CHDI’s 12th Annual HD Therapeutics Conference took place April 24 – 27, 2017, in Malta. This unique conference series focuses on drug discovery and development for Huntington’s disease, and draws participants and speakers from the biotech and pharmaceutical sectors as well as academia and research institutions. The conference is intended as a forum where all participants can share ideas, learn about new disciplines, network with colleagues and build new collaborative partnerships. We are indebted to all of the conference speakers, and especially grateful to those who are able to make their presentations available here for a wider audience.
- Genetic modifiers implicate the DNA damage response in the trinucleotide repeat disorders Lesley Jones, PhD, Cardiff University
- Identification of genetic variants associated with Huntington’s disease progression Davina J. Hensman Moss, MA, MBBS, University College London
- Haplotype analysis of the Huntington disease mutation reveals insights to therapy, ancestry, and biologyChris Kay, University of British Columbia
- Molecular and temporal dynamics of Huntington’s disease: Impact on stress response systemsChristian Neri, PhD, University Pierre and Marie Curie
- Proteomics technologies for biomedical applications in neurodegenerative diseasesMatthias Mann, PhD, University of Copenhagen
- HTT genetic discovery Marcy MacDonald, PhD, Massachusetts General Hospital
- Cryptic polyglutamine repeat sequence variation and somatic instability in Huntington disease: Drivers of pathology? Darren G. Monckton, PhD, University of Glasgow
- Toward allele-selective RNA-targeted ligands for Huntington’s disease Kevin M. Weeks, PhD, University of North Carolina at Chapel Hill
- Developmental pathogenesis of Huntington’s disease Mark F. Mehler, MD, Albert Einstein College of Medicine
- Isogenic Huntington’s disease human embryonic stem cell lines reveal novel development phenotypes Albert Ruzo, PhD, The Rockefeller University
- Introduction Andy Howard, CHDI
- Biofluid biomarkers for HD: The State of the Union Edward Wild, PhD, University College London
- Htt reduction: Challenges and concerns in development of HD therapeutics, a perspective of a SCA1 investigator Harry T. Orr, PhD, University of Minnesota
- HTT-CRISPR editing with the Kamicas9 self-inactivating system Nicole Deglon, PhD, Lausanne Univeristy Hospital(CHUV)
- Screening approaches to identify small-molecule modulators of huntingtin protein levels Elizabeth M. Doherty, PhD, CHDI
- Introduction Roger Cachope, MD, MS, CHDI
- Characterizing neuronal dysfunction in clinical and preclinical studies Andrew Leuchter, MD, University of California, Los Angeles
- Exploring the therapeutic potential of the endocannabinoid system in Huntington’s disease Joseph F. Cheer, PhD, University of Maryland School of Medicine
- Abnormal STN-GPe network activity in the Q175 knock-in mouse model of Huntington’s disease Joshua W. Callahan, PhD, Northwestern University
- Progenitor cell mobilization and replacement for striatal repair in Huntington’s disease Abdel Benraiss, PhD, University of Rochester Medical Center
- Modeling duality between clinical progression and circuit dysfunction in Huntington’s disease to inform therapeutic intervention James Kozloski, PhD, IBM Research
- Neuroinflammation: What is it? What does it have to do with HD? Richard M. Ransohoff, MD, Biogen
- CYP46A1 gene therapy to restore brain cholesterol metabolism in Huntington’s disease Nathalie Cartier-Lancave, MD, University of Paris Sud/University Paris Saclay
- Cell replacement therapy for Huntington’s disease: Progress and challenges Anne Rosser, MA, PhD, MB BChir, Cardiff University
- Modeling Huntington’s disease with patient neurons generated by direct neuronal reprogramming Andrew Yoo, PhD, Washington University School of Medicine
Genetic modifiers implicate the DNA damage response in the trinucleotide repeat disorders
Lesley Jones, PhD, Cardiff University
The polyglutamine diseases, including Huntington’s disease (HD) and multiple spinocerebellar ataxias are caused by expanded CAG tracts, encoding glutamine, in unrelated genes. Longer CAG repeat tracts are associated with earlier ages at onset in HD and the other trinucleotide repeat diseases. CAG repeat length, however, does not account for all of the difference in age at onset, and the existence of additional genetic modifying factors has been suggested in these diseases. A recent genome-wide association study in HD detected association between age at onset and genetic variants in the DNA damage response genes. This implies that such modifiers might act on the CAG expansion mutation, rather than on mutant huntingtin-specific pathways. In order to test this hypothesis we examined the HD associated modifying variants in the DNA damage response in other repeat expansion associated disorders, to investigate whether they also modified age at onset in these diseases.
We assembled an independent cohort of 1,462 subjects with HD and CAG repeat expansion spinocerebellar ataxias, and genotyped single-nucleotide polymorphisms selected from the most significant hits in the genetic modifiers of HD study. In the analysis of DNA repair genes as a group, we found the most significant association with age at onset when grouping all polyglutamine diseases (HD+spinocerebellar ataxias; p = 1.43 × 10–5). In individual SNP analysis, we found significant associations for rs3512 in FAN1 with HD+spinocerebellar ataxias (p = 1.52 × 10–5) and all spinocerebellar ataxias (p = 2.22 × 10–4) and rs1805323 in PMS2 with HD+spinocerebellar ataxias (p = 3.14 × 10–5). All associations were in the same direction as in the HD GWAS.
This study shows that genes of the DNA damage response significantly modify age at onset in HD and spinocerebellar ataxias, suggesting a common pathogenic mechanism which operates on the CAG expansion in the DNA of the genes. This could potentially operate through the observed somatic expansion of repeats that is known to be modulated by genetic manipulation of genes and proteins of the DNA damage response in disease models. This finding potentially offers novel therapeutic opportunities in multiple neurodegenerative diseases.
