Advances in Disease Modeling for ALS and FTD Workshop: Executive Summary

Satellite Symposium of the 2010 Society for Neuroscience Meeting
Manchester Grand Hyatt Hotel, San Diego
November 12, 2010
Workshop   Agenda   Participant List

Introduction
Recent advances in defining the genetics and pathology of Amyotrophic Lateral Sclerosis (ALS) and Frontotemporal Dementia (FTD) have provided the opportunity to develop new animal models to study disease pathophysiology and test therapeutic strategies. Investigators have utilized mutations in genes encoding TAR DNA-binding protein (TARDBP; TDP-43) and Fused in Sarcoma/Translocated in Liposarcoma (FUS/TLS) to develop disease models in a spectrum of diverse organisms including yeast, roundworm, fly, zebrafish and rodents.

In light of these advances, The ALS Association (ALSA) and the National Institute of Neurological Disorders and Stroke (NINDS) convened a workshop that focused on emerging models for ALS and frontotemporal lobar degeneration with ubiquitinated inclusions (FTLD-U), the most common form of FTD. Organizers of the workshop aimed to review the technical approaches used to develop new ALS and FTLD-U animal models, exchange information about their pathological, molecular and behavioral phenotypes, and discuss strategies to facilitate their broad availability and appropriate use in translational research. Invited participants included investigators who are currently developing and characterizing new animal models for the above-mentioned diseases. Registration for the workshop was open to the broader research community to attend as general participants.

Introductory Presentations
Five presenters summarized the field’s current understanding of TDP-43 proteinopathies, the genetics and molecular neuropathology of ALS and FTD, as well as TDP-43 protein function.

Dr. Virginia Lee, University of Pennsylvania School of Medicine, Philadelphia, provided an overview of misfolded protein inclusions in neurodegenerative diseases, such as aggregates of beta amyloid in senile plaques, tau in neurofibrillary tangles, a-synuclein in Lewy bodies, polypeptides with expanded polyglutamine (polyQ) domains in inclusion bodies, TDP-43 in cytoplasmic inclusions and prion protein in plaques. In addition, she reviewed the clinical features, genetics and pathological hallmarks of ALS and different forms of FTD. Recent advances in ALS/FTD research support the concept that ALS and FTLD-U are syndromic variants of a clinicopathological spectrum of the same neurodegenerative disease. Furthermore, it is now evident that TDP-43 inclusions represent a new type of proteinopathy.

Dr. Robert Brown, University of Massachusetts Medical School, Worcester, updated attendees on the search for gene defects that cause familial and sporadic ALS (FALS and SALS), familial FTD and FALS-FTD. He reviewed the growing list of genes and chromosomal loci that have been linked to these diseases, and noted that these gene mutations implicate defects in multiple cellular pathways culminating in neuronal death that include axonal transport, vesicle trafficking, RNA processing and susceptibility to environmental toxins. In this context, he emphasized that there is a remarkable convergence in the pathophysiologies of SALS, FALS and FALS-FTD.

Dr. Ian Mackenzie, University of British Columbia, Vancouver, discussed the molecular neuropathologies of genetically distinct subtypes of ALS and FTD. TDP-43 pathology is found in FTLD-U and most forms of ALS but not in SOD1-linked FALS. He noted that cell and animal models may not fully recapitulate the age-dependent phenotypes of adult-onset neurodegenerative diseases as a general caveat. Furthermore, disease susceptibility may differ between animals and humans. He concluded that these complexities warrant the use of multiple complementary model systems.

Drs. Don Cleveland and Magdalini Polymenidou, University of California, San Diego, discussed the physiological roles of TDP-43. TDP-43 is involved in multiple steps of RNA metabolism, including transcription, splicing or transport of mRNA, as well as microRNA metabolism. The presenters described current data demonstrating that TDP-43 regulates the levels or splicing of more than 900 RNA transcripts. TDP-43 binds to UG nucleotide repeats in distal introns, thereby regulating the levels of long pre-mRNA transcripts. Furthermore, TDP-43 auto-regulates its synthesis in part by binding to the 3’ untranslated region of its own transcript. It is plausible that a feed-forward auto-regulatory mechanism leading to increased TDP-43 levels drives disease progression.

