Skip to content

Using Stem Cells for Biological and Therapeutics Discovery in Mental Illness


Location: Bethesda, Maryland

Convened by:
National Institute of Mental Health (NIMH)
Foundation for the National Institutes of Health (FNIH)

Sponsored by:
Janssen Pharmaceuticals
Eli Lilly and Company
One Mind for Research

On April 24-25, 2012, scientists from academia, government and the private sector met to discuss advances and challenges in using patient-derived reprogrammed cells, such as induced pluripotent stem cells (iPSCs) and induced neuronal cells (iNCs), to identify molecular, cellular, and developmental alterations relevant to mental illnesses and possible utility in diagnostics and drug development. The meeting opened with the recognition that the heterogeneity of mental illnesses, the uncertain etiology, and the recently publicized failures of pre-clinical data replication necessitated a cultural shift in the way science is done, from academic ‘silos’ toward a more collaborative effort. With the increasing popularity of iPSCs, it is critical to quickly establish rigorous, transparent, and reproducible methods to avoid a proliferation of poorly designed studies.

Patient Selection Criteria: Clinical Phenotyping, Genetics

A critical first step is determining the criteria used for stratifying patients for cell-based analysis. Many clinical features are shared between schizophrenia, bipolar disorder, autism spectrum disorder (ASD) and other illnesses, requiring that diagnosis be made by exclusion. However, patient drug response, biomarkers, imaging abnormalities, risk factors, and heritable traits may be used to select patients. Two opinions emerged at the meeting.  One side argued that the best approach would be to focus on highly penetrant single gene disorders, such as Fragile X and Rett Syndromes, along with copy number variants, with a focus on early-onset cases. Others argued that resolving heterogeneity for some diseases will require large sample sizes and an approach that targets treatment response and common molecular or developmental pathways, rather than specific genes. However, all agreed that studies should include comprehensive genomic characterization and clinical assessment of patients.

Streamlining Cell Reprogramming and Validation for Larger Scale Studies

Great progress has been made in optimizing reprogramming methods, including selection of source material. While fibroblasts are well-established and expandable source cells, patients may not consent to the invasive skin biopsy procedure, and evidence was presented that somatic mutations occur during the cell expansion process. Alternatively, blood that has already been banked could be a potential supply of source cells. Blood cells are easily obtained with less potential heterogeneity than fibroblasts. Their disadvantage is the lack of expandability and a limited track record for reprogramming. Some source material has been found to be genetically mosaic, yielding reprogrammed cells with substantially different properties, though this can be exploited when the mosaicism relates causally to the disorders.

A trend toward larger studies will require development of current good manufacturing practice (cGMP) by eliminating variation in the iPSC derivation and validation process. Many groups have optimized non-integrating reprogramming methods and utilize fully defined growth medium. Reprogramming technologies that bypass the pluripotent state include iNCs and a new protocol for induced neural progenitor cells (iNPCs). Several speakers commented that conventional validation assays based on PCR, immunocytochemistry, germ layer generation, and teratoma assays are being replaced by genome-wide analysis. One “big dataset” approach to phenotypic comparison of iPSC lines is the Stem Cell Matrix, which provides a collection of 8000 samples consisting of fetal/adult tissues and cell lines, and that can be augmented by datasets from other labs to avoid sampling bias. PluriTest was proposed to replace older methods of pluripotency validation, such as teratoma assays, which are time/resource-intensive and provide poor discrimination of line-to-line variation.

Getting the “Right” Cells and Validating the Pathophysiological Relevance of Assays

A major pitfall of in vitro studies is the assumption that limited characterization of differentiated cell types is sufficient for researchers to validate that they are working with a cell type as it exists in vivo. While meeting participants agreed that neurons with a glutamatergic phenotype could be readily generated from iPSCs, they could express a variety of markers, suggesting a poorly defined regional identity. Thus, conditions must be optimized for the regional specification and maturation state appropriate to the disorder being studied. An alternative is to systematically define the neuron phenotype using gene reporters as a complement to cell surface marker-based selection. Site-directed integration of transgenes into iPSCs reduces variability between lines, thus allowing for a more robust comparison of control versus disease states. Many workshop participants advocated for the generation of a broad repertoire of fluorescent reporter lines for regional or cell type specification to effectively purify neuron subtypes.

A high-risk assumption of patient-derived cell reprogramming studies is that causal abnormalities can be resolved using assays that reconstitute a molecular or cellular process that is predictive of systems-level pathophysiology. Studies of Timothy Syndrome (TS) iPSCs found that, compared to controls, TS neurons assumed upper cortical layer fates at the expense of lower cortical layer fates, as measured by regional identity genes.1 Similar results were found in mice with the CACNA1C calcium channel mutation that causes TS. Another promising approach is a three-dimensional culture paradigm, involving self-organizing cell aggregates that recapitulate features of the developing forebrain.2 This assay system may provide a more physiologically relevant system to assay aberrant developmental processes.