Identification of genetic variants associated with Huntington’s disease progression
Davina J. Hensman Moss, MA, MBBS, University College London
Huntington’s disease (HD) is a devastating neurodegenerative disease, caused by a CAG repeat expansion in HTT. AAO has previously been used as a quantitative phenotype in genetic analysis of modifiers of HD, but AAO is hard to define and not always available. Here we therefore aimed to generate a novel measure of disease progression, and identify genetic markers associated with this measure.
We generated a progression score based on principal component analysis of prospectively acquired longitudinal changes in motor, behavioural, cognitive and imaging measures in the TRACK-HD cohort.
We generated a parallel progression score using 1773 previously genotyped subjects from the European REGISTRY study. 216 subjects from TRACK-HD were genotyped. Association analyses were conducted using mixed linear models, with gene-wide analysis using MAGMA. Meta-analysis of TRACK-HD and REGISTRY was performed using METAL. Co-localisation of signals between studies, and with expression data, were investigated.
Longitudinal motor, cognitive and imaging scores were correlated with each other in TRACK-HD subjects, justifying a one-dimensional cross-domain measure (principal component score) as a unified progression measure in both studies. The TRACK-HD and REGISTRY progression measures were correlated with each other (r=0·674), and with AAO (r=0·315, r=0.234 respectively). A meta-analysis of progression in TRACK-HD and REGISTRY gave a genome-wide significant signal (p=1.12×10-10) on chromosome 5 spanning 3 genes, MSH3, DHFR, MTRNR2L2. The lead SNP in TRACK-HD is genome-wide significant in the meta-analysis (p=1.58×10-8), and encodes a protein change in MSH3. In TRACK-HD, each copy of the minor allele at this SNP is associated with a 0.4 (95% CI=0.16,0.66) units per year reduction in the rate of change of the Unified Huntington’s Disease Rating Scale (UHDRS) Total Motor Score, and 0.12 (95% CI=0.06,0.18) units per year in the rate of change of UHDRS Total Functional Capacity. Notably, these associations remained significant after adjusting for AAO.
The multi-domain progression measure in TRACK-HD is associated with a functional variant that is genome-wide significant in a meta-analysis. Observing such a strong association in only 216 subjects implies that this measure is a sensitive reflection of disease burden, that the effect size at this locus is large, or both. As knock out of MSH3 reduces somatic expansion in HD mouse models, this highlights somatic expansion as a potential pathogenic modulator, informing therapeutic development in this untreatable disease.
Haplotype analysis of the Huntington disease mutation reveals insights to therapy, ancestry, and biology
Chris Kay, University of British Columbia
The expanded CAG repeat causative of Huntington disease (HD) is associated with specific genetic variants in the surrounding chromosomal region. Variants closest to the CAG repeat are known to occur as haplotypes that are coinherited with the CAG repeat over many generations. We performed dense SNP genotyping across the Huntingtin (HTT) gene in a range of ethnically distinct HD patient cohorts to characterize haplotypes of the HD mutation in different ancestry groups. A1, A2, and A3a represent the three most common HD mutation haplotypes in patients of European ancestry, allowing rational target selection for allele-specific HTT silencing in these populations. Different European populations have different frequencies of these haplotypes, with implications for allele-specific gene silencing strategies. Strikingly, the HD mutation also occurs most frequently on the A2 haplotype in patients of Middle Eastern and South Asian ancestry, whereas the HD mutation occurs on ethnically distinct haplotypes in African and East Asian patients. In the mixed subpopulation of South Africa, we show that the HD mutation has both major European and minor African origins. In Latin American patients, the HD mutation occurs most frequently on the A1 HTT haplotype as in Europeans, but on a distinct A1 haplotype variant also found in Amerindian controls, supporting multiple indigenous origins of the disease. We also reveal African haplotypes of the HD mutation in Latin America. The global diversity of A1 and A2 haplotypes was subsequently analyzed within the 1000 Genomes Project, and subtype variants genotyped in different HD patient populations. We reveal genetic traces of HD founder events in some patient populations, as previously suggested by genealogical evidence.
In addition to HTT haplotype characterization, we examined all common missense mutations in HTT to identify potential human modifiers of HTT post-translational modification (PTM). We reveal a SNP and HTT haplotype specific to East Asian populations that modifies post-translational myristoylation of HTT, resulting in downstream alterations to HTT proteolysis. This is the first SNP shown to functionally modify a PTM in HD, offering a candidate modifier acting through altered HTT biology.
Molecular and temporal dynamics of Huntington’s disease: Impact on stress response systems
Christian Neri, PhD, University Pierre and Marie Curie
The stress response hypothesis postulates that stress response is genetically regulated and biologically important to the reproduction, development and maintenance of living organisms. In so far, understanding how the brain may use specific stress response mechanisms to resist Huntington’s disease (HD) and maintain function over time, compensation, has significant therapeutic potential.
Compensation can be viewed as a biological program prescribing the temporal dynamics of molecular- and cellular-network remodelling in response to HD and age. We investigated this phenomenon on a system level. Using spectral decomposition (BioGemix platform), we generated two models of molecular dynamics in HD, including one model in which datasets are integrated across species, covering the developmental and adult phases of the HD process, and one model in which datasets are integrated across CAG-repeats and age points in the allelic series of HD knock-in mice, covering the presymptomatic-to-symptomatic stages of disease in these mice. We also generated a trans-network model of HD pathogenesis in which three kinds of networks are integrated, including HD networks obtained from analysis of RNA-seq data in the allelic series of HD knock-in mice by using spectral decomposition, weighted gene co-expression network analysis (WGNCA) or bayesian network inference (GNS Healthcare’s approach), achieving a strong level of gene prioritization.
