Long-term expression of optogenetic proteins including channelrhodopsin-2 (ChR2) is widely used to study neural circuit function, but whether ChR2 expression itself perturbs circuits is not known. We expressed a common construct, CAG::ChR2 (H134R)-EYFP-WPRE, in L2/3 pyramidal cells in rat somatosensory cortex via in utero DNA electroporation (IUE). L2/3 pyramidal cells expressed ChR2-EYFP, but histology revealed abnormal morphology and targeting of ChR2-EYFP expressing axons, beginning at postnatal day (P) 33 and increasing with age. Axonal abnormalities included cylinders that enveloped pyramidal cell proximal apical dendrites, and spherical, calyx-like structures that surrounded neuronal cell bodies, including in L4. These are abnormal subcellular and laminar targets for L2/3 pyramidal cell synapses. Abnormalities did not occur in cells expressing GFP instead of ChR2, or in intermixed ChR2-negative axons. Long-term viral-mediated expression (80 d) did not cause axonal abnormalities when the CAG promoter was used, but produced some abnormalities with the stronger αCaMKII promoter (albeit much less than with in utero electroporation). Thus, under some circumstances high-level, long-term expression of ChR2-EYFP can perturb the structural organization of cortical circuits.

Summary Local recurrent networks in neocortex are critical nodes for sensory processing, but their regulation by experience is much less understood than long-distance (translaminar or cross-columnar) projections. We studied local L2/3 recurrent networks in rat somatosensory cortex during deprivation-induced whisker map plasticity, by expressing channelrhodopsin-2 (ChR2) in L2/3 pyramidal cells and measuring light-evoked synaptic currents in ex vivo S1 slices. In columns with intact whiskers, brief light impulses evoked recurrent excitation and supralinear inhibition. Deprived columns showed modestly reduced excitation and profoundly reduced inhibition, providing a circuit locus for disinhibition of whisker-evoked responses observed in L2/3 in vivo. Slower light ramps elicited sustained gamma frequency oscillations, which were nearly abolished in deprived columns. Reduction in gamma power was also observed in spontaneous LFP oscillations in L2/3 of deprived columns in vivo. Thus, L2/3 recurrent networks are a powerful site for homeostatic modulation of excitation-inhibition balance and regulation of gamma oscillations.

Inhibitory synapse development in sensory neocortex is experience-dependent, with sustained sensory deprivation yielding fewer and weaker inhibitory synapses. Whether this represents arrest of synapse maturation, or a more complex set of processes, is unclear. To test this, we measured the dynamics of inhibitory synapse development in layer 4 of rat somatosensory cortex (S1) during continuous whisker deprivation from postnatal day 7, and in age-matched controls. In deprived columns, spontaneous miniature inhibitory postsynaptic currents (mIPSCs) and evoked IPSCs developed normally until P15, when IPSC amplitude transiently decreased, recovering by P16 despite ongoing deprivation. IPSCs remained normal until P22, when a second, sustained phase of weakening began. Delaying deprivation onset by 5 days prevented the P15 weakening. Both early and late phase weakening involved measurable reduction in IPSC amplitude relative to prior time points. Thus, deprivation appears to drive two distinct phases of active IPSC weakening, rather than simple arrest of synapse maturation.

No abstract