Highlights from Presentations on New Model Systems
The workshop proceeded with a session on non-mammalian models of TDP-43- and FUS/TLS-dependent cellular toxicity. Dr. Aaron Gitler, University of Pennsylvania School of Medicine, Philadelphia, opened the session with a presentation describing the identification of genetic modifiers of TDP-43 toxicity in yeast and Drosophila. These studies revealed that intermediate-length polyQ expansions in Ataxin 2 are a risk factor for ALS. He noted that ongoing research focuses on the characterization of additional genetic modifiers of TDP-43 toxicity and the possible involvement of other RNA-binding proteins in neurodegeneration. Dr. David Morton, Oregon Health and Science University, Portland, described studies of TDP-43 function in Drosophila larvae, where depletion and over-expression of TBPH, the fly ortholog of TDP-43, result in distinct defects in locomotion and changes in neuromuscular junction activity. This suggests that TBPH loss-of-function and over-expression have different cellular consequences. Dr. Jane Wu, Northwestern Medical School, Chicago, Randal Tibbetts, University of Wisconsin, Madison, and Fen-Biao Gao, University of Massachusetts, Worcester, reviewed models of TDP-43 proteinopathy in adult Drosophila. Over-expression of wild-type and mutant human TDP-43 (hTDP-43) in flies recapitulates neuropathological and clinical features of TDP-43 proteinopathy. Dr. Tibbetts reported that the Notch pathway is up-regulated in hTDP-43 transgenic flies, and noted that gene expression signatures from these fly models may yield further insights into TDP-43-dependent neurodegeneration. Dr. Gao also described efforts to generate and characterize TDP-43 mutation-specific induced pluripotent stem cells.

Dr. George Jackson, University of Texas Medical Branch, Galveston, discussed the suitability of different tissue-specific expression drivers for modeling TDP-43 proteinopathy in Drosophila. He also reported that expression of hTDP-43 in wings yields robust phenotypes of aberrant wing compartmentalization and veination. These wing phenotypes might be a more sensitive phenotypic read-out for modifier screens than eye phenotypes.

Dr. Christopher Link, University of Colorado, described a Caenorhabditis elegans (C. elegans) model of TDP-43 toxicity. Pan-neuronal expression of hTDP-43 impairs the swimming ability of C. elegans and, at the cellular level, alters motor neuron synapses. Deletion of TDP-1, the C. elegans ortholog of TDP-43, leads to the formation of vacuoles, which is suggestive of excitotoxic neuronal death. Dr. Pierre Drapeau, University of Montreal, reviewed TDP-43- and FUS/TLS-based models in Danio rerio (zebrafish). Mutant zebrafish lines exhibit motor deficits. A series of rescue experiments suggests that TDP-43 and FUS act in the same pathway. Dr. Drapeau also highlighted general advantages of zebrafish as a model system for motor neuron disease including fast response times and the presence of single motor neuron axons that are quantifiable.

The afternoon session of the workshop focused on rodent models of TDP-43- and FUS/TLS-dependent neurodegeneration. Dr. Lucie Bruijn, ALSA, introduced the session by highlighting the research advances attained by modeling ALS in SOD1-mutant rodents. Even as new, genetically distinct, in vivo and in vitro models emerge, mutant SOD1-based models remain an important tool in ALS research.

Dr. Jada Lewis, University of Florida, Gainesville, reviewed the behavioral and pathological phenotypes of transgenic mouse models that over-express hTDP-43. Constitutive over-expression of hTDP-43 yields phenotypic features that are reminiscent of ALS and/or FTLD-U, although none of these models recapitulate the human disease process fully. To understand the impact of hTDP-43 over-expression on normal development, her team generated an inducible model. Induction of hTDP-43 during development results in gross brain atrophy, while induction after weaning causes progressive neurodegeneration and produces TDP-43 aggregates similar to the ones in the human condition.

Dr. Brian Kraemer, University of Washington, Seattle, discussed the phenotypes of homozygous and heterozygous TDP-43 knock-out mice. Complete loss of TDP-43 causes lethality in early embryogenesis. Heterozygous knock-out mice exhibit age-related phenotypes including increased weakness; however, the heterozygous mice show no obvious signs of neurodegeneration and have normal muscle morphology.

Dr. Philippe Parone, University of California, San Diego, summarized ongoing efforts by Dr. Don Cleveland’s research team to generate mouse models of TDP-43- or FUS/TLS-mediated neurodegeneration. The technical approaches used by this group include the expression of prion promoter-driven transgenes and bacterial artificial chromosome clones of monogenic human genomic fragments. Both wild-type and disease-mutant forms of hTDP-43 and hFUS/TLS are being tested. Dr. Parone reported that mice expressing genomic hTDP-43 develop gait abnormalities in adulthood. Both wild-type and mutant forms of hTDP-43 cause this phenotype. Mice expressing prion promoter-driven hTDP-43 exhibit progressive loss of motor function. This phenotype is restricted to mutant hTDP-43 and not observed when wild-type hTDP-43 is being expressed. The behavioral and histological analyses of FUS/TLS transgenic mice are still in the early stages.