Important questions remain as to how easily iPSC- and iNC-based results can be validated against other assays of pathophysiology. Comparisons against human gene variants in mouse are one such opportunity. Human post-mortem analysis has the advantage of utilizing actual patient tissue, but is disadvantaged by its late stage in the disease trajectory and by variable tissue quality; nonetheless, these assays could be used as partial validation of pathophysiological relevance. Clinical measures such as behavioral assessment, drug response, or imaging could be compared with cell data, particularly where such measurements provide mechanistic insight (e.g., myelin tracts). Engraftment assays are valuable for studying human cells in the context of circuits, but proper integration of human cells in mouse remains a methodological hurdle. Some attempts have been limited by incomplete maturation of human neurons into the mouse cortex, while preliminary success was reported in the hippocampus.

Molecular Discovery, Target Identification, Screening

While acknowledging the challenges, workshop participants described efforts they have begun to identify novel molecular targets (e.g., by transcriptional and phosphoproteomic profiling) and adapt cellular assays for high throughput screening. A key argument in favor of iPSC and iNC-based screening assays, relative to previous drug pipeline assays, is that they utilize human neurons/glia and thus recapitulate gene expression, signaling isoforms and cellular processes that may be species-, and tissue-specific. Furthermore, there is a shortage of good cellular phenotypes for mental illness pathophysiology; approaches that utilize unbiased profiling of thoroughly validated cell types may yield novel hypothesis-generating datasets that can be used to generate screening assays. One such platform is the NIH Library of Integrated Network-based Cellular Signatures (LINCS) project, which generates gene signatures through high throughput screening of cellular responses to various perturbogens. Patients with especially strong drug responsiveness may provide a unique opportunity to generate a robust cellular signature of disease-relevant pathways. Other participants cautioned that the utility of any screening assay for psychiatric disorders depends critically on further validation and that high throughput primary screens (e.g., signaling, synaptic vesicle cycling, or neurotransmission) must include secondary circuit-based validation assays involving human cell engraftment and/or human disease-associated gene alleles in mouse.

Coordinating Resources and Future Efforts

The meeting included a brainstorming session to address the question: “How can we move more rapidly towards new patient treatments through the application of cell reprogramming technologies?” Much discussion focused on a need for greater collaboration and interdisciplinary training, particularly in skills that spanned academic and industry research. Suggestions to facilitate technology were mainly in the context of generating standardized, scalable and shared resources such as less costly, more efficient cell differentiation, gene targeting and reporter techniques. All agreed that cellular analysis should be sufficiently powered and integrated with whole genome sequencing and careful clinical phenotyping of patients, incorporating measures that can validate cellular phenotypes. While some advocated for a new large-scale collection of iPSCs for different psychiatric diseases, others confirmed that a number of these studies were ongoing, and a collaborative culture would begin with transparent sharing of cell lines and protocols to contain costs and enhance replicability. Centralized collection of biomaterials and re-consenting for more comprehensive genetic analysis and clinical phenotyping, where needed, could provide leverage for more collaborative future efforts. Similar large public-private collaborations, such as the Innovative Medicines Initiative in Europe, are underway and provide contingencies for international coordination; several participants advocated for joining efforts across borders.

Efforts are underway at NIH to enhance research through the standardization and sharing of resources. In 2011, NIMH established the new Stem Cell Resource at the NIMH Center for Collaborative Genomic Studies of Mental Disorders, to make source cells and iPSCs more widely available. The genetic component of this resource already houses over 100,000 patient samples of blood, lymphoblastoid cell lines, and DNA, along with genetic and clinical phenotyping data. The NIH Center for Regenerative Medicine (CRM) pilots new technologies such as optimizing iPSC and cell screening methods, generating reference and reporter cell lines, harmonizing informed consent language, and working with reagent vendors to enhance freedom to operate. CRM coordinates with the National Center for Advancing Translational Sciences to integrate iPSC technology into the therapeutic pipeline. The NIH Molecular Libraries Program provides access to 350,000 small molecules for qualified screening projects by NIH-funded investigators; this is already being utilized in iPSC-based screening projects to identify novel probes relevant to psychiatric disorders. The meeting concluded with general agreement that this discussion should serve as a foundation for collaborative relationships among public and private stakeholders to promote the translational and clinical utility of reprogrammed cell research.

For more information, please contact David M. Panchision, PhD at

 1 Pasca, S.P. et al. (2011). Using iPSC-derived neurons to uncover cellular phenotypes associated with Timothy syndrome. Nat Med 17, 1657-62.

 2 Mariani, J. et al. (2012). Modeling human cortical development in vitro using induced pluripotent stem cells. Proc Natl Acad Sci U S A 109, 12770-5.