These models consistently define a sequence of events in which specific stress-response systems may be altered in the course of the HD process. In particular, this indicates that intercellular communication may be persistently altered from the earliest stages of the HD process, during neuronal differentiation. The same applies to FOXO signalling, a stress response system central to the regulation of cell survival. The models involving the RNA-seq data collected in the HD knock-in mice indicate that cell survival mechanisms under FOXO control may respond to HD in the adult brain, gaining importance as HD mice are symptomatic or close to developing symtoms. These observations provide a platform for discussing how to target neuronal compensation pathways in HD, and we are currently examining the biological significance of the top hypothezes generated by these models.
Proteomics technologies for biomedical applications in neurodegenerative diseases
Matthias Mann, PhD, University of Copenhagen
Mass spectrometry (MS)- based proteomics has evolved into a powerful tool for investigating protein abundances, modifications and interactions on a global scale (Aebersold & Mann, Nature, 2016). It is now routinely possible to quantify essentially the entire proteome in cells and tissues, even from relatively low amounts of starting material. Post-translational modifications such as phosphorylation can be measured quantitatively for tens of thousands of sites and can even be followed over time in vivo (Humphrey, Mann,. Nature Biotechnology, 2015). We and others have also constructed proteome wide interaction maps, which provide important guidance towards protein function (Hein, Mann. Cell, 2015).
In this presentation, I will give a brief overview of where this still rapidly evolving technology stands at this point, giving examples from each of the main areas of application. I will then turn to translational and clinical applications, which are becoming ever more realistic. For example, our group has recently determined the hitherto elusive substrate of the Parkinson’s kinase LRRK2 in a mouse model of the disease (Steger, Mann, eLife 2016). We have also used MS-based proteomics to construct cell type specific proteomics maps of various mammalian organs and currently we are using the brain atlas to study proteomic differences in the brains of Alzheimer patients that had presented with symptoms vs. those that had not as well as controls.
In the context of the EU funded initiative ToPAG (partners Hartl, Klein and Baumeister, http://www.topag. mpg.de/), we have used proteomics to study aggregates in a severe mouse model of Huntingtin’s disease. For the first time, we established a quantitative inventory of these aggregates, which we compared to the soluble brain proteome and the CSF. This allowed us to home in on likely loss of function candidates, which we investigated in cell based rescue experiments.
The direct analysis of body fluids, especially plasma and CSF, in human patients would be especially attractive but has been hampered by great technological challenges. However, as I will show, these obstacles are now falling away. Using streamlined preparation techniques and new scan modes, we routinely identify and quantify 1000 proteins in CSF, and we are starting to use these capabilities in clinical investigations of patient cohorts in neurodegenerative diseases.
HTT genetic discovery
Marcy MacDonald, PhD, Massachusetts General Hospital
An unstable CAG trinucleotide repeat in HTT is the sole root genetic cause of Huntington’s disease. Observations from genotype-phenotype studies with Huntington’s disease individuals are guiding the discovery of the mechanism that initiates the disease process and the critical biological processes that underpin the onset and progression of symptoms. The impact of HTT genetic variants on huntingtin structure, posttranslational modifications and normal and abnormal functions will be presented. The findings will illustrate how HTT- focused genetic knowledge forms a fundamental foundation on which to build therapeutic efforts that aim to influence the primary cause of the disorder.
Cryptic polyglutamine repeat sequence variation and somatic instability in Huntington disease: Drivers of pathology?
Darren G. Monckton, PhD, University of Glasgow
It has been known for many years that in addition to repeat length variation, the exact sequence of the polyglutamine repeat tract and the adjacent polyproline also vary in both non-pathogenic and pathogenic HTT alleles. Likewise, it has been known for many years that the expanded CAG is somatically unstable in a process that is age-dependent, tissue-specific and highly expansion biased. Notably, very large alleles exceeding 1,000 repeats are observed in a subset of striatal neurons. These data strongly suggest that somatic instability contributes toward the tissue specificity and progressive nature of the symptoms. Indeed, it has been shown that the frequency of large expansions in cortical cells correlates with variation in age at onset not accounted for by inherited repeat length. Most recently, it has been demonstrated that DNA repair genes lie under some of the association peaks for genome wide analysis of variants contributing to variation in age at onset. In order to further address these issues we have developed a high-throughput DNA sequencing and bioinformatic pipeline that allows us to determine the sequence of the polyglutamine and polyproline tracts in HTT in large numbers of individuals with HD and the general population. These studies have revealed an unexpectedly high frequency of atypical alleles in the general population with four previously observed variants being common. We have also detected a novel allele containing CCG interruptions within the CAG array in a large ‘premutation’ length allele. The majority of expanded HD alleles we have sequenced retain the expected structure of CAG and CCG repeats, but a subset of atypical alleles have been detected in patients. Using this approach we can also estimate the degree of somatic mosaicism present in the blood DNA of each participant. These data confirm that as expected, somatic mosaicism in the blood DNA of HD individuals is age- and allele length-dependent. We are currently attempting to understand how atypical alleles and individual-specific mutational dynamics contribute toward phenotypic variability in HD. We are also investigating how variants in the DNA mismatch repair genes contribute toward variation in somatic mosaicism.
