13 April 2011. Neurons have been successfully grown from induced pluripotent stem cells (iPSCs) derived from people with schizophrenia, according to a study in Nature appearing online 13 April 2011. Fred Gage of the Salk Institute in La Jolla, California, and colleagues report that, though these neurons resembled those from people without schizophrenia in many ways, they had deficits in forming connections with other neurons, and exhibited differences in gene expression—the researchers highlighted cAMP and Wnt pathway genes—when compared to control neurons. The connectivity and some gene expression aberrations could be "normalized" by treating the neurons with the antipsychotic loxapine.
The study adds schizophrenia to the few diseases that have so far been modeled using iPSCs derived from actual patients. With techniques to reprogram adult, readily obtainable tissue like skin cells into iPSCs, researchers can now try to generate cell types of interest from patient populations to better simulate at a cellular level what may be happening in a particular disease, even in a specific person. While recent studies of patient-specific iPSCs have focused on single gene disorders such as Rett’s syndrome (e.g., Marchetto et al., 2010), the new study marks one of the first attempts to study a disorder with heterogeneous genetic origins.
A brief report in Molecular Psychiatry in February described iPSCs derived from schizophrenia patients (Chiang et al., 2011), but the new study starts to get a handle on features potentially related to disease, comparing the connectivity patterns, synaptic markers, physiology, and gene expression profiles of neurons grown from iPSCs derived from control and schizophrenia subjects.
First author Kristen Brennand and colleagues started with fibroblast samples from four schizophrenia patients: one with childhood onset of the disorder, and three others with an affected parent. Control samples came from age- and ancestry-matched individuals with normal psychiatric evaluations. The fibroblasts were transformed into iPSCs using a lentivirus to introduce genes that reprogrammed the cells into a pluripotent state. The iPSCs were then differentiated into neural precursor cells, and then neurons, over the course of three months. Most turned out to express VGLUT, a marker of glutamatergic cells, about 30 percent expressed GABAergic neuron markers, and less than 10 percent were positive for tyrosine hydroxylase, an enzyme required to make dopamine.
Interconnected neurons derived from induced pluripotent stem cells (iPSCs) from schizophrenia patients. hiPSC neurons are shown expressing the neuronal proteins Beta-III-tubulin (red) and MAP2AB (green). Nuclei are stained with DAPI (blue). Magnification is 20x. Image credit: Kristen Brennand, Salk Institute for Biological Studies
When grown with astrocytes in the dish, the neurons formed connections with each other. Using a modified rabies virus to trace the number of direct inputs received by a given neuron (Wickersham et al., 2007), the researchers measured a decrease in connectivity, with schizophrenia-derived neurons receiving inputs from about half the number of neurons as controls did. Treatment with a variety of antipsychotic agents with affinity for both dopamine and serotonin receptors did not affect connectivity, with one exception: adding loxapine, which targets both dopamine and serotonin receptors about equally (Kapur et al., 1997), to the dish for three weeks boosted connectivity in the schizophrenia neurons.
The researchers also measured slightly fewer neurites, the processes destined to become dendrites or axons, in the schizophrenia hiPSC neurons compared to controls—something the authors compare to the reduced dendritic arborizations found in postmortem brain. The neurons from individuals with schizophrenia also had less staining for PSD95—a protein involved in anchoring proteins at glutamatergic synapses—than control neurons did.
These changes did not seem to compromise synaptic function, however. The researchers report that the schizophrenia hiPSC neurons exhibited normal action potentials, spontaneous excitatory and inhibitory synaptic activity, and spontaneous calcium signals. This overall picture of decreased connectivity with normal synaptic function runs counter to the synaptopathic view of schizophrenia and other disorders (Südhof, 2008) in which the number of synapses is postulated to remain normal, but synaptic function is compromised. The authors suggest that further analysis may, in fact, reveal some functional differences in the neurons derived from individuals with schizophrenia.
With gene expression microarrays, the researchers detected deviations in expression of 596 genes in the schizophrenia neurons that were at least 1.3 times greater or less than the level found in controls. Of these genes, 25 percent had been previously linked to schizophrenia, either through genetic association or postmortem studies. The authors write that gene ontology analysis of the altered expression highlighted glutamate receptor genes, and cAMP and Wnt pathway genes. Other schizophrenia-related genes, including NRG1 and ANK3, had significantly elevated expression in schizophrenia-derived neurons compared to controls. Interestingly, the NRG1 increase was detected only in neurons, and not in fibroblasts or iPSCs from the schizophrenia patients, which argues that it is critical to look at the cell type relevant to a disease. Further study with qPCR verified patterns of altered expression for these and other schizophrenia suspects, and loxapine treatment usually boosted expression of these genes.
However, patients varied in their patterns of gene expression, which may reflect differences in the underlying genetic component contributing to each individual's schizophrenia. To address this, the researchers analyzed copy number variations (CNVs)—losses or gains of segments of DNA—which have been reported to substantially increase risk of schizophrenia (Walsh et al., 2008). They found 42 genes affected by CNVs among their four patients, none of which occurred at regions where CNVs have been previously associated with schizophrenia. Strikingly, only 12 of the genes affected by CNVs showed changes in neuronal expression that correlated with whether a copy of a gene was lost or gained. This suggests that compensatory mechanisms could be at work in these neurons, and indicates that neurons grown from iPSCs may deliver a reality check for ideas gleaned from human CNV studies, which often spur animal models based on an observed deletion or duplication of a particular gene.
Some of the results echo reported pathophysiology in schizophrenia; for example, NRG1 expression was elevated in the neurons grown from iPSCs derived from schizophrenia patients, similar to the increased levels found in postmortem brain tissue (see SRF related news story). Other results suggest new avenues of research, such as finding altered expression in genes related to axon guidance and NOTCH signaling (interestingly, NOTCH4 currently has a positive meta-analysis in SZGene).
A stem cell watershed
Despite the heterogeneous genetic risk factors likely at work in this small patient sample, it is interesting that some consistent results—such as the decrease in connectivity—were obtained. In fact, the authors predict that a narrower, more consistent pattern of expression changes affecting a smaller number of genes will emerge as the number of individuals with schizophrenia studied with iPSCs increases. This is consistent with a "watershed" model that proposes that a vast variety of gene malfunctions could contribute to schizophrenia by converging on the same key biological pathways.
The study marks the beginning of an era of stem cell research of schizophrenia. Future work will refine the description of these neurons and delineate how drugs may change them, and researchers will have to grapple with the interpretation of any results coming from the schizophrenia-derived neurons that happen to resemble, or diverge from, alterations noted in the brains of people with schizophrenia.—Michele Solis.
Brennand KJ, Simone A, Jou J, Gelboin-Burkhart C, Tran N, Sangar S, Li Y, Mu Y, Chen G, Yu D, McCarthy S, Sebat J, Gage FH. Modelling schizophrenia using human induced pluripotent stem cells. Nature. 2011 April 13.