The brain is the most complex organ in the human body. Rough estimates suggest that the central nervous system of higher mammals is comprised of 10^11 neurons. There neurons communicate with each other by 10^15 contact sites termed synapses. At present, our knowledge about synaptic function is largely based on the highly detailed analysis of a small number of “model synapses”, such as the neuromuscular junction and the calyx of Held. However, the function of cortical synapses is just beginning to emerge.

My group will focus on the hippocampus, which plays a key role in spatial information processing and episodic memory. The hippocampus is comprised of three subfields, the dentate gyrus, the CA3 (cornu ammonis 3) region, and the CA1 (cornu ammonis 1) region. Two general types of neurons are present in all regions: principal neurons, which use glutamate as transmitter and have an excitatory effect on their postsynaptic target cells, and interneurons, which release g-aminobutyric acid (GABA) and have inhibitory postsynaptic effects. Both types of cells exist in various subtypes, such as granule cells, pyramidal neurons, and various types of GABAergic interneurons. Understanding the rules of signal processing in such complex networks requires information about both connectivity (the “connectom”) and function of identified synapses.

The central goal of our research team is to determine the rules of communication at identified synapses in hippocampal microcircuits at a quantitative level. This may ultimately allow us to understand how higher brain functions are implemented in brain circuits and how these functions are perturbed in neurological and psychiatric diseases. Within the next three to five years, my group at ISTA will address the following questions:

1. Nanophysiology of fast-spiking, parvalbumin-expressing GABAergic interneurons

Rationale. In contrast to the large amount of information available on the signaling properties of pyramidal neurons, comparatively little is known about the molecular, subcellular, and cellular properties of GABAergic interneurons. Principal neurons and interneurons fundamentally differ in signaling characteristics, but the reasons for these differences are incompletely understood.

Experimental plan. One of our major goals is to precisely define the spatiotemporal signaling properties of GABAergic interneurons at nanometer and microsecond resolution. We will concentrate on a major GABAergic interneuron subtype, the fast-spiking, parvalbumin-expressing basket cell (BC), because of its salient functional properties and its uniquely powerful output. We will examine the functional properties of dendrites, axons, and presynaptic terminals by subcellular patch-clamp techniques (pioneered by the Jonas group). Furthermore, we will study the distribution of ion channels and receptors on the surface of these GABAergic interneurons, using subcellular patch-clamp recording and transmitter uncaging.

Significance. The long-term goal is to understand signaling in the fast-spiking, parvalbumin-expressing BC at the same level of depth as that in hippocampal and neocortical pyramidal cells, which served as a “gold standard” for the analysis of neuronal function over many years.

2. Analysis of synaptic mechanisms of information storage in hippocampal CA3 cell networks

Rationale. Several lines of evidence suggest that the hippocampal network is involved in higher-order processing of spatial information. For example, the hippocampus contributes to pattern separation, amplifying differences between subtly different input signals. Likewise, the hippocampus is involved in pattern completion, restoring original patterns from degraded or incomplete inputs. How these complex network functions emerge from the elementary properties of neurons and synapses is, at present, unclear. Several lines of evidence suggest that CA3-CA3 cell synapses (also referred to as commissural / associational or c/a synapses), the most abundant glutamatergic synapses in the hippocampus, play a key role in pattern completion. However, the functional properties of these synapses are largely unknown.

Experimental plan. A second major aim of my lab is to determine the synaptic mechanisms of pattern completion in the CA3 circuit of the hippocampal network. We plan to characterize both the functional and the morphological properties of unitary synaptic signals at CA3-CA3 cell synapses. The problem of low connectivity between CA3 neurons will be overcome by simultaneous recording from ensembles of up to ten neurons. Functional parameters will be correlated with structural properties of synapses. The group will also study the rules of induction of synaptic plasticity at CA3-CA3 pyramidal cell synapses. Finally, as the active electrical properties of dendrites of CA3 pyramidal cells shape the induction rules of synaptic plasticity, we will characterize the dendrites of CA3 pyramidal neurons by subcellular patch-clamp techniques.

Significance. From these experiments we expect to obtain the first quantitative functional characterization of CA3-CA3 cell synapses, the most abundant synapses in the hippocampus. These results will form the basis for understanding the synaptic mechanisms of higher-order information processing in the hippocampus.

3. Analysis of glutamatergic and GABAergic transmission in hippocampal synapses in vivo

Rationale. Analysis of activity of neurons under in vitro conditions in brain slices led to a greatly improved understanding of the molecular, subcellular, and cellular events underlying brain signaling under precisely defined experimental conditions. However, if we want to understand how synaptic activity shapes higher brain functions, analysis of synaptic events has to be correlated with rhythmic activity in vivo, and ultimately with behavior of the animal. To address this issue, the Jonas lab has begun to explore the properties of synaptic transmission in the hippocampus in vivo.

Experimental plan. First, we will characterize the properties of synaptic signaling at various types of input synapses in vivo in anesthetized rats or mice and compare their properties with those previously described in vitro. Input synapses will be selectively stimulated electrically and optogenetically, expressing channelrhodopsin under the control of specific promoters and activating the infected neurons by light pulses. Next, whole-cell recordings from various types of neurons of the hippocampal formation will be obtained in awake animals. Experiments will be done either in freely moving animals or in immobilized animals placed in virtual environments. The properties of neuronal activity will be compared between entorhinal cortex neurons (the input neurons to the hippocampus), dentate gyrus granule cells, CA3 pyramidal cells, and CA1 pyramidal cells. Finally, we will examine the behavioral consequences of the low- and high-frequency activation of single hippocampal neurons by intracellular stimulation.

Significance. The present results will provide key information about the role of excitatory and inhibitory synapses in higher-order computations carried out in the hippocampus. In particular, whole-cell recording from entorhinal cortex neurons will reveal the synaptic mechanisms of grid cell activity, while recording from dentate gyrus granule cells will allow us to determine the mechanisms of grid cell – place cell conversion and pattern separation.

Techniques. To address these questions, my lab will use cutting edge electrophysiology, imaging, and computational techniques. In particular, subcellular patch clamp (presynaptic and dendritic recording) and multiple cell recording techniques in rat and mouse brain slices will be used. As in vivo patch-clamp recording techniques become iteratively refined, we expect that the focus of the lab will increasingly shift to in vivo experiments.

Long-term goals and perspectives. The long-term goal is to integrate the information about neurons and synapses into realistic models of neuronal networks (“build it, and you understand it”, Hopfield). The assembly of a full-sized network model of the hippocampus based on experimental data is in reach. ISTA will be a particularly suitable environment for implementing such a large scale computational approach, because of both the excellent computational infrastructure and the outstanding expertise in computer science.

The combined approach comprised of in vitro experiments, in vivo analysis, and modeling will allow a rigorous test of current ideas about the mechanisms underlying higher-order information processing in neuronal networks of the hippocampus. More generally, the results will shed light on one of the central questions in modern neuroscience, which is to understand how elementary synaptic properties shape complex network functions.

In the long term, our results may be also relevant for understanding the pathophysiology of brain disorders. For example, evidence accumulates that fast-spiking, parvalbumin-expressing BCs are involved in a variety of brain diseases. The growing list includes epilepsy, schizophrenia, autism, and neurodegenerative diseases. Thus, a detailed understanding of the functional signature of these cells may, at some point, help to develop new therapeutic strategies for neurological and psychiatric disorders.