July 14, 2014. Stem cells derived from people with schizophrenia carrying 15q11.2 deletions reveal structural problems that could result in misplaced neurons in the brain. The study, published in Cell Stem Cell on July 3, implicates the gene CYFIP1 in these mishaps, which encodes a protein involved in shaping the cytoskeleton.
The fruit of a collaboration led by Guo-li Ming and Hongjun Song of Johns Hopkins University, Baltimore, Maryland, the study combines results from stem cells, mouse experiments, and human genetics to argue that deficits in CYFIP1 derail early brain development in a way that increases risk for schizophrenia. Specifically, neural progenitor cells carrying 15q11.2 deletions were confused about which way was up or down, and did not develop proper cell-cell junctions. In mice, knockdown of CYFIP1 led to migration errors, with newborn neurons going to the wrong cortical layers. In humans, the researchers found that single nucleotide polymorphisms (SNPs) tagging genes involved in cytoskeleton dynamics, including one for CYFIP1, interacted to increase risk for schizophrenia—evidence for epistatic effects.
CYFIP1 is one of four genes lost in 15q11.2 deletions, which increase risk for schizophrenia, autism, and intellectual disability (Malhotra and Sebat, 2012). Recent studies delving into its function show CYFIP1’s involvement in inhibiting protein translation and promoting remodeling of actin, the building block of a cell’s cytoskeleton (De Rubeis et al., 2013). A study from Josef Kittler's group at the Institute of Psychiatry in London, UK, published earlier this year, found that CYFIP1 influenced dendrite structure in rodent neurons in culture: Knocking it down resulted in neurons with sparser dendrites, fewer dendritic spines, and somewhat unmoored glutamate receptors (Pathania et al., 2014). Loss of one copy of CYFIP1 in mice resulted in similarly sparse dendrites in the hippocampus. These types of changes could contribute to disruptions in neural connections that have been suggested for schizophrenia.
The new study points to roles for CYFIP1 early in development, too, before a neuron sprouts its dendrites. Starting with induced pluripotent stem cells (iPSCs), the researchers could track the earliest stages of development as they morphed into neural progenitor cells, and later, cortical neurons. Though iPSCs have been made from cells obtained from people with schizophrenia before (e.g., see SRF related news report), the new study is one of the first to look at people with the same genetic lesion. Song and Ming’s group briefly reported in 2011 the establishment of iPSCs from skin samples collected from two people with schizophrenia sharing a deletion within disrupted-in-schizophrenia 1 (DISC1) (Chiang et al., 2011). It is challenging to assemble multiple people with the same rare genetic lesion, but having such a homogeneous subset could more precisely connect the dots between genes and brain development.
This way up
First author Ki-Jun Yoon and colleagues began by establishing iPSCs from skin cells from the three subjects with 15q11.2 deletions, using their family members as controls. In making neural progenitor cells (NPCs) from the iPSCs, the researchers did not observe any differences in proliferation or differentiation. But when they induced the cells to become cortical neurons, forming the flower-shaped assemblies of cells called rosettes, differences emerged: while control neurons were topped with PKCλ, a marker of the top, or apical, side of cells, neurons with 15q11.2 deletions were not, having instead scattered PKCλ labeling. Similarly, the structure of the contact points between the cells, called the adherens junctions, was also disorganized. These distortions were not found in NPCs made from people with the DISC1 deletions, however, indicating that it was specific to 15q11.2 deletions.
These structural disruptions could be blamed on CYFIP1. Increasing CYFIP1 in 15q11.2 deletion neural progenitor cells rescued the mislocalized expression of PKCλ. Likewise, knockdown of CYFIP1 in control cells resulted in the scattered pattern.
The pathway view
In mice, knockdown of CYFIP1 in the embryonic brain resulted in cells going to the wrong places. For example, radial glial cells, which spawn newborn neurons and glia, were found outside of their usual ventricular zone layer in the nascent brain. This spelled trouble for newborn neurons, which typically follow the long processes of radial glial cells on their way to the correct layer: neurons meant to reside in the deep cortical layers found their way to upper layers, while neurons destined for upper layers went to deep layers.
CYFIP1 interacted with the WAVE complex, a set of proteins that regulates actin polymerization—the putting together of actin segments to form the filaments that make up the skeleton of a cell. In fact, knocking down one of the members, Abi1, or two other genes downstream, ARP2 and ARP3, led to radial glial cells vacating their usual VZ site.
Overall, the findings implicate an actin-regulating pathway involving CYFIP1, WAVE complex proteins, and ARP2 and ARP3, in these developmental displacements. Taking such a pathway-based view, the researchers teamed with Daniel Weinberger's group at the Lieber Institute in Baltimore to test whether a handful of selected SNPs in different components of this pathway increased risk for schizophrenia in four case-control samples. Though no single SNP showed an association with schizophrenia, considering them in combination did: having one allele at a SNP tagging ACTR2, which forms a complex with ARP2 and ARP3, together with another allele at a SNP tagging CYFIP1(or nearby NIPA2) was associated with an increased chance of schizophrenia (OR=11.14).
This result argues that the insights gleaned from people with the rare 15q11.2 deletion may apply to others without the deletion, but who do have schizophrenia. Although finding ways to remedy cytoskeletal anomalies may be too late to fix the mislaid neurons from early development, it may be possible to facilitate the dendrites of mature neurons taking on their fully branched forms, perhaps providing another avenue to pursue for schizophrenia treatment.—Michele Solis.
Yoon KJ, Nguyen HN, Ursini G, Zhang F, Kim NS, Wen Z, Makri G, Nauen D, Shin JH, Park Y, Chung R, Pekle E, Zhang C, Towe M, Hussaini SM, Lee Y, Rujescu D, St Clair D, Kleinman JE, Hyde TM, Krauss G, Christian KM, Rapoport JL, Weinberger DR, Song H, Ming GL. Modeling a Genetic Risk for Schizophrenia in iPSCs and Mice Reveals Neural Stem Cell Deficits Associated with Adherens Junctions and Polarity. Cell Stem Cell. 2014 Jul 3;15(1):79-91. Abstract