Optogenetics applied on
multiple circuit levels
Sicco D. de Knecht, B.Sc. – 0522368 Supervisor: Prof. Wytse J. Wadman
Co-assessor: Dr. Natalie Cappaert
Brain and Cognitive Sciences, Track Neuroscience
Abstract
Circuit analysis is a discipline of neuroscience that investigates neural connectivity on multiple levels. Much about neuronal circuitry remains unclear due to limited (experimental) possibilities and the complicated nature of important questions. Optogenetics is gaining ground in modern neuroscience and is being applied at many relevant circuital levels. The cell-type specificity, in vivo and in vitro applicability and bidirectional control of neurons make it a suitable tool for studying complicated and integrative neural systems.
This review aims to point out the significant and sometimes limited additions of the optogenetic approach in circuit analysis. The optogenetic methods discussed in this review have the potential to reduce and/or overcome some of the limitations of widely used techniques in circuit analysis on different levels. In this review three of these levels are identified: the subcellular level, small functional circuits and global circuits, and the most important progress made within these areas is discussed.
Introduction
Optogenetics entails the field of genetic and optical methods that allow the experimenter to gain control over cellular processes with the use of light. These genetically guided mechanisms are able to ensure a cell-specific expression of light-sensitive proteins, allowing for cell type specific
manipulation of neuronal populations. These intrinsic properties of the optogenetic toolbox broaden the possibilities for answering neuroscientific questions in a novel, non-invasive, cell-type
dependent, and controllable manner.
In 2005 a light sensitive opsin receptor was successfully transfected into a neuronal cell for the first time (Boyden et al. 2005). Since then the development and application of these opsins (overview in (Rein & Deussing 2012)) has taken a flight and optogenetic approaches and have been used in a growing number of experimental paradigms (overview in (Deisseroth 2011; Andre Berndt et al. 2011)). In this review the individual advantages of these applications on multiple levels of circuit analysis will be discussed, as well as their limitations.
Circuit analysis in neuroscience is the field that studies the interconnectedness of the brain on a functional level. Circuit analysis is concerned with understanding how parts of a circuit operate within a larger network. This is a crucial step to understanding the impact single units can have on the functioning of the whole circuit and together combine to result in – often complex – emergent properties. In order to study communication within networks the researcher needs applicable, non-invasive and precise tools to interrogate the system. The current toolbox for the investigation of neuronal circuitry does not always allow the experimenter to manipulate neurons in a precise and cell-type specific manner. Optogenetic manipulations offer a welcome expansion to this toolbox (see Box 1 & 2).
Optogenetics opens up the possibility to selectively activate or inhibit distinct neuronal cell-types in complex tissue containing multiple cell populations. In most optogenetic experiments excitation and inhibition can be delivered with the temporal precision required to study the properties of a
biologically relevant circuit component. Optogenetic techniques create the possibility to influence activity in specific neuronal cells and even subcellular areas in a non-invasive and controlled manner, changing the field of electrophysiology and neuroscience as a whole.
As with any method optogenetics has its own specific limitations. Most of the limitations are caused by its heavy reliance on gene expression, molecular and dynamic properties of different opsins and light sensitivity of opsins and cells. Although many improvements have been made to the originally isolated opsins there are still obstacles to be overcome as well as definite limitations.
In this review the virtues and limitations of the optogenetic toolbox are discussed for a number of functional levels. The levels are, starting off with the smallest level: the subcellular level, small functional circuits and the global level. This review aims to pinpoint and illustrate the unique properties of the optogenetic approach and how it can be used to answer neuroscientific questions.
Levels of Circuitry
Circuits consist of connected neurons that integrate signals by their spatial and temporal distribution. The integrated signals can be regarded as to hold an informational value. This
underlies the emergent property that different integrated signals can evoke a different response in specific neuronal circuits. In this section circuit analysis is discussed over three different levels: the subcellular level, the level of small functional circuits and the global level. These circuit levels are used to highlight the uses of optogenetic approaches to questions that require a specific spatial and/or temporal resolution.
Subcellular level
Neuronal cells are extremely functionally regionalized (Hammond, 2008) in the sense that different sites on a neuron are critically endowed with specific tasks and thus present site specific properties. Differences in molecular trafficking, organelle and membrane (receptor) build-up and ion-channel permeability form the molecular basis of this specialisation. All of these properties can have a crucial influence on synaptic connectivity and hence on the functioning of neural circuits.
