| Shatz Home | Research | People | Publications | Images | Movies | Contact Info |  

 

Research Interests of the Shatz Lab Neural Signaling and Connectivity in Visual System Development The human brain is the most complex computational machine imaginable. There are well over a hundred billion nerve cells and even more connections. It is this complexity of connectivity, and its precision that determines how the brain works. But how does this wiring happen during development? How do nerve cells know where to grow their connections? To what extent does the brain have to function in order to achieve its adult precision of connectivity? The major aim of research in this laboratory is to elucidate the cellular and molecular mechanisms that establish this precision during neuronal development. Requirement for Neural Activity in Visual System Development When connections first form within a target structure, they are not established in the adult precise pattern. For example, connections between retina and its target nucleus in the thalamus, the lateral geniculate nucleus (LGN) are highly ordered with respect to eye input in the adult, such that ganglion cells from opposite eyes form connections within separate but adjacent eye-specific layers in the LGN. In contrast, in the fetus there are initially no layers: the ganglion cell inputs from the two eyes are intermixed and then gradually sort out to form the layers. Research in the laboratory demonstrated that the process of sorting requires neural activity. An important question currently under study is, how can neural activity "instruct" ganglion cell axons to sort into eye-specific layers? Before we could answer this question, we had to determine the source of the neural activity. The eye-specific layers form even before the rods and cones are mature. Therefore vision cannot provide the signals. Using special techniques that allow us to monitor the activity of hundreds of neurons simultaneously, we discovered that even before rods and cones are present, retinal ganglion cells are spontaneously sending signals to the brain. These signals are highly coordinated such that neighboring cells signal at the same time, and a "wave" of activity sweeps across the eye periodically. The waves are generated by an early-functioning circuit within the retina that involves both the retinal ganglion cells and a special type of interneuron, the cholinergic amacrine cells. These periodic waves are present only during development when the layers are forming, and disappear just before eye opening. What is more, blocking these waves prevents the formation of the eye-specific layers in the LGN. Thus, the brain can supply its own activity early in development to help in the process of forming precise connections; later on, vision takes over. For ganglion cell axons from each eye to sort into layers, LGN neurons must be able detect the the correlated firing of cells within one eye and strengthen those synapses that are coactive: "cells that fire together wire together". By means of whole cell recording in slices of LGN in vitro, we demonstrated that the synaptic connections between ganglion cells and LGN neurons can undergo strengthening similar to Long Term Potentiation in the hippocampus when stimulated with the same firing patterns as those generated in vivo by the ganglion cells. We also discovered, in a special "reduced brain preparation", that the spontaneous waves of activity generated in the eye are not only sent on to the LGN in the form of action potentials, but they are then relayed across synapses where they are powerful enough to make the LGN neurons fire action potentials. Neurotrophins as Signaling Molecules in Activity-Dependent Development The process of forming precise connections and eliminating imprecise ones during development involves not only short term changes in the strength of synaptic connections between the retinal ganglion cells and LGN neurons, but also long term structural changes in the branching patterns of ganglion cell axons. How is this long term change effected? One idea is that the target neuron releases a retrograde growth factor that is taken up by the presynaptic axon and acts to promote long term growth and stabilization of the input. We have been investigating this idea by studying the next set of connections in the visual pathway: from the LGN neurons to neurons of the primary visual cortex in the occipital lobe. In the adult, LGN axons representing each eye are segregated from each other in the primary visual cortex into columns (rather than into layers as in the LGN). Neural activity again is needed for columns to form, but during a later time in development when vision (rather than waves) provides the input. If growth factors are involved in the formation of the ocular dominance columns, then supplying an excess of the right factor(s) should prevent column formation. This is because LGN axons might be stimulated to grow and branch without regard to the correct patterns of neural activity, which would normally have regulated the availability of the factor. This is exactly what we found when certain specific neurotophins were infused into the cortex: Brain-derived growth factor (BDNF) and Neurotrophin-4/5 (NT-4/5) prevented column formation while Nerve Growth Factor (NGF) and Neurotrophin-3 (NT-3) had no effect. Because extra neurotrophins were added to the developing brain, these experiments could not reveal whether endogenous neurotrophins are normally involved in the formation of columns. So, in an additional set of experiments, competitive antagonists of the neurotrophins were infused. Results again indicated a requirement for BDNF, NT-4/5 or some other as yet unidentified endogenous ligand of TrkB, the high affinity receptor for these neurotrophins. These observations imply that specific members of this family of nerve growth factors play an essential role in the activity-dependent formation and remodeling of synaptic connections during the development of the mammalian central nervous system. Activity-dependent regulation of Class I MHC in visual system development and adult hippocampus. To learn more about the molecular mechanisms responsible for activity-dependent synaptic remodeling, we have started molecular screens for genes in the developing visual system whose level of expression is regulated by neural activity. The strategy is to compare gene expression in the presence or absence of endogenous action potential activity during the development of connections between retina and LGN. In a differential display experiment, we discovered that mRNAs within the LGN encoding class I major histocompatibility complex antigens (class I MHC) decreased following activity blockade during the period of layer formation. Class I MHC is expressed in times and places of known synaptic plasticity, and can be down-regulated, not only by blocking spontaneous retinal waves early in development, but also by blocking visually-driven activity in postnatal life during the period in which ocular dominance columns form in cortex. Class I MHC levels can be increased in hippocampal neurons following kainate-induced seizures in adults. This discovery implies a novel role for this family of molecules--not thought to be present normally in neurons-- in nervous system development and plasticity. Class I MHC molecules, previously thought to mediate cell-cell interactions exclusively in immune function, are present in the CNS where they may mediate neuronal signaling and activity-dependent changes in synaptic connectivity. These observations provide insights into possible cellular and molecular mechanisms of activity-dependent development. They also imply that these mechanisms, operating during development, may be rather similar to those thought to occur during memory and learning. This is exciting because a good deal is known about the signaling pathways from synaptic activation to regulation of gene expression in the phenomenon of hippocampal LTP and we believe that similar pathways are involved during development. Thus, our studies pertain not only to developmental mechanisms of brain wiring, but may also help in the understanding of brain mechanisms of learning and memory.