Dr. Zuoshang Xu, University of Massachusetts Medical School, Worcester, described a new transgenic mouse model of FUS/TLS over-expression. This model exhibits early-onset hyperactivity. Furthermore, he discussed a general caveat of the Cre-loxP recombination technology. Although this technology is a powerful experimental tool for cell type-specific genetic manipulations, loxP sites that are inverted relative to one another can result in unexpected phenotypes (e.g., microcephaly), and this can occur if multiple transgenes integrate into the genome in random orientation. Dr. Xu emphasized that such non-specific effects must be taken into consideration when designing and interpreting mouse model studies that use the Cre-loxP recombination technology.

Dr. Abraham Acevedo, Medical Research Council, Harwell, UK, provided an update on mutant TDP-43 and FUS/TLS mouse lines of the Harwell N-ethyl-NB-nitrosourea (ENU) Chemical Mutagenesis Program. The Program conducts ENU-based mutagenesis and performs phenotype- and genotype-driven screens to study gene function. The facility houses a large archive of sperm and DNA samples from ENU-mutagenized mice that can be used for gene-driven screens. The archive includes multiple TDP-43 and FUS/TLS point mutants that were bred and are currently being characterized. Among other point mutations, a nonsense mutation in TDP-43 was found (Q101X). Homozygous Q101X mutants are not viable, while heterozygotes and compound heterozygotes with another point mutation exhibit hind-limb grasping and reduced body tone. Other behavioral measures including Rotarod and grip-strength remain unchanged over 1.5 years. FUS/TLS-mutant lines are currently being characterized. So far (at 7 months of age) they do not show an overt phenotype. The Harwell Mutagenesis Program has also generated a Sod1D83G-mutant line. These mice, currently under longitudinal characterization, are smaller, weaker and present motor neuron degeneration. ENU-mutant mice in other ALS-related genes such as senataxin and progranulin have also been generated and are freely available to collaborators (http://www.mousebook.org/catalog.php?catalog=mnd).

Dr. Robert Baloh, Washington University, Saint Louis, concluded the session by describing transgenic mice that express hTDP-43 containing a familial ALS mutation under the control of the mouse prion protein promoter (Prp-TDP-43A315T), the first published TDP-43-based mouse model. The mice develop features that are reminiscent of ALS and FTLD-U, including gait abnormalities, motor axon loss and ubiquitin pathology in the spinal cord and frontal cortex. Of note, most ubiquitin-positive inclusions in these mice do not stain for TDP-43. In addition, the Prp-TDP-43A315T mice exhibit sexual dimorphism. Both genders develop disease but females live about 30 days longer than males. This model is currently being bred a Jackson Laboratory (JAX) using a speed congenic approach to standardize the background. The two sites use different breeding strategies and so far, there are some differences in the observed survival times for females depending on whether they have been subject to mating.

Dr. Baloh also noted that there is a strong dose-dependence in TDP-43 over-expressing models yielding a steep dose-response curve between phenotype and transgene dose. This needs to be taken into consideration when interpreting phenotypic data from these models.

Perspective of JAX
Dr. Cathleen Lutz, JAX, Bar Harbor, provided an overview of the JAX Centralized Husbandry and Distribution Core for mouse models. JAX imports mouse models via rederivation, markets and assesses community interest, builds colonies, provides quality-controlled maintenance, validates the phenotypes and serves as a distribution core. 1,200 lines are currently being maintained as live colonies and 5,000 lines are archived under cryopreservation. Routine services include allele-specific genotyping, copy number assessment, phenotypic assessment, stabilization of the genetic background and protection against accidental loss or contamination by archiving cryopreserved lines.

JAX has entered partnerships with multiple advocacy groups to standardize and facilitate the distribution of disease-specific mouse models, including models for Spinal Muscular Atrophy and Parkinson’s Disease. Recently, three advocacy groups in the ALS field, ALSA, the ALS Therapy Alliance (ATA) and the Tow Foundation, have come together to support a National ALS Mouse Model Repository at JAX. Goals are to standardize new ALS models and to provide genetic and phenotypic quality control while working closely with the ALS research community as the models are being established. JAX also hopes to develop best practice guidelines for the use of these models. Dr. Lutz highlighted general benefits of depositing models early and noted that, if requested, data will be kept confidential until publication. JAX also offers the opportunity to produce embryonic stem cells from mouse models. The Prp-TDP-43A315T mice produced by Dr. Robert Baloh’s team are already available through JAX and are currently being backcrossed.

