart). Complement studies in biolistics system with transgenic models of HD.  Improve automation of software for analyzing neurite degeneration (Reinhart, Kelly). Perform co-transfections using huntingtin and HD suppressors using biolistics (Signer). Use biolistics technique in vivo (Constantine-Paton).

f.        Filter compound libraries using ‘smart’ chemistry criteria (Reinhart). Pursue study of mechanism-based, enzyme-specific inhibitors (Kelly).

g.      Use organotypic slices from mice expressing GFP-labeled huntingtin to monitor aggregation (Bates, Lindquist).

h.      Develop assays that more closely mirror the disease process in human brains (Johnson).

i.        Examine medical records of HD patients for undiscovered correlations (Young), but limitations of approach should be considered (Housman).

4.      Therapeutic candidates

a.      Continue characterization of HDAC inhibitors as potential therapeutic agents (Thompson, Bates).

*Examine effects of SAHA on gene expression using microarrays (Bates)

*Determine whether other acetylated proteins, besides histones, contribute to SAHA’s effects (Lindquist)

*Identify methods for sustained dosing of HDAC inhibitors in humans (Kelly)

*Test new HDAC inhibitors in HD mice (Bates)

*Investigate the use of HDAC inhibitor cocktails to reduce doses of individual drugs (Thompson).

b.      Extend studies of the effects of doxycyline and other tetracyclines on aggregation (Bates).

c.      Cross chaperone transgenic mice with HD mice (Bates)

d.      Continue to investigate effects of geldanamycin (Bates, Lindquist)

e.      Test effects of L-type calcium channel inhibitors, such as nifedepine (Surmeier)

f.        Design new CoQ10 trial with increased dosages and numbers of patients (Young)

g.      Recruit chemists to design improved variants of candidate compounds (Kelly)

5.      Logistics

a.      Improve availability of mouse models. Most models are freely available, but participants should contact HDF if problems arise (Johnson). Set up repository for mouse models (Goldstein).

b.      Develop criteria to streamline movement of therapeutic candidates into clinical trials:

*Establish minimum number of experimental repetitions in independent labs (Goldstein)

*Keep criteria flexible, assess situations on a case-by-case basis (Kelly)

*Increase disclosure and communication between drug screeners (Leventhal)

c.      Provide suggestions to Hi Q Foundation for setting up a pre-testing facility for candidate compounds

d.      Develop criteria to terminate projects (Tobin).

e.      Continue the attack of HD on multiple fronts (Lemke)