The actual result of differences on synaptic connectivity of different regions in a neuron are hard to determine for many stimulation and recording techniques in themselves have a significant impact on the intracellular milieu of a cell, as well as its membrane properties. This is a major challenge in the sense that many crucial and relevant cellular processes in neurons are hard to study using traditional techniques. Optogenetic approaches can circumvent some of these limitations, two examples of this are described in the following studies on subcellular circuit mapping (Petreanu et al. 2009) and the molecular changes in synaptic plasticity (Y.-P. Zhang & Oertner 2007; Y.-P. Zhang, Holbro, et al. 2008; Schoenenberger et al. 2010).
Subcellular circuit mapping
Determining the exact location and relative strength of a synapse can teach us more about the
functional connectivity within a cortical circuit (Douglas & Martin 2007). In the neocortex, pyramidal (PY) cells in different layers are thought to follow a preferential pattern of connectivity between neuronal populations. Different layers of PY cells are thought to act as weakly coupled
compartments (Mainen & Sejnowski 1996) that receive inputs from specific neuronal populations. A recent study shows how optogenetic approaches can be used to map these highly regionalized connections in the cortex (Petreanu et al. 2009).
The main goal of this study was to determine whether different cortical populations project preferentially to pyramidal cells in layer 3 and 5 (A/B). In a series of experiments ChR2 was
expressed in neurons of the ventral posterior medial nucleus (VPM), primary whisker cortex (M1), posterior medial nucleus (POm) and local excitatory neurons in pyramidal cell populations (layer 4 and 2/3). Followingly, blue laser photostimulation was used to excite the axonal projections of these neuronal populations at different layers of the barrel cortex. Simultaneously recorded somatic EPSC’s in the soma of layer 3 and 5 pyramidal neurons revealed the location and strength of the synapses from the target population onto the respective layers of the barrel cortex.
The results from this experiment revealed a high specificity in the subcellular organisation of synaptic connections. Different neuronal populations connect preferentially to specific layers of the neocortex – and henceforth onto specific subcellular parts of a PY cell – with axons from more central origins projecting to increasingly higher regions of the apical dendrites. Although this is mostly an observational study it does strengthen the view that neurons are highly functionally regionalized. This has a significant impact on the functional connectivity and properties of a neuronal circuit. Furthermore, this protocol provides an alternative for studying functional connectivity superior to studying axodendritic overlap, which isn’t always a good predictor of actual synaptic connectivity (Shepherd et al. 2005).
These studies provide a good example of how optogenetics can be used to improve upon existing protocols, making the mapping of functional connectivity between genetically and topologically defined neuronal populations more efficient and reliable. Additionally, one might argue that the use of a photostimulation grid – which stimulates a field of the cortex without knowledge of the
connectivity – allows for a more unbiased (blind) approach to stimulation as the experimenter isn’t the one preferentially targeting measurement sites.
Synaptic transmission
Long term potentiation (LTP) can be regarded as a relevant adaptation of different nodes within a circuit. LTP induction on a single postsynaptic site can even in itself be regarded as a small
functional circuit for it is an electrochemical feedback mechanism that can be influenced by outside input. LTP is a well studied phenomenon in neurophysiology and the tetanus stimulation protocol (Bliss & Lomo 1973) is a commonly used method to evoke LTP in synaptic connections. This
manipulation can be applied site-specifically and without high intrusion to the internal milieu of the synapse. However, the classic electrophysiological tetanus stimulation techniques do hold
restrictions. The main restrictions are the difficulty of isolating the responses in single synapses (Schoenenberger et al. 2010) and the wash-out of soluble proteins and second messengers, thought to be essential to maintaining LTP (Tanaka et al. 2008; Chen et al. 2001).
These restrictions impose limitations on the type of processed that can be studied. One particular interest goes out to the molecular and structural changes that are induced by LTP on a time scale relevant to protein production and morphologic adaptation. Whereas regular protocols need to be completed within 5-10 minutes, the low invasiveness and spatial precision of novel optogenetic ChR2-mediated protocols can significantly broaden the scope and time frame of LTP research. In a number of consecutive studies by the Oertner group (Y.-P. Zhang & Oertner 2007; Y.-P. Zhang, Holbro, et al. 2008)(review (Schoenenberger et al. 2010)) the molecular and structural (volumetric) changes in synaptic plasticity and LTP were studied using different ChR2-mediated protocols. The first study (Y.-P. Zhang & Oertner 2007) set out to determine whether ChR2 excitation in the presynaptic terminal would lead to presynaptic glutamate release and, coupled with postsynaptic depolarization, to increased synaptic connectivity – as LTP is supposed to work. After initial success to create light evoked action potentials in the presynaptic terminals, selective influx of calcium at the
postsynaptic site indicated glutamate release and induction of LTP. Calcium influx was visualized using Ca-sensitive dyes. This compellingly showed an increase in Ca2+ levels in single targeted spines as a result of presynaptic activity.
