Abstract:
The inferior olive is a brainstem nucleus which receives input from a broad spectrum of sources ranging from visual and auditory to proprioceptive and nociceptive to vestibular and motor-sensory inputs as well as cortical innervation but also inhibitory innervation from the cerebellar nuclei via the olivo-cerebellar loop. Olivary neurons project to Purkinje cells in the cerebellar cortex via the climbing fibers and to the cerebellar nuclei via climbing fibers collaterals. While the input is manifold and integrates many modalities, the output climbing fiber signal consists of a short high frequency burst with a variable number of burst components but an amazingly low burst frequency of about 1-2 Hz. The climbing fiber signal elicits a massive calcium influx in the dendritic tree of Purkinje cells which gives rise to the somatic complex spike. Although Purkinje cell function has drawn a lot of attention, the purpose of the complex spike and its source, the climbing fiber signal, remain only poorly understood. This has consequences for the understanding of the role of the cerebellum at large. A key towards better insight into the nature of cerebellar function may therefore lie in the investigation of the function of the inferior olive in general and of the spatio-temporal nature of olivary neuron activity in particular. Neurons of the inferior olive are remarkable as they are densely connected to each other via dendro-dendritic electrical synapses –the highest density of electrical synapses in the central nervous system – resulting in clusters of electrotonically coupled neurons. It has been shown that these electrical synapses are primarily composed of the gap junction protein connexin 36 (Cx36). Olivary neurons exhibit sub-threshold oscillations (STO) which are based on the interplay of calcium conductances. These oscillations are highly synchronous among neighboring olivary neurons and it is assumed that electrical coupling provides the means of this synchrony. Current interpretations of the functional relevance of the STO assume that the probability of regenerative events is higher at the peak of the STO phase when the membrane is more depolarized. This would allow a temporal and spatial control of activity among coupled neurons. Moreover, recent work has indicated that the strength of coupling among olivary neurons is dynamically regulated on different time scales. In particular synaptic input located at the electrical synapses seems to be able to modulate coupling and thus provide control over the extent of coupled neuron clusters and their synchrony on a short time scale. Previous studies have attempted to understand the purpose of the electrotonic coupling by using Cx36 knockout mice. In these animals olivary neurons were able to maintain STO in the absence of Cx36 but the oscillations were not in synchrony any more. Yet, only a mild phenotype with subtle impairment of motor learning but not of motor coordination could be described. This may be due to the fact that olivary neurons compensate for the lack of Cx36 on the cellular level. To overcome the issue of developmental compensation lentiviral vector mediated dominant negative inhibition of Summary 2 Cx36 in rats was used as a tool to transduce olivary neurons. This work showed that olivary neurons in areas with high transduction density were no longer able to maintain STO. Unfortunately animals used in this work were not subjected to behavioral assessment and thus no conclusion on possible motor impairment could be drawn. Hence, it also remains unexplored what possible consequences on locomotion may arise from postnatal reduction of Cx36 based coupling. We tried to address the role of Cx36 mediated coupling in the rat inferior olive by a strategy resting on two pillars. The first pillar was formed by investigating the change of gap junctional coupling during the first postnatal weeks. We recorded the presence and properties of spontaneous regenerative events (spontaneous action potentials and spikelets) as well as STO in olivary neurons in acute brain slices from animals at postnatal week two, three or five with the whole-cell patch clamp technique. Furthermore, we diffused neurobiotin, a gap junction permeable dye, to the patch clamp recorded neurons and reconstructed the extent of the coupled networks in three dimensional confocal laser scanning microscopy image stacks. The second pillar was formed by RNA interference mediated knockdown of Cx36 by specific short-hairpin RNA delivered by lentiviral vector. At three weeks (postnatal day 21) viral vector solution was stereotaxically injected in the center of the inferior olive under electrophysiological guidance. Animals (5w-shCx36) were allowed to recover for several days and were subjected to free locomotion analysis before being sacrificed at postnatal week five for whole-cell patch clamp recordings of olivary neurons / neuron pairs. Recordings of pairs of olivary neurons enabled us to assess the electrotonic coupling in 5w-shCx36 neurons compared to the wildtype or non-functional control and to investigate differences in STO presence and synchrony. Moreover, we examined the occurrence of spontaneous regenerative events and their kinetics in respect to the phase and amplitude of the STO cycle in the prospect to identify features that may code for the oscillatory state of the respective neuron – putative important information to be transmitted to the cerebellar cortex and the cerebellar nuclei. We found that gap junctional coupling among olivary neurons undergoes dynamical changes during the first postnatal weeks with highest coupling around postnatal week three. Whereas no case of neurobiotin dye-coupling could be observed at postnatal week two, the incidence of dye-coupling and the number of coupled cells per dye-filled neuron increased at postnatal week three but dropped again by postnatal week five. Robust STO were not found until postnatal week three, the time point of maximal coupling, and continued to be present from then on. The probability of occurrence of spontaneous action potentials as well as of spikelets in olivary neurons followed the dynamics of gap junctional coupling and so did the spontaneous firing frequencies of both event Summary 3 types. We further found a general maturation of the regenerative events during postnatal development. Spontaneous action potentials as well as spikelets became faster from postnatal week two to five, indicating substantial changes in membrane channel composition and density. RNAi was found to effectively reduce the amount of Cx36 immunoreactivity in transduced olivary neurons indicating that a substantial proportion of gap junctions was eliminated. This manifested in highly reduced dye-coupling and almost complete diminishing of electrical coupling. The largely absent coupling led to an increased probability of observing spontaneous action potentials. Moreover, the firing frequency of spontaneous action potentials was higher than in the age matched wildtype, resembling frequencies found in olivary neurons from postnatal week three. Interestingly, spikelets were still observed in 5w-shCx36 neurons which is counterintuitive to the idea of them being the echo of action potentials transmitted across gap junctions, questioning the prevailing interpretation. Uncoupled 5w-shCx36 neurons were less able to maintain STO synchrony with neighboring neurons. Yet, most of these neurons were stable oscillators during the full time of recording and we even observed an increase in dominant STO frequency as a result of uncoupling. We observed that spontaneous action potentials and spikelets occurred independent of the STO phase at postnatal week three. At postnatal week five however, both event types were aligned to the peak of the STO phase. In contrast, uncoupling led to a loss of spikelet alignment to the STO phase while the alignment of spontaneous action potentials to the STO phase was undisturbed. Parameters of the kinetics of both regenerative event types were sensitive to the STO phase or amplitude and may be capable to convey information about the oscillatory state of the neuron. While several electrophysiological changes on the neuronal level could be attributed to the loss of gap junctional coupling, we identified only a mild phenotype during free locomotion. Classifying animals based on the affected olivary sub-nuclei, we could identify most alterations of locomotion parameters when substantial parts of the medial accessory olive (MAO) were transduced. We conclude that gap junctional coupling is necessary for the proper maintenance of STO synchrony and the timing of spikelets in respect to the STO phase. Moreover, spontaneous activity of IO neurons seems to positively depend on coupling. Furthermore, features of action potential and spikelet kinetics appear to convey information on the STO phase and amplitude at event occurrence. Yet, on the level of locomotor coordination, only moderate changes were observed and it seems that the impact was strongest if the medial accessory olive was affected to a great extent.
Cerebellum