Toward allele-selective RNA-targeted ligands for Huntington’s disease
Presentation not currently available
Kevin M. Weeks, PhD, University of North Carolina at Chapel Hill
An ideal, but unrealized, therapeutic for Huntington’s disease would involve selective inactivation of the disease-associated allele, while leaving the normal allele untouched and able to express the essential huntingtin protein. In healthy individuals, the first exon of each of the two alleles of the huntingtin (htt) gene contains a relatively short repeating CAG segment. The most common healthy allele has 17 repeats. In HD patients, one htt allele is abnormally expanded to between 36 and 70 CAG repeats, although shorter and significantly longer repeat regions have also been reported. Studies employing synthetic in vitro transcripts spanning the first exon of the htt gene show RNAs containing expanded CAG repeats clearly form complex structures that are distinct from the structures of transcripts with normal numbers of repeats. More recently, we have pursued studies showing that the expanded, disease-associated allele likely forms a three-dimensional architecture with a cleft targetable by small molecules. Ongoing studies examining the structure of htt mRNAs from native cellular RNA and their targetability by small molecules will be discussed. A strategy, using proprietary mutational profiling (MaP) technologies, for the identification and development of allele-specific small-molecule ligand hits will also be outlined. The net result of this work suggests that the unique three-dimensional RNA tertiary structure of the disease-associated htt transcript constitutes a compelling target for allele-specific small molecule therapeutic intervention.
Developmental pathogenesis of Huntington’s disease
Mark F. Mehler, MD, Albert Einstein College of Medicine
Neurodegenerative diseases exhibit developmental components to pathogenesis that encompass an active prodromal phase associated with stressor states that promote relaxation of lineage constraints and aberrant reprogramming of cell fate. This may enhance cellular vulnerability to diseases of aging. We have previously shown that HD knock-in (KI) mice (Hdh-Q111) display impaired developmental stem cell- mediated striatal medium spiny neurogenesis/subtype specification, expansion of intermediate progenitor species, derepression/aberrant signaling of core pluripotency factors, Nanog and Sox2 in generative
and differentiation zones and loss of striosome/matrix compartmentalization. To demonstrate that the developmental diathesis is linked to HD pathogenesis, we employed BACHD:CAG-CreERT2 mice, which carry mutant huntingtin (mHtt) modified to harbor a floxed exon1 containing the pathogenic CAG expansion. Following tamoxifen administration after striatal neurogenesis, the floxed mHtt-exon1 was removed and mHtt expression was terminated. These conditional mice displayed similar profiles of impairments to mice expressing mHtt throughout life: (i) striatal neurodegeneration, (ii) early vulnerability to NMDA-mediated excitotoxicity, (iii) impairments in motor coordination, (iv) temporally distinct in vivo abnormalities in striatal electrophysiological activity, and (v) altered corticostriatal functional connectivity and plasticity. Employing a similar paradigm with conditional huntingtin hypomorphic (Hdhd.hyp) mice, we observed: (i) striatal/cortical neurodegeneration, (ii) neurological/skeletal muscle alterations, (iii) white matter tract impairments/axonal degeneration, and (iv) subpallial heterotopias, aberrant striatal maturation and deregulation of gliogenesis. These findings reveal that loss-of-function effects are involved in HD developmental pathogenesis, abnormal ventral telencephalon patterning and gliogenesis. Additional developmental studies of Hdh-Q200 mice revealed intricate deregulation of the second dorsolateral (Gsx2) wave of subpallial oligodendrogliogenesis with compensatory enhancement of the first medioventral (Nkx2.1) subpallial wave, and a corresponding ventral to dorsal gradient of myelin-dependent abnormalities. Moreover, using huntingtin knockout and KI
(Hdh-Q111) ESC lines, we observed that huntingtin also promotes embryonic stem cell maintenance, germ layer formation, organogenesis, neural induction, primitive and definitive neural stem cell specification, organ and tissue-specific lineage elaboration, and mutant huntingtin impairs all of these earlier developmental processes. These cumulative observations suggest that HD gives rise to previously unanticipated brain and body-wide developmental deficits that may establish complex matrices of cellular vulnerability to disease (co-morbidities, inverse co-morbidities) involving both undifferentiated and differentiated cell types.
Isogenic Huntington’s disease human embryonic stem cell lines reveal novel development phenotypes
Presentation not currently available
Albert Ruzo, PhD, The Rockefeller University
Despite the fact that HTT was among the first disease-causing genes to be cloned over 20 years ago, the molecular mechanism of disease is still unknown, and more strikingly, even the wild-type function of HTT is still not clearly defined. Animal models of HD have been useful in understanding some aspects of cellular dysfunction, but unfortunately they do not fully reproduce the full symptomatology of human HD, and therefore better models systems are required. Here we describe the generation of a set of isogenic, CRISPR/Cas9 gene-edited, human embryonic stem cell (hESCs) lines with different CAG lengths, representing different levels of disease severity and onset. In order to understand the mechanisms of the disease, we also generated HTT+/- and HTT-/- isogenic hESC lines to determine whether HD phenotypes arise due to a gain- or a loss-of-function of HTT. We found that the simple alteration of the CAG length produced a constellation of early human developmental alterations at multiple stages: We found that when cultured on micropatterned substrates and induced to differentiate with BMP4, our isogenic lines self-organize and generate distinct CAG length-dependent, radially symmetrical patterns of germ fates, thus providing the earliest reported phenotypic signature for human HD. When submitted to neural induction, all lines generated neural rosettes and neural progenitors, but self-organization was perturbed in the HD and HTT-/- lines. Finally, at later time points during neural differentiation, examination of post-mitotic cortical neurons revealed the intriguing appearance of multinucleated giant progenitors and neurons. The frequency of these abnormal cells, which are caused by a failure in cytokinesis, increased proportionally to the CAG length, and were phenocopied in HTT-/- lines. Our results provide for the first time highly quantitative, human phenotypic signatures of HD, and strongly suggest that HD mutations cause abnormalities during embryonic development of the human brain, with the devastating consequences manifesting symptoms decades later.