It has been shown that αCaMKII is both necessary and sufficient to induce synaptic plasticity (Lledo et al. 1995). However, the question whether the synapse selective maintenance of LTP is also
dependent on αCaMKII has remained unanswered (Chen et al. 2001). To study these second messenger dynamics, a larger window of time is necessary, an optogenetic protocol created a possibility to study this.
The second study was set up to track changes in dendritic spine volume and molecular trafficking in CA1 and CA3 pyramidal neurons during LTP (Y.-P. Zhang, Holbro, et al. 2008). To this extent a different protocol for LTP was used where ChR2 was expressed at the postsynaptic site (pyramidal cells in CA1), coupled with synaptic 2-photon glutamate uncaging. The ChR2 positive pyramidal cells in CA1 were filled with a cytosolic (red) dye as to create a reference for synaptic volume. When αCaMKII is fluorescently labelled (green), the ratio between red and green fluorescent signal reveals information on the concentrations of αCaMKII or Ca2+. The results from this study showed a
persistent increase (lasting up to 40 min after LTP induction) of soluble αCaMKII only at the spines persistently increased in volume and not at neighbouring unaffected spines, indicating a role of αCaMKII in maintaining LTP.
These applications of opsins combined with glutamate uncaging broaden the possibilities for studying synaptic plasticity over an extended period of time in a non-invasive manner. In a certain sense these approaches are thus more suitable than others for studying the long-term subcellular molecular effects of synaptic transmission on a single synapse level. However, important questions need to be raised on the overall applicability and validity of these approaches. It remains uncertain whether either of these protocols (pre- or postsynaptic ChR2) is a truly physiological equivalent of LTP. Influences on diffusion and cellular trafficking of molecules might impose restrictions to the true physiological validity of these approaches. Even though caged glutamate is used to locally target receptors, uncaged glutamate still might travel to other, nearby, glutamate receptors. Models of glutamate dynamics do suggest that glutamate diffusion is restricted within the narrowest parts of the synaptic cleft, but in wider areas its diffusion is similar to that in free solution (Cory & Glavinović 2006).
The ChR2 opsin itself might also influence the physiological validity of these synaptic plasticity studies for ChR2 is a non-selective cation gated channel, which can carry Ca2+ ions as well. This has been shown to result in significantly altered calcium dynamics as compared to the physiological situation (Schoenenberger et al. 2010), undoubtedly having a profound impact on intercellular molecular signalling. This issue could be resolved by using an altered opsin that selectively conducts Na+ ions. The basic approaches for creating such an opsin have been described previously
Small functional circuits
Multiple neurons forming a network conveying information can be regarded as small functional circuits. As to determine the true contribution of a (population of) neuron(s) it is essential to specifically study targeted alterations in activity with as little collateral stimulation to other
(possibly-related or influential) neuronal populations. The isolated contribution can then be taken as a starting point to further unravel the functioning of a larger circuit. A number of examples of the experimental benefits in studying small functional circuits will be described in this chapter. The olfactory circuit
Optogenetic approaches can supersede the cell specificity and locality of electrophysiological techniques. In a study by Arenkiel and colleagues on the olfactory system (Arenkiel et al. 2007) the benefits of the in vivo genetic isolation of specific cell types and the site specifically evoked responses were first demonstrated. In this study the leading question was to what extend mitral cells - which are amongst other cells involved in odour sensitivity - form an integrative network amongst them. Mitral cells are hypothesized to integrate information in an intercellular manner, creating a
calculating network that ‘gates’ information to the piriform cortex. This would mean that increased lateral excitation of mitral cells would lead to an integrated and therefore altered overall response. In this study a transgenic mouse expressing a ChR2 under the Thy1 promoter making expressed ChR2 specifically in the CNS. Currents were evoked in surface mitral cells in vivo with a light source conducted by an optic fibre. Increasing numbers of cells were excited – by increasing the diameter of the beam – with lgith as to compare the responses of individual mitral cells under increased lateral excitation of neighbouring cells. This revealed that, contrary to the aforementioned prediction, the activity of single units was not significantly altered by lateral interactions in the olfactory bulb. Another hypothesis – stating that there is a threshold for the relay of signalling between the populations of mitral cells to the piriform cortex was tested in a second experiment. Using similar optogenetic control over the input, the relay within the circuit mitral cell to the piriform cortex circuit was studied. Simultaneous excitation in the mitral cell population and measurement in the piriform cortex revealed that the relay of signalling from the mitral cells to the piriform cortex is strictly relative to spot diameter of excitation. A distinct number of mitral cells is required to evoke a response in the piriform cortex. These findings support the threshold hypothesis and reveal a
meaningful and functional integration of neuronal signalling.
