21 November 2007. It may be a surprising idea, but it is becoming clearer from work done in several labs that complement factors and MHC Class I proteins are regulators of synapse formation during fetal development, synapse plasticity, and synapse repair. This function occurs without apparent T cell involvement or any presentation of self or non-self peptides. It also now appears that these proteins may be relevant to neural disorders such as autism and schizophrenia, as discussed at the 2007 meeting of the Society for Neuroscience, held 3-7 November in San Diego, California. Lisa Boulanger of the University of California, SD, led a symposium titled "The New Neuroimmunology: Immune Proteins in Synapse Formation, Plasticity, and Repair." Parts of the research presented appeared on ARF before, see Bar Harbor Report 2006, ARF related news story, and ARF news story). Below is an update of this emerging story.

Carla Shatz of Harvard University first demonstrated that MHC Class I proteins were functionally required for development and plasticity of the CNS in 2000 (Huh et al., 2000). Today Shatz presented work from her laboratory—which primarily focuses on the visual cortex ocular dominance models of plasticity—showing that Class I proteins colocalize with PSD-95, considered by many a “master organizer of synapses” (see Goddard et al., 2007). Shatz demonstrated that Class I proteins are expressed at high levels in neurons of both somatosensory cortex as well as hippocampus. Using a double knockout mouse that lacks both β2-microglobulin (β2M) and the transporter associated with antigen processing 1 (TAP1), which dramatically reduces the surface expression of all MHC Class I proteins, Shatz and her colleagues showed that Arc (activity-regulated cytoskeletal-associated protein) induction in the visual cortex is abnormally widened after visual stimulation of the double knockout mice. These data suggest that MHC Class I proteins regulate the process of synaptic plasticity in the ocular dominance model used. Shatz proposed that the gene PirB (paired immunoglobulin-like receptor B) encodes the receptor for MHC Class I in neuronal synapses, which was shown previously to be expressed in neurons in the brain, and functions to limit experience-dependent plasticity in the visual cortex (Syken et al., 2006).

Staffan Cullheim of the Karolinska Institute in Stockholm, Sweden, used the same double β2M/TAP1 knockout mice to examine the role that Class I proteins play in the elimination of synapses following nerve injury (Thams et al., 2007; Cullheim and Thams, 2007). The data presented focused on the role of activated microglia and MHC Class I protein specificity in the synapse removal process after axotomy, and suggested that there may be a differential effect of these proteins in excitatory (NMDA) versus inhibitory (glycine or GABA) synapses, with more elimination of inhibitory synapses.

Ben Barres of Stanford University, Palo Alto, California, shifted the focus from MHC Class I proteins to the role of components of the complement cascade C1q and C3 to act as “punishment signals” and cause axon atrophy and withdrawal. Using a new imaging method called array tomography (Micheva and Smith, 2007), Barres showed that C1q colocalizes to developing CNS synapses using immunofluorescent staining of 70 nm sections of developing mouse brain. Barres further showed that in C1q or C3 knockout mice, synapse refinement that normally occurs in early postnatal development (P5 through P30) was defective, resulting in more synapses, not more neurons. Barres presented a model in which immature astrocytes clustering near the developing synapses release C1q and C3 into the synapse to prune and eliminate synapses. An important implication of Barres’s presentation is the tantalizing idea that inhibitors of the complement cascade may have the potential to block neurodegeneration.

The final talk of the symposium was by Lisa Boulanger. Boulanger has continued on with the work she had done as a postdoctoral fellow in Shatz’s laboratory, and has pursued studies of MHC Class I proteins in synaptic pruning and plasticity to explore the potential implications for both autism and schizophrenia. Boulanger is no stranger to the autism field, having published a thoughtful article on abnormal development of brain connectivity in autism in 2004 (Belmonte et al., 2004). In her talk, Boulanger examined the electrophysiological responses of β2M/TAP1 double knockout mice in paired pulse inhibition (an experimental model of sensory gating deficits in schizophrenia that is widely used, if not widely accepted), AMPA receptor fEPSP, and NMDA-induced chemical LTD, which is similar to low frequency stimulation induced LTD.

Surprisingly, the last test, which results in a stable LTD in wild-type mice, instead induced a robust LTP in the β2M/TAP1 double KO mice. Pursuing these studies further, Boulanger demonstrated that NMDA treatment caused a dramatic increase in surface AMPA receptor expression. Boulanger proposed that the increase in AMPA receptors was a result of increased internalization of AMPA, accompanied by a dramatic increase in recycling AMPA receptors back to the synaptic surface to cause a homeostatic shift of net increase.

Boulanger proposed that MHC proteins are tied to neural diseases, speculating that maternal immune challenge increases the risk of the unborn fetus to such diseases (see Patterson, 2007 and Schizophrenia Research Forum news story). In the model that Boulanger proposed, induction of a maternal immune response in mice increases maternal cytokines that are capable of entering the fetal blood circulation. These cytokines, if exposed to the developing nervous system of the fetus, may regulate MHC Class I levels in neurons. The question remains, Do changes in neuronal MHC Class I expression mediate changes in the development of the fetal brain?—Gwendolyn T Wong.