Andy Howard, CHDI
Biofluid biomarkers for HD: The State of the Union
Edward Wild, PhD, University College London
With the first human trial underway, the need for objective biological markers of successful target engagement and meaningful biological effect by huntingtin-lowering therapeutics has never been more pressing. Significant progress has been made in recent years, especially in the quantification of mutant huntingtin and protein biomarkers of neuronal death in CSF.
Many challenges remain. We need to replicate proposed top-tier candidates in CSF, study them longitudinally to understand their predictive power, and validate each combination of biomarker and assay to the standards expected by regulators. We need more specific CSF markers to dissect the molecular pathogenesis of HD in the human CNS, and biomarkers that can tell us about the regional distribution of pathology throughout the course of the disease. And each new treatment approach – whether huntingtin-lowering or further downstream – needs its own bespoke palette of pharmacodynamic biomarkers.
In this introductory talk, I will briefly review the state of biofluid biomarkers and consider our readiness for testing and delivering novel therapeutics; I will give an update on current CSF collection efforts through the multisite HDClarity study; and I will present some recent findings of interest from the TRACK-HD study.
Htt reduction: Challenges and concerns in development of HD therapeutics, a perspective of a SCA1 investigator
Harry T. Orr, PhD, University of Minnesota
Along with Huntington’s disease (HD), spinocerebellar ataxia type 1 (SCA1) is a genetic neurodegenerative disorder due to an expansion of a CAG tract encoding glutamines in the affected protein. A major approach to developing disease-modifying treatments for such neurodegenerative conditions has been to identify cellular pathways that drive pathology, the manipulation of which would serve as potential therapeutic targets. Here I will discuss the possibility that the efficacy of pathway-based approaches is likely to be complicated by aspects of polyQ-mediated pathogenesis. Notably the expanded polyQ protein impacts many cellular pathways, reducing the likelihood that targeting any one pathway will have a substantial therapeutic effect. Furthermore, there is increasing evidence suggesting that RNA from the mutant gene may contribute to pathogenesis by mechanisms that are distinct from its translation to a mutant expanded polyQ protein. Therefore, I will argue for focusing initial therapeutic efforts on approaches that target expression of RNA from the affected gene. Yet, it continues to be important to understand the normal biology/function of the disease proteins to be targeted in order to effectively monitor potential deleterious outcomes of chronically lowering the targeted proteins. Moreover, it will be important to initiate such therapies early in disease progression, perhaps ideally in presymptomatic individuals.
HTT-CRISPR editing with the Kamicas9 self-inactivating system
Nicole Deglon, PhD, Lausanne Univeristy Hospital(CHUV)
Neurodegenerative disorders are a major public health problem because of the high frequency of these devastating diseases in the population. Genome editing with the CRISPR/Cas9 system is making it possible, for the first time, to modify the sequence of genes linked to these diseases directly in the adult brain. However, one major challenge for CNS applications is the development of a system allowing transient expression of the Cas9 nuclease to improve the biosafety and limit off-target events. Here we developed the kamicas9 self-inactivating system ensuring transient expression of the Cas9 protein with on-target performance similar to CRISPR/Cas9. We then used the non-homologous end-joining (NHEJ) pathway, active in post-mitotic cells, to inactivate the gene implicated in Huntington’s disease (HD) and demonstrate high HTT editing in neurons and glial cells derived from HD-iPS and in the striatum of HD mice. HTT inactivation was associated with a drastic reduction of HTT aggregate formation and reduced neuronal dysfunctions, confirming the potential of HTT editing. Sequencing of potential off-targets sites in CRISPR/Cas9 samples revealed a very low incidence with only one site above background level. Importantly, the editing frequency at this specific sequence was drastically reduced with the kamicas9 system. These results demonstrate the potential of the Kamicas9 self-inactivating editing technology for applications in neurodegenerative diseases.
Screening approaches to identify small-molecule modulators of Huntingtin protein levels
Elizabeth M. Doherty, PhD, CHDI
In the HD community there is considerable excitement about clinical trials of promising new huntingtin lowering therapeutics that are now underway or planned. The most advanced clinical candidates are the antisense oligonucleotides which, as large molecules, pose a challenge in terms of delivery and distribution throughout all compartments of the brain. In contrast, small-molecule drugs can be designed to enter the brain readily and, moreover, distribute widely throughout the brain and periphery. Consequently, the discovery of small-molecule modulators of HTT is a major strategic goal for CHDI Foundation.
In the past, screening efforts have resulted in the identification of small-molecule modulators of HTT levels that act through general protein synthesis, degradation, or autophagy mechanisms (e.g. mTOR, Hsp90, see Bard et al J. Biomol. Screen. 2014, 19(2), 191). These pathways act broadly on all proteins, are not selective for HTT, and therefore introduce a significant risk of off-target toxicity. To identify small molecules that modulate HTT protein levels specifically or, ideally, are selective for mutant HTT (mHTT) protein levels, we conducted an unbiased, cell-based, phenotypic screening campaign coordinating the work of multiple contract research organizations. We opted to use HD patient-derived embryonic stem cells (ESC) and induced pluripotent stem cells (iPSC) which express mHTT at endogenous levels for the screen, rather than rely upon engineered over-expression systems. These stem cell lines were also differentiated into neuronal precursor cells, medium spiny neurons and cortical neuronal populations for screening and profiling purposes. We used a time-resolved fluorescence resonance energy transfer (TR-FRET) detection assay with antibodies paired to distinguish between mHTT and total HTT (mHTT plus wild-type HTT protein) to achieve the sensitivity needed to measure changes in endogenous protein levels, and adapted to a microtiter plate for high-throughput screening. Importantly, we also developed a series of secondary assays to assess cytotoxicity, selectivity for mHTT over wild-type HTT, selectivity for HTT over other proteins (including another polyQ expanded protein), and effect in neuronal and stem cell-derived neuronal assays. Medicinal chemistry resources were engaged to support hit validation, mechanism of action studies, and confirmation of putative molecular targets. Our overall goals are to identify small molecules that lower mHTT, to rigorously validate novel molecular targets that modulate mHTT levels, and to use this information to select the best avenues for therapeutic intervention.