The optogenetic gain in this study is the cell-type specific control over mitral cells within the olfactory system, opening up new opportunities of studying neural circuits in the olfactory circuit. Especially, the targeting of these neuronal populations with a high control over the number of excited cells was formerly unfeasible with chemical or electrophysiological techniques (Arenkiel et al. 2007). Extending on the possibilities optogenetics offers it might also be interesting to study the ceiling for activation. One might imagine that a certain ceiling of activation can be reached in terms of odour sensitivity. This could be represented by, for example, a maximum number of excited cells that still results in a different cortical signal.
Though the stimulation of an area with increasing diameters might be very precise, this study should still cannot be considered to be a quantitative study. Only the number of cells (without exact measurement) and not the frequency or strength of the signal were subject of investigation. This tells us little about the effect of differential spatial and/or temporal summation of imputs and their
respective representation in the network. Furthermore, optogenetic approaches for studying the responsiveness of the olfactory system and its subsequent signalling clearly allow for remarkable cell specificity. However, the physiological relevance of evoking simultaneous responses in
increasing numbers of neighbouring cells might not be as high as suggested. Due to the nature of the experiment, the relation to the physiological situation remains unclear.
Oscillatory circuits
Cortical gamma oscillations (20 – 80 Hz) in are thought to have a functional impact on processes ranging from sensory information processing to higher cognitive functions. One long held hypothesis is that fast-spiking (FS) interneurons form a neural circuit that is responsible for the development of gamma oscillations (Traub et al. 1997). It has been shown that networks of FS cells can provide synchronous inhibitory postsynaptic currents (IPSC) to local excitable cells in the prefrontal cortex (Hasenstaub et al. 2005) and network models suggest that these IPSC’s could drive gamma oscillations (X. J. Wang & Buzsáki 1996; Traub et al. 1997). However, causal in vivo
investigation of these hypotheses in specific neural circuits has been unfeasible due to difficulties in stimulating these predefined neuronal (FS) populations.
In the two studies described below, optogenetic approaches were used to specifically target FS cells and to test whether these populations have a role in driving gamma oscillations in the mouse cortex (Cardin et al. 2009; Sohal, F. Zhang, Yizhar & Deisseroth 2009b). The first of these studies causally tested the hypothesis that synchronous activity of FS interneurons induces gamma rhythms in the mouse barrel cortex (Cardin et al. 2009). The second study further investigated the role of different frequencies of FS spiking and inhibition of FS cells on the emergent gamma oscillations within the FS – pyramidal cell circuit in the neocortex (Sohal, F. Zhang, Yizhar & Deisseroth 2009b). Both of these studies further elucidate the role of the FS interneurons and attempt to determine whether they are sufficient and necessary to promote gamma oscillations.
Driving γ-oscillations
Sensory input from the whiskers of a mouse evoke stimulus related potentials in the mouse barrel cortex (Crochet & Petersen 2006) through primary axons reaching the barrel cortex via the thalamus. Axonal projections of inhibitory FS interneurons, project directly onto specific populations of regular spiking (RS) neurons in the mouse barrel cortex. According to cortical in vivo recordings and
neuronal modelling mentioned above, driving synchronous IPSC’s in the FS cells should be able to evoke gamma oscillations as measured in the Local Field Potential (LFP) within the barrel cortex. To this effect FS cells expressing ChR2 were stimulated with light pulses at 40 Hz. When activated with light populations of FS cells created synchronous IPSP’s. These IPSC’s inhibited spiking of RS neurons in layers 2/3 and 4 of the barrel cortex via direct synaptic inhibition. The gamma oscillations
were found to gate sensory input from whisker stimulation only in distinct phases of their
oscillation. The resulting signal in the RS cells was measured and the gamma phase was shown to modulate the strength and timing of the signal, with a diminished response under the peak of
inhibition and larger responses under the opposite phase (Cardin et al. 2009). These findings indicate that the evoked gamma oscillations significantly alter the amplitude and precision of the relay of sensory information to the cortex.
