Skip to content

Using Stem Cells to Model Psychiatric Disorders: NIH Grantees Meeting

Meeting Summary
Bethesda, Maryland

Sponsored by:
National Institute of Mental Health (NIMH)

In October 2010, the NIMH convened a meeting of NIH grantees who use stem cells, particularly those derived from patients, as tools to identify fundamental molecular and cellular defects and the role of altered developmental processes in psychiatric disorders. Invited participants represented 17 research groups funded by NIMH, the National Institute of Neurological Disorders and Stroke (NINDS) and the Eunice Kennedy Shriver National Institute on Child Health and Human Development (NICHD), primarily through the American Recovery and Reinvestment Act of 2009 funding or other 2009 initiatives. In his welcome remarks, NIMH Director Thomas Insel, M.D., noted that this stem cell effort could build upon the successes of the previous three years in identifying genes associated with schizophrenia, autism spectrum disorder (ASD) and other disorders. However, progress would require the cooperative efforts of a “village” of interdisciplinary scientists and clinicians, such as those attending this meeting.

The introductory presentation noted that the relative utility of cell-based tools for studying psychiatric disorders, including human embryonic stem cells (hESC), induced pluripotent stem cells (h-iPSC), olfactory epithelium, and induced neuronal cells (iNC), depends on many variables. These variables include the appropriateness of patient selection to reflect common pathophysiology, the efficiency of cell line generation and validation, the faithfulness with which cells mimic in vivo developmental processes, the relevance of experimental assays to specific diseases, and the adaptation of these assays for target/therapeutic discovery and screening. Realizing the full potential of these tools and technologies for accelerating discoveries in psychiatric disorders will also require sufficiently powered sample sizes and appropriate linkage with associated clinical, phenotypic and genotypic data.

Patient Selection Criteria: First Steps in Reverse Translation
The first session focused on how classification of psychiatric disorders affects efforts to develop a meaningful molecular/cellular/developmental assay for schizophrenia, bipolar disorder, or ASD. It was noted that these disorders are heterogeneous and that diagnostic classification remains observational rather than based on underlying pathophysiology. The symptoms of schizophrenia, for example, overlap with a number of other disorders, making for diagnosis based on exclusion rather than inclusion. Since minimizing patient heterogeneity is a critical first step in developing new assays, Diagnostic and Statistical Manual (DSM) diagnosis was viewed as an uncertain criterion for grouping samples together for molecular and cellular analysis.

Participants agreed that most traction would lie with selecting patients with specific rare variants of large effect, whereas idiopathic cases would be much more difficult to study. This issue was also linked to statistical power; disorders that could be linked to rare variants of large effect would require fewer patients to be sufficiently powered, whereas genetically heterogeneous cases could require hundreds, if not thousands, of samples to achieve meaningful results. On the other hand, some participants noted cases where non-genetic parameters (e.g., head size in ASD, lithium response in bipolar disorder) could reflect common mechanistic pathways affected by multiple gene variants of small effect and thus serve as useful patient selection criteria.

iPSC Derivation and Validation Strategies
The second session addressed ongoing efforts to generate the best cellular tools, particularly iPSC from human subjects. A vigorous dialog centered on the advantages and disadvantages of using viral integration versus episomal-based or other methods of reprogramming. While many participants argued that the potential for cell transformation was a strong disadvantage of integration, it was noted that the ability to lineage-mark cells was crucial to studying distinct cell types in heterogeneous populations, either for modeling disorders or where cell replacement therapy was a goal. The use of excisable constructs was discussed as a way of combining the high efficiency of integrating viruses while limiting the potential for transformation. Other promising methods such as protein and RNA-based methods were discussed. It was agreed that increasing derivation efficiency while retaining the genomic integrity of the cells remained a major goal.

Participants also discussed validation methods at length, including methods to track cells through the reprogramming process to select only those clones that are fully reprogrammed. Methods ranged from immunocytochemistry to fluorescent reporter systems and comprehensive genetic, epigenetic and transcriptional analysis. Differences in the genomic stability of iPSCs and ESCs were identified, although there was not agreement as to whether these differences are systematic. Molecular phenotyping data suggests that the pluripotency “signature” is actually a spectrum of phenotypes, albeit a spectrum that has defined boundaries.  One concern was whether teratoma assays were necessary as an assay for pluripotency and reconstitution, given the substantial requirement of time, money, and animal resources. While it was noted that alternative assays could validate pluripotency and reconstitution at a lower cost, other participants argued that more detailed developmental and organotypic analysis of teratomas could be highly informative, but were rarely done. Although most of the discussion focused on exploration and optimization, it was noted that an important long term goal is standardization: finding the right balance between validation methods that are comprehensive, expensive, and time-consuming versus those methods that can be used broadly but provide limited information.