Discovering disease-modifying treatments for Huntington’s disease: Lessons learned from Alzheimer’s disease
Presentation not currently available
Steven M. Paul, MD, Voyager Therapeutics
Despite strong preclinical and clinical evidence establishing the genetic etiology and both the molecular and cellular pathogenesis for a number of inherited neurodegenerative disorders, effective therapeutic interventions have been very slow to emerge. In fact, to date, and despite massive investments by the biopharmaceutical industry, virtually all potential disease-modifying therapies for Alzheimer’s disease and Parkinson’s disease have failed. Why have these R&D efforts failed, especially given compelling biological, and in many cases genetic data, strongly supporting the therapeutic approach being taken? What have we learned from these failed R&D efforts and associated clinical trials that can be applied to current and future R&D efforts to discover and develop disease-modifying therapies for Huntington’s disease? In this presentation I will examine the multiple reasons for prior failed research programs and clinical trials in Alzheimer’s disease. Relevant issues from drug target validation and selection, to candidate drug optimization and selection, to the timing of intervention, to the appropriate use of biomarkers to enroll patients and accurately stage disease progression, to challenges in monitoring treatment effects (both for safety and efficacy) in patients will be discussed and exemplified using real-life examples from the recent R&D experience in Alzheimer’s disease. This list of “lessons learned” from past experience in Alzheimer’s disease may prove relevant as we plan for and embark on disease-modifying R&D efforts, including clinical trials, in Huntington’s disease.
Roger Cachope, MD, MS, CHDI
Characterizing neuronal dysfunction in clinical and preclinical studies
Andrew Leuchter, MD, University of California, Los Angeles
Exploring the therapeutic potential of the endocannabinoid system in Huntington’s disease
Joseph F. Cheer, PhD, University of Maryland School of Medicine
Huntington’s disease (HD) is an inherited neurodegenerative disorder caused by a polyglutamine (CAG) expansion in the huntingtin gene. Striatal and cortical dysfunction are the hallmark neuropathological features underlying dyskinesia (e.g., chorea) and akinesia at later stages. However, prominent psychiatric symptoms manifest prior to motor dysfunction. We have recently shown in the Q175 mouse model of HD that suppressed motivation – one of the earliest indicators of HD – is associated with compromised dopaminergic signaling and network dynamics in the nucleus accumbens (NAc). We will discuss how pharmacological elevation of endocannabinoid levels in Q175 mice rescues compromised neurobiological markers and ameliorates motivational deficits. Our findings indicate that modulating NAc abnormalities with endocannabinoid-based therapies may be beneficial in treating the prodromal psychiatric deficits of HD.
Abnormal STN-GPe network activity in the Q175 knock-in mouse model of Huntington’s disease
Joshua W. Callahan, PhD, Northwestern University
Huntington’s disease (HD) is a hereditary, trinucleotide repeat disorder that is associated with involuntary movement and impulsive behavior. The subthalamic nucleus (STN) and reciprocally connected external globus pallidus (GPe) are key components of the movement/action suppressing hyperdirect and indirect pathways of the cortico-basal ganglia-thalamo-cortical circuit. Abnormal STN-GPe network activity may therefore contribute to circuit dysfunction and symptoms in HD. Although ex vivo cellular electrophysiogical studies have revealed abnormal cellular and synaptic properties in the STN and GPe of HD mice, the impact of these abnormalities on the operation of the STN-GPe network in vivo has not been addressed with modern cell class-specific recording and optogenetic interrogation approaches. Thus, STN-GPe network activity was recorded in urethane-anesthetized 6-month-old Q175 knock-in (KI) HD and WT mice during cortical slow-wave activity (SWA) and periods of cortical activation/desynchronization (ACT) using silicon tetrodes and optrodes. Opsins were virally expressed in neurons to identify and manipulate the activity of specific cell classes in vivo.
D2-expressing striatal projection neurons (D2-SPNs) were hypoactive during SWA in Q175 KI mice, consistent with reduced dendritic excitability of D2-SPNs. Prototypic GPe neuron activity was elevated during SWA and ACT in Q175 KI mice, consistent with D2-SPN hypoactivity and elevated autonomous GPe activity. STN neuronal activity was reduced during SWA and ACT in Q175 KI mice, consistent with hyperactivity of prototypic GPe neurons. Furthermore, during optogenetic silencing of prototypic GPe neurons, STN activity remained diminished in Q175 KI mice, consistent with impaired autonomous STN activity. Finally, stereological analysis of vesicular glutamate transporter 1 (vGluT1)-immunoreactive axon terminals revealed that the density of cortico-STN axon terminals was reduced in Q175 KI mice, arguing that loss of synaptic excitation may contribute to STN hypoactivity. Together these data suggest that in Q175 KI mice dysfunction of multiple circuit elements underlies abnormal prototypic GPe neuron hyperactivity and STN hypoactivity.