Altering gamma-oscillations
Building on the observation that stimulation of FS cells can drive gamma-oscillations the question becomes whether it is possible to manipulate the frequency and relative strength of these
oscillations. In the mouse neocortex a microcircuit consisting of parvalbumin (PV) expressing FS interneurons receives axonal excitatory projections from pyramidal (PY) cells. In this circuit PV interneurons are thought to be involved in driving gamma oscillations which gate input from PY cells.
In a study with bidirectional optogenetic control Sohal and co-workers drove inhibitory eNpHR expression in PV interneurons in the prefrontal cortex of transgenic mice (Sohal, F. Zhang, Yizhar & Deisseroth 2009b). Optogenetic stimulation of pyramidal cells (PY), expressing ChR2 under an αCaMKII dependent promoter, revealed that non-rhythmic stimulation of PY cells with blue light drove activity in interneurons and was sufficient to generate emergent gamma-oscillations. As expected, yellow light-induced inhibition of PV interneurons suppressed the power of the gamma oscillations measured in the LFP.
To further investigate the role of PV interneurons in creating gamma oscillations in downstream PY cells, a feedback inhibition circuit was imposed on the network. In this setup dynamic-clamp was used to create simulated EPSCs (sEPSCs) in PY cells. Simultaneously, light flashes based on the spiking patterns of local PY cell were delivered to ChR2-PV neurons. The question was whether these inputs could impose different oscillation frequencies. Enhanced gamma oscillation power was observed which was elicited by dynamic sEPSCs, but only in the gamma-range, indicating
selectivity in gamma oscillation generation in this circuit. Under all frequencies imposed on the PV interneurons the output spike rate in PY cells increased with higher sEPSC rate, but notably the gain between input and output was higher under gamma oscillations.
These results provide compelling evidence that PV interneurons can generate cortical gamma oscillations within these defined circuits, and that these oscillations have an impact on sensory relay and processing of different inputs within these cortical circuits. The experiments were performed in vivo under active network conditions and optogenetic approaches allowed for the targeted
stimulation of neuronal cell types within a relevant range of frequencies.
The genetic basis for manipulating distinctive neuronal populations is held within the use of specific promoters and local application of viral vectors. Interestingly, the parvalbumin promoter fragment used by Sohal and colleagues, only targets the specific GABAergic subset of (fast-spiking)
parvalbumin expressing interneurons (Meyer et al. 2002). To achieve functional expression levels, a Cre-dependent, double floxed, specific targeting of PV interneurons was necessary to achieve detectable and functional levels of ChR2-YFP expression This still only drove expression in 41 (+/- 7) % of PV immunoreactive cells (Sohal, F. Zhang, Yizhar & Deisseroth 2009a), indicating that for – unknown – reasons less than half of the targeted PV population was hit. It is important to regard this limitation for it cannot be determined what causes this specific expression pattern, and it could potentially influence the outcome of the experiment.
A closing remark on these studies should be that both studies clearly and compellingly show that it is indeed possible to impose the desired activation on, and generate specific activation patterns in these neural circuits. In turn the simultaneous control over input and output data streams enables the researcher to enact the activity that is desired based on the limitations of the targeted cells and the techniques used.It should be noted that this only goes as far as to show that it is possible to manipulate networks in such a fashion, not that it is a naturally occurring phenomenon. Parkinsonian circuitry
In recent years deep brain stimulation (DBS) has developed into a serious therapeutic option for treating advanced Parkinson’s disease (PD) (Weaver et al. 2009). Although this high frequency stimulation clearly alleviates PD symptoms is has remained unclear what the principal reasons for its effectiveness are. In DBS treatment of PD electrodes implanted in the subthalamic nucleus (STN) deliver high frequency (>90 Hz.) electrical stimulation to this area.
Due to the heterogeneous cellular composition of the STN it remains unclear which of the neuronal populations in the STN is responsible for the therapeutic effect of DBS in PD. In a study by
Gradinaru and colleagues (Gradinaru, Mogri, Thompson, Henderson & Deisseroth 2009b) an optogenetic approach was used to investigate the role of specific neuronal populations within the STN in DBS alleviation of PD symptoms. Different neuronal populations of the STN: excitable STN cells, astroglia and afferent fibres originating in the motor cortex, were subjected to opsin mediated excitation or inhibition and the resulting effects were measured on an electrophysiological and behavioural level.
