Currently Online (9)
The first talk by Daniel Geschwind walked us through some of the main features of genome-wide association studies with regard to psychiatric disorders such as autism. Their experimental approach includes using human fetal neurons and looking, e.g., at their gene expression profiles over several time points. What they, perhaps not surprisingly, found was that many of the genes up-regulated during development are associated with psychiatric disorders. He went on to tell us about Weighted Gene Co-Expression Network Analysis which provides an estimate of which genes are co-expressed with regard to their expression levels. Essentially, this technique seems to be a correlational analysis of gene expression profiles over time and grouping those genes the expression of which co-vary over time. One such analysis showed that genes associated with autism are five times more connected than average. In another study they used post-mortem samples from autistic patients (three candidate regions). Using gene expression analysis, they could predict which sample came from an autistic brain and which from a control brain, indicating that there is a shared molecular signature in a subset of autism cases. Another result from this study was that a strong differential pattern of gene expression in different brain regions found in control brains was lost in autistic brains. Looking at which genes are the ones most contributing to these effects, they found evidence that microglia upregulation is involved in autism and that this upregulation might be involved in synaptic formation/pruning in the developing brain. The interesting part of this talk for me was that the genetic heterogeneity of autism spectrum disorders might possibly be reducible to a relatively small set of common molecular pathways and networks.
The second talk by Daniel Weinberger was about schizophrenia. The talk started out with some data on identical twins where only one individual had schizophrenia. From samples such as these, they estimate the genetic risk for schizophrenia. The genes they find also map on general cognitive development. They combine these genetic studies with fMRI to establish connections between genetic information and brain activity. They found that genetic risk factors for schizophrenia are linked to inefficient processing during cognitive tasks - in schizophrenics and their healthy siblings. One point he stressed from his studies was that information processing in the brain is subserved by distributed, degenerate networks. He told the story of a transcription factor, ZNF804a, which had previously discovered to be involved not in any specific region activating during a cognitive task or not, but in how well prefrontal activity was coupled to hippocampal activity, in schizophrenia. Mirroring the degenerate netowrk actions in the brain, he told us that even simple behaviors are subserved by complex, degenerate genetic networks. Not surprisingly, he emphasized the important role of epistasis in such genetic networks. So far, everything he said was right down my alley To my great pleasure, he even quoted the PNAS paper I shared on Google Plus recently. He went on to rattle through a whole bunch of genes which are involved in associative learning and hippocampal synaptic plasticity and part of a genetic network. None of these genes by themselves scores on risk for schizophrenia. However, the gene-network does score highly on genetic risk for schizophrenia. This means that risk factors in individual genes can be compensated for by the degeneracy of the network. However, when the whole pathway is affected due to several hits in this pathway, the buffering capacity of the network is severely challenged and schizophrenia might result. This pattern was also found in the three genes NRG1-ERBB4-ACT1: individual genes and pairs of genes had low to moderate risk for schizophrenia, a genome with all three risk-associated genotypes had a 27-fold increase in risk for schizophrenia.
These first two talks were excellent examples of how the genetic variability in humans is starting to look more and more like nature's way to do genetic manipulations for us that we otherwise would introduce artificially in our genetic model systems. Given the access to huge samples of individuals, this area of research looks very promising and will have a great future.
Albert Galaburda, in the third presentation of this session talked about developmental dyslexia. In contrast to the previous talks, there seem to be reasons to believe that a candidate gene approach might be fruitful. The genes he talked about were Dyx1c1, Kiaa0319 and Dcdc2. Two phenotypes of a Dyx1c1 knock-out in mice are hydrocephaly and situs inversus (organs being formed on the wrong side of the body). These phenotypes share a common underlying phenotype, ciliopathy. Apparently, ciliary action is required to establish asymmetric organ formation (e.g., having the heart on the left side) and to establish a normal flow of spinal fluid. These phenotypes could be linked mechanistically to some of the processes thought to underly developmental dyslexia.
The final talk of this session Allan Reiss talked about Fragile X and Williams Syndrome, two disorders with opposite social phenotypes. After showing us some very interesting videos with patients, he told us a little about the underlying genetics. In Fragile X the responsible gene is FMR1 , encoding an mRNA binding protein, binding mRNAs involved in synapse formation. In Willimas Syndrome, there are 26-27 genes deleted on chromosome 7. The behavioral phenotypes are essentially opposite: Fragile X patients avoid eye contact, while Williams patients seek eye contact, exceeding healthy controls. In Fragile X, social anxiety is lager than non-social anxiety, while in Williams syndrome it's the opposite. However, both groups have high levels of anxiety. In language, Fragile X patients are delayed while Willliams patients are enhanced. Both groups show an enhances stress response. There were some more features I couldn't write down fast enough. Neuroanatomical phenotypes include larger brains for Fragile X patients and smaller for Williams. In Fragile X the frontal cortex is smaller than normal, while in Williams it's slightly larger. The Superior Temporal Gyrus is smaller in Fragile X and larger in Williams. The caudate is much larger than normal in Fragile X while it is smaller in Williams syndrome. In Fragile X, the amygdala is small, while in Williams it's large. Moving on to functional imaging during tasks involving faces, he showed that amygdala activation to faces is increased in Fragile X but decreased in Williams with opposite activations in the fusiform cortex. Prefrontal cortex is activated more in Williams syndrome and less in Fragile X. All these neuroanatomical and neurofunctional phenotypes are obviously very consistent with the behavioral phenotypes.
Render time: 0.7341 sec, 0.1514 of that for queries.