Progenitor cell mobilization and replacement for Striatal repair in Huntington disease
Abdel Benraiss, PhD, University of Rochester Medical Center
Huntington’s disease (HD) is characterized in part by the loss of neostriatal medium spiny neurons (MSNs). In previous studies, we found that new MSNs might be produced from endogenous neural stem cells, by mobilizing these cells from the ventricular subependyma by intraventricular delivery of viruses expressing BDNF and noggin. Importantly, this strategy substantially delayed disease progression in R6/2 huntingtin mutant HD mice (J. Clin. Invest. 117:2889, 2007). Pallidal backfills followed by patch clamp confirmed the maturation and circuit integration of new striopallidal neurons, while survival analysis revealed that AAV4-BDNF/noggin-treated R6/2 mice exhibited delayed motor deterioration and substantially increased survival (Cell Stem Cell 12:787, 2013). Rabies virus tracing further established that newly generated striatal neurons attracted corticostriatal afferent inputs, and thereby mediated the establishment of new corticostriatopallidal circuits in treated adult animals. These findings suggested that BDNF and noggin-induced neuronal addition might provide a promising means of disease modification in HD. On that basis, we next asked if modification of the resident glial environment might further potentiate the survival of medium spiny neurons, both resident and newly-generated, in the HD environment. To that end, we transplanted normal HTT WT human glial progenitor cells (GPCs) into neonatal R6/2 x rag1-/- HD mice, using a donor cell phenotype and injection protocol that resulted in the substantial replacement of host mutant HTT striatal glia by WT human glia. These animals lived significantly longer and manifested slower motor deterioration than their untransplanted controls, while exhibiting a normalization of striatal potassium homeostasis, attenuated HD-associated hyperexcitability, and improved cognitive performance (Nature Comm., 7:11758 DOI: 10.1038, 2016). Together, these findings suggest that a combined strategy of mobilizing endogenous neural stem cells to produce new neostriatal neurons, paired with restoration of a wild-type glial environment using transplanted glial progenitor cells, may offer significant benefit in the treatment of Huntington’s Disease.
Modeling duality between clinical progression and circuit dysfunction in Huntington’s disease to inform therapeutic intervention
James Kozloski, PhD, IBM Research
The progression of Huntington’s disease (HD) is marked by clinical changes in cognitive, affective, and motor function among mutant gene carriers, as well as changes in gross morphological measures in patient brain imaging. Animal disease models reveal widespread cellular, synaptic, microcircuit and brain circuit changes due to the mutant Huntington protein (mHTT). Each of these changes emerges at multiple scales, often simultaneously, and their measures therefore overlay the primary risk of mHTT, which accumulates locally, with compensation for this risk, which emerges globally in the brain’s circuitry resulting in clinical phenotypes. Identifying when, where, and how to intervene with agents that reverse the effects of mHTT therefore requires first a mapping from clinical observations and disease categories onto related brain circuits and their neurophysiological states. Next, within this modeling duality, therapeutic compensation for mHTT can proceed as a formal reversal of phenotypes at the clinical level of the model, effected by inputs deemed necessary using simulations at the circuit-level.
We have undertaken at IBM Research a multiscale modeling approach to HD that incorporates disease progression modeling across clinical phenotypes and brain imaging biomarkers with brain circuit neurophysiological models to implement this duality. We present results of a population model of the disease, with pre- and post-symptomatic cognitive, functional, and motor states and their transitions inferred from recent HD longitudinal studies. These states are compared to novel structural and functional brain biomarkers extracted by machine learning from MRI, fMRI, and DTI data to inform a clinical model of circuit degeneration and dysfunction. Finally, this model informs our computational neuroscience models of neuronal and circuit electrophysiology, validated against preclinical HD animal model data. We show how cellular, synaptic, microcircuit, and brain circuit functional regulatory units can be discovered using these simulations and then manipulated to effect and predict disease progression, and various outcomes from therapeutic inputs to the models at each scale.
Neuroinflammation: What is it? What does it have to do with HD?
Richard M. Ransohoff, MD, Biogen
Neurodegenerative disorders such as Huntington’s disease (HD) are among the most challenging of medical and scientific problems due to the complexity of central nervous system (CNS) cells, network circuitry and the outputs: subjective experience and objective behavior. Neurons are the primary CNS cells, carrying out functions such as signal transmission and network integration, and are the main targets of neurodegenerative disease. It’s crucial also to consider how the neuron’s environment contributes to neurodegeneration. Maintaining an optimal milieu for neuronal function rests with supportive cells termed glia and the blood-brain barrier. For monogenic disorders like HD, one needs to keep firmly in focus that all cells of the affected organ harbor the mutant gene. Within the broad field of neuroscience research, accumulating evidence indicates that neurodegeneration occurs in part because the neural environment is affected in an assemblage of processes collectively termed neuroinflammation. Notably, therapies targeting glial cells might provide benefit for those afflicted by neurodegenerative disorders.