Assays with Relevance to Disease
The next sessions dealt with perhaps an even greater challenge: to develop valid molecular, cellular and developmental assays that relate meaningfully to disease pathophysiology. Most participants agreed that, for psychiatric illnesses, this is a particularly challenging task due to the absence of an obvious lesion in patients. Most presentations centered on efforts to generate neural derivatives efficiently from human ESCs or iPSCs, which is fundamental to identifying CNS differences between patients and control subjects. From this foundation, investigators discussed a number of different methodologies planned or in progress, including: (1) enrichment of specific cell subtypes using varied culture conditions and/or cell sorting; (2) use of aggregate culture paradigms to reproduce three-dimensional tissue structure; (3) moderate- to high-content phenomic analysis (e.g., transcriptome, proteome) for validation and target identification; (4) electrophysiological analysis of differentiated neurons; (5) analysis of other cell functions (e.g., calcium dynamics, signal transduction pathways); (6) reporter systems for cell subtype and lineage tracing; (7) cell engraftment for chimera analysis, coupled with optogenetics, functional magnetic resonance imaging or other functional analysis; (8) genetic or pharmacological rescue experiments to establish causality; (9) comparison of iPSCs from humans and mice with similar human alleles (e.g., copy number variations), coupled with multi-modal, in vivo analysis of these mice; and, (10) systems for high throughput screening (e.g., automated morphometry, cell signaling reporters).

The participants agreed that while methods for generating and assaying enriched populations of cells were improving, substantial exploratory work and validation remains to be done in order to optimize assays that meaningfully and specifically address the basis of psychiatric disorders. In the future, a particular combination of functional assays and phenomics could eventually strike the right balance between maximal information and ease-of-use to be a “gold standard” across many labs in the field.

Resource & Data Sharing, Future Priorities
A major long-term goal, particularly for NIH, is to make standardized reagents, methods, and data widely available to ensure sufficiently powered datasets and facilitate improvement in technologies and assays. For this reason, participants were encouraged to think together about standardization as a goal, particularly as the field matures. An overview of NIH efforts in data sharing and bioinformatics was presented to encourage thinking about how results in stem cell research could be broken down by data types and data definitions. Michael Huerta, PhD, the Director of the NIH National Database for Autism Research  (NDAR), described how NDAR is an example of how the universe of data sharing could be expanded to incorporate new scientific disciplines as they develop, such as stem cell modeling of psychiatric disorders. The key to success in such an endeavor is to engage the logistical expertise of NIH and the cutting edge experimental expertise of investigators to build a bioinformatics structure that will be accepted by and useful to the entire community.

Resource sharing was viewed as critical to provide standardized reagents and build sufficiently powered sample sizes; because iPSC derivation techniques are still being optimized, most participants thought that the focus should be on banking source cells such as fibroblasts. It was suggested that this function be linked comprehensively to other data on the relevant patient populations (e.g., genetics, diagnostic parameters). However, a poll of participants showed that procedures vary by institution, and barriers to centralized banking of reagents, such as fibroblast cell lines, still remained. Since informed consent procedures are authorized by each institutional review board, it was suggested that a template consent form may be helpful as way to standardize the process. Additionally, participants wanted to enable investigators to share cells readily, without repositories exerting a proprietary hold over the banked reagents.

The meeting concluded with a discussion of the role that public-private partnerships can play in promoting the translational and clinical utility of this research. In particular, it was noted that assays—adaptable to medium and high throughput screening—that have flexibility for multiple disorders, while also having the potential for measuring disease-specific outcomes, will be of great value. The vigorous day-long discussion showed that while iPSC research is still a young, conceptually risky, and rapidly evolving technology, stem cells promise to become standard and integral tools in the psychiatric research toolbox.

For more information, please contact David M. Panchision, Ph.D. at