We seek to determine the epigenetic mechanisms through which environmental factors influence development of the nervous system. In particular, we are interested in both the coupling between electrical and synaptic activity in neural systems and the growth, differentiation, and maturation of neurons and their connections in synaptic circuits. Elimination of redundant, non-functional synapses is a major process involved in nervous system development, and we have demonstrated some of the molecular and cell-biologic mechanisms responsible for activity-dependent synapse elimination in an in vitro model system. We seek to extend our studies to some neurodevelopmental disorders such as autism.

Activity-Dependent Synapse Loss and Stabilization at the Neuromuscular Junction
Nelson, Jia, Li, Viloria, Yang, Lanuza; in collaboration with Fields
The Hebbian, activity-dependent modulation of synapse efficacy, in which activated pathways are strengthened and inactive ones weakened, depends on an integrated pattern of activation of protein kinases, such that appropriate phosphorylation reactions occur that differentially affect the stability of activated and non-activated synapses and their associated receptors. One step in the activation of the kinases, in particular protein kinase C (PKC), involves the action of the serine protease, thrombin, and its associated receptor, the protease-activated receptor or PAR. Thrombin’s action can be reproduced by a peptide, the thrombin receptor–activating peptide (TRAP), which can activate the PAR. Collaborative experiments with Dr. Lanuza and her group in Spain have shown that the thrombin receptor (ThR) is localized at the neuromuscular junction in vivo and that application of the TRAP increases the rate of synapse elimination that occurs naturally in the Lanuza group’s system. Thus, the molecular apparatus for thrombin’s participation in the synapse modification process is present and functional at the neuromuscular junction. The natural elimination process can be divided into three stages: an early stage in which significant loss of polyneuronal innervation occurs with relatively little change in the distribution of post-synaptic receptors, an intermediate stage in which both axonal loss and receptor changes are prominent, and a final stage in which the innervation has become primarily mono-neuronal, although substantial changes in the receptors occur. It is during the early/intermediate stage that pharmacological and genetic prevention of PKC action blocks synapse elimination. Genetically modified (PKC theta knock-out) animals show a delay of synapse loss developmentally, but eventually the muscle innervation becomes completely mononeuronal. The use of PKC blockers in vivo over the same period also completely prevented synapse elimination. We have investigated what aspect of the neural activation process might be involved in the kinase-mediated change in synapse effectiveness. Carbachol can be used to provide cholinergic stimulation of skeletal muscle, and we have shown that such treatment of cultured skeletal myotubes results in activation and membrane translocation of PKC theta.

A variety of experiments have implicated PKA as playing a crucial role in the stabilization of activated synapses. In view of the redundant mechanisms underlying activity-dependent synapse elimination implied by the studies of PKC block or knock-out, we were interested in alternative messengers or mediators of the elimination/stabilization. In collaboration with Douglas Fields, we have examined the possibility that ATP might be involved in the process. For this purpose, we have used apyrase, an enzyme that breaks down ATP, and suramin, a blocker of the P3 purinergic receptor that mediates some of the effects of ATP. In the presence of these agents, stimulation produces a significant decrement in synapse strength (stimulation of control, untreated preparations do not produce a decrement of the stimulated inputs). The depression lasts for over an hour after the drugs are washed out and stimulation of the preparation terminated. Both nerve and muscle in our preparations release large amounts of ATP, with the release modulated by activation of the nerve or muscle. Thus, it seems possible that ATP may play a role in the stabilization of activated synapses, possibly through activation of a purinergic receptor.

Activation of the PI3 kinase by Wortmannin does not affect synapse stability in either resting or stimulated synapses, leading to the conclusion that phosphorylations in general are not involved in synapse modulation. It seemed worthwhile to explore further the possibility that kinases other than PKA and PKC were involved in activity-dependent synapse modulation. The ras-mitogen–activated protein kinase (MAPK) and cAMP/PKA-mediated pathways play important roles in neuronal plasticity. We have used different blockers of the MAPK to test for its involvement in synapse modification in our system. Two such blockers are PD98059 and UO128. Neither of these produces synapse loss in the absence of synapse activation, but, in conjunction with stimulation, both do produce loss. In the presence of synapse stimulation, the inactive analog of UO128, UO124, produces no synapse loss. The results suggest that MAPK may be involved in the Hebbian synapse modulation that occurs in our model system.

Molecular Bases for Neurodevelopmental Disorders
Nelson, Kohler, Satyanarayana, Krek-Larner;* in collaboration with Grether
We have used a variety of immunoaffinity analytical methods to determine the levels of several neuro-active molecules in blood samples from controls and children with different neuro-developmental disorders. Through collaboration with members of the California Department of Health Services, we have access to blood specimens that were drawn at birth from a large sample of children and archived with subsequent diagnosis of some of the children as having autism or other developmental disorders. In very-low-birth-weight children with and without diagnoses of infection or cerebral palsy, we found no correlation of clinical condition with levels of a number of cytokines. Our finding stands in contrast to earlier results showing significant differences in cytokine levels that correlated with clinical condition in term infants. We have measured the neonatal blood levels of the neurotrophins NT-3, NT-4, and Brain Derived Neurotrophic Factor (BDNF) and of several cytokines in controls and in children subsequently shown to be autistic. We observed no significant differences between cases and controls for IL-1 or IL-8, but the level of BDNF was about 25 percent lower in samples for autistic cases than in normals. We obtained our result with the double antibody sandwich immunoaffinity method used with the bead flow–based technique of the Luminex system. Given that our result did not show the increase in BDNF (as well as some other analytes) found earlier with the single antibody recycling immunoaffinity chromatographic method, we are attempting to develop a single antibody version of the Luminex method to determine if methodological issues explain the discrepancy.