CYP46A1 gene therapy to restore brain cholesterol metabolism in Huntington’s disease
Nathalie Cartier-Lancave, MD, University of Paris Sud/University Paris Saclay
Cholesterol metabolism dysfunction exerts devastating effects on developing and adult neurons both in vivo and in vitro. Most importantly, cholesterol homeostasis defects in the adult brain are linked to neurodegenerative diseases, such as Niemann-Pick type C, Alzheimer (AD) and Huntington’s disease (HD). In AD and HD, cholesterol homeostasis defects involve a general perturbation in the expression of cholesterol biosynthesis enzymes. 24S-hydroxycholesterol (24OH-Chol), the catabolite of cholesterol, is decreased in the plasma of HD patients. Levels of CYP46A1, the rate-limiting enzyme which catalyzes the production of 24OH-Chol in neuronal cells, are decreased in the striatum of HD mice models and in the putamen of HD patients. CYP46A1 plays major roles in activating brain cholesterol turnover and thus increasing the mevalonate pathway, with consequent beneficial effects on synaptic plasticity and function. Restoring CYP46A1 expression in vivo by adeno-virus-mediated (AAV-CYP46A1) delivery in the striatum of R6/2 mice results in significant improvement in motor behavior associated with decreased aggregates of mutant Huntingtin and neuroprotection. Cholesterol biosynthesis pathway is restored in the targeted brain regions, leading to a normalization of cholesterol biosynthesis intermediates, desmosterol and lanosterol, but also cholesterol itself and 24OH-Chol. By contrast, CYP46A1 knockdown using AAV carrying a short hairpin (sh) RNA directed against CYP46A1 in the striatum of wild-type mice results in neurodegeneration and behavior abnormalities. Doses-response studies in mice and translational steps in non-human primates are ongoing to optimize the clinical protocol. The feasibility and tolerance of AAV gene therapy to treat neurodegenerative diseases have been increasingly demonstrated in human patients. Our goal is to propose a phase I/II clinical application to evaluate the efficacy and safety of a single administration of AAV-CYP46A1 in the striatum of patients with Huntington’s disease at an early stage of disease progression.
Cell replacement therapy for Huntington’s disease: Progress and challenges
Anne Rosser, MA, PhD, MB BChir, Cardiff University
Cell replacement therapy aims to replace cells lost to the disease process, with the rationale that they will re-innervate the host brain and repair the damage neural circuitry. In HD, the cells most affected are the medium spiny neurons (MSNs), and thus these are the principle target cells for cell replacement in HD. Cell replacement is only successful if the donor cells differentiate precisely into target cells (in this case, MSNs). In normal development MSNs differentiate within the part of the foetal brain called the ganglionic eminence (GE) and animal studies have shown that GE cells transplanted into the striatum of adult HD rodents can alleviate motoric and cognitive deficits. There is also pilot data suggesting that transplantation of human foetal GE cells can improve function in people with HD. However, the scarcity of foetal tissue of sufficient quality severely limits clinical application, quite apart from the ethical controversies surrounding this source. Thus, there is an imperative to identify a reliable renewable source of donor cells.
Many potential sources exist, but most attention recently has been on deriving MSNs from human embryonic stem cells (hESCs). hESCs are an attractive potential donor cell source as they can be expanded in vitro, cryopreserved, and differentiated into mature somatic cells. However, transplantation of hESC-derived MSNs is a complex multistep process. Repair-HD is a consortium of neuroscientists and clinicians who are working to bring together all the elements required for clinical translation of hESC-derived MSNs for HD. This includes optimisation of MSN differentiation protocols, GMP translation, rodent and primate transplant studies, development of improved clinical assessment tools, and consideration of novel trial designs to accommodate the constraints of this complex intervention. This presentation will present the current status of this work and will consider some of the challenges still to be overcome in order to bring stem cell therapies to clinical application for HD.
Modeling Huntington’s disease with patient neurons generated by direct neuronal reprogramming
Andrew Yoo, PhD, Washington University School of Medicine
Recent development in cell-fate reprogramming techniques demonstrated the possibility of generating human neurons by directly converting non-neuronal somatic cells. We recently established a reprogramming approach that generates human neurons using neuronal microRNAs (miRNAs), miR-9/9* and miR-124 (miR-9/9*-124) that when ectopically expressed, directly reprogram human fibroblasts to neurons. The miR-9/9*-124-induced neuronal state can be guided into specific neuronal subtypes with additional transcription factors. As such, we recently showed that miR-9/9*-124 with CTIP2, DLX1/2 and MYT1L converted human adult fibroblasts to a neuronal population highly enriched with striatal medium spiny neurons (MSNs), primary cell type affected in Huntington’s disease (HD) (Victor et al., Neuron, 2014). In addition, an interesting feature of converted MSNs is the maintenance of age stored in original fibroblasts, as demonstrated by the retention of age-associated epigenetic, transcriptomic and cellular marks (Huh et al. eLife, 2016). These findings suggest the potential of modeling HD in tissue culture, using patient MSNs that reflect the age status of adult donors. The applicability of MSN conversion to HD fibroblasts remains to be determined.
Our studies so far indicate that the MSN-specific neuronal conversion is highly efficient in fibroblast samples from symptomatic HD patients (HD-MSNs). Interestingly, HD-MSNs displayed protein aggregation of mutant HTT (mHTT), increased DNA damage, and spontaneous cell death in comparison to age- and gender-matched control MSNs. In addition, we found that the age of patient fibroblasts is an important component of recapitulating the mHTT aggregation in HD-MSNs. If we experimentally “rejuvenated” HD-fibroblasts to an embryonic stage (through deriving iPSCs and re-deriving fibroblasts) and conducted neuronal reprogramming, these MSNs showed drastically reduced HTT aggregation and this is largely due to the differential level of proteolysis inherent in adult and embryonic cells. We also found that spontaneous cell death in HD-MSNs occured due to mHTT-induced DNA damage as the cell-death phenotype could be intervened by an ATM inhibitor. Lastly, we found neuronal subtype-dependency of HD phenotypes as cortical neurons generated from HD fibroblasts (HD-CNs) showed differential vulnerabilities to HD phenotypes. Our results overall suggest the potential of directly converted neurons as a cellular model to study HD.