Summary of Panel Discussion
A panel of experts discussed best practices related to in vivo modeling of ALS and FTLD-U. Dr. Leonard Petrucelli, Mayo Clinic, Jacksonville, opened the discussion by asking which behavioral phenotypes should be characterized. While motor phenotypes can be readily assessed in rodent models, it is more challenging to characterize cognitive phenotypes. The frontal cortex of mice is proportionally much smaller than in humans, and it is therefore difficult to model executive dysfunction in this and lower species. It was suggested that the research community should take an “agnostic” approach as the new models emerge and characterize their behavior as fully as possible. Some contract research organizations and large repositories provide comprehensive phenotyping services that could be taken advantage of. Regarding the pathology, neuronal and non-neuronal cells of the spinal cord and cortex need to be analyzed as well as the cortical neurocircuitry. It was also noted that pathology per se will likely not suffice as a phenotypic outcome measure since preclinical drug discovery usually relies on behavioral phenotypes as primary outcome measures.

The panelists also discussed how to decide which mouse models should be deposited in public repositories. Dr. Lutz commented that although JAX does not want to maintain redundant mouse models as live colonies, it is worthwhile and cost-effective to archive “back-up” lines in cryopreserved form while higher priority lines are being characterized. In this context, it was noted that it would be useful if a web forum were created where investigators could share details about their mouse models that were not published but would help standardize the use of the models across research laboratories. Dr. Lutz noted that JAX maintains lines on a congenic background and checks transgene copy numbers to avoid genetic segregation. In addition, only well characterized mouse models should be broadly distributed.

Paul Taylor, St. Jude Children’s Research Hospital, Memphis, discussed the value of non-mammalian animal models as tools for ALS and FTD research. Species like yeast, roundworm, zebrafish and fly are amenable to genetic manipulations since generation time and expense are low. Both genetic and chemical screens are feasible. Fundamental aspects of gene function are often conserved, and lower species can therefore be used to study the basic physiological function of disease genes and to identify genetic interactions in an unbiased way. Possible caveats include the uncertainty of whether the temporal features of late-onset human diseases can be recapitulated in the short life spans of simpler species. The panelists agreed that this is an important issue but noted that at least in Drosophila methods have been developed to compensate for these temporal effects. Other caveats include limitations in modeling cell type-specific aspects of human diseases in simpler species. Based on these considerations, the panelists concluded that these models represent a powerful starting point to study human disease but that more complex phenotypes must be modeled in higher species.

Dr. Steve Perrin, ALS Therapy Development Institute, Cambridge, led the discussion on best practice guidelines for preclinical trial design. He noted that the utility of the SOD1G93A mouse model is well established, and that the phenotype of the model has been relatively stable over the past 15 years. Experience gained from using this model has shown that it is imperative to develop standards and ways to control biological variables in preclinical study design. At this point, the new ALS/FTD animal models still need to be better characterized before they can be used in therapy development. In addition, it is important to recognize that not all animal models that are currently being developed will be appropriate for translational research. In general, only some models exhibit phenotypes that are practical for therapy development. Given the progressive nature of ALS, one also needs to ensure that the way the animals are being monitored doesn’t, in fact, exacerbate disease.

Elizabeth Fisher, University College London, UK, introduced the topic of large mammalian animal models for ALS and FTD research. Although mouse models are important research tools, drawbacks include the fact that the neuroanatomy and neurophysiology in mice are not fully understood. Furthermore, there are significant limitations in modeling complex behaviors in mice. Non-human primates serve as a model species in other disease areas including spinal cord injury. Despite the obvious advantages of modeling human diseases in more closely related species, the use of non-human primates and other large mammals as model species is associated with significant ethical and practical issues. Panelists also discussed the possibility of using canine models for ALS research. Two naturally occurring diseases with ALS-like features are currently being characterized in dogs. Despite the promise of these canine diseases as new models for ALS, the cost of running preclinical trials in dogs is high. The panelists concluded that the generation of ALS/FTD models in non-human primates, and other large mammals, faces significant ethical and economic challenges that need to be taken into consideration before developing such models.

Concluding Goals
Amelie Gubitz, NINDS, summarized next steps as identified by the workshop participants. These include:

• Develop a consensus/priority list regarding which new mouse models should be transferred to JAX to begin building the new ALSA/ATA/Tow Foundation-supported National ALS Mouse Model Repository. It is expected that this will be a dynamic, research community-driven process.
• Develop recommendations regarding the phenotyping of the new models. JAX will validate the new repository mouse models according to such recommendations.
• Once models with homogeneous and “workable” phenotypes are available, formulate guidelines for the optimal use of these models in preclinical research, similar to the guidelines for SOD1-linked ALS models
• Continue to analyze the strengths and limitations of the existing and emerging models. It is expected that many of these models will serve as complementary tools for ALS/FTLD-U research.