First, in two experiments the excitable cells in the subthalamic nucleus (STN) were inhibited by and excited with, respectively, eNpHR and ChR2, expressed under the glutamatergic neuron specific αCaMKII promoter. In both cases excitation of cells and inhibition of spiking could be manipulated with the appropriate light pulses, resulting in increased and decreased spike frequencies. However, no significant alleviation of PD symptoms was observed during the behavioural testing, as studied with a rotation assay, head position bias, climbing and track length (Gradinaru, Mogri, Thompson, Henderson & Deisseroth 2009a). This indicated that a principal hypothesis suggesting a role for excitable cells in the STN could not be confirmed nor falsified. A second hypothesis stated that excitation of astroglia could be responsible for the alleviation of PD symptoms in DBS. To test this hypothesis, astroglia within the STN expressing ChR2 were successfully stimulated with trains of blue light. However, this stimulation was unsuccessful in alleviating the behavioural symptoms of
PD. Although preliminary validation of excitation and inhibition in cellular recordings and consequent validation of light penetration revealed a clear electrophysiological response on these neuronal populations, they did not alleviate PD symptoms.
A final experiment was performed to test the hypothesis that stimulation of the afferent fibres from layer 5 of the motor cortex, that project to the STN, can bring about the ameliorating effect found in DBS. High frequency optical stimulation of ChR2 expressing neurons in M1 altered the spiking pattern of the afferent fibres and alleviated PD symptoms. These results were recurrent over a line of different behavioural tests and suggest a major role for afferent fibres of excitatory cells from layer V of the motor cortex in the effectiveness of DBS.
The use of optogenetics in these experiments is essential for its high cellular specificity forms the unique opportunity to test separate hypotheses on the contributions of neuronal populations in DBS alleviation of PD symptoms. Even though no permanent conclusions can be drawn on the impact and functions of all circuit components that might be involved, major advances have been made in understanding determining crucial contributors to PD circuitry. The fact that stimulation of afferent fibres from M1 could bring about the ameliorating effect on PD in this animal model of PD, doesn’t guarantee that this finding holds for the treatment of humans as well. However, should these findings correspond directly to the human situation, deep brain stimulation in the STN could be matched with “shallow brain stimulation” in M1.
This study sheds light on which STN neuronal populations can be targeted to alleviate PD
symptoms. It remains important to determine whether optical stimulation is comparable to electrical stimulation applied in high-frequency DBS. For one, HF stimulation at 130 Hz in STN excitable cells increased the spike frequency from 41 (±2) to 85 (±2) Hz. In an earlier DBS study using electrodes, HF stimulation at 140 Hz caused the spike frequency in the STN to fall by 77% (Welter et al. 2004). It cannot be determined from these two separate experiments what causes this difference in evoked spiking patterns. However, it does go to show that interpreting the findings from an optogenetic study requires a thorough comparison with the known electrophysiological effects. Especially because different opsins have unique properties in terms of reliability, conductance and inactivation dynamics (Andre Berndt et al. 2011).
The Amygdala
The amygdala forms an integrative circuit mediating multiple processes including emotional responses to external stimuli and emotional memory (LeDoux 2003). Studying the intricate network within the amygdala has been a major challenge because the subnuclei are functionally and
morphologically heterogeneous. In two recent studies the functional microcircuit of the amygdala and the amygdala–nucleus accumbens circuitry were subject of investigation. The researchers show that it is possible to identify the properties of distinct neuronal populations and specific projections in amygdala circuitry using optogenetics (Tye, Prakash, Kim, Fenno, Grosenick, Zarabi, Thompson, Gradinaru, Ramakrishnan & Deisseroth 2011a; Stuber et al. 2011).
The basolateral amygdala – central amygdala circuit: control of anxiety
In the first study the role of amygdala circuitry in the control of anxiety was studied. The amygdala microcircuit consists of the basolateral amygdala (BLA) and the central amygdala (CeA) and the CeA is subdivided into a lateral and medial part (CeL and CeM), both receiving input from the BLA. The CeA microcircuit is known to mediate fear and anxiety but it remains unclear which neuronal populations and projections constitute this involvement (Kalin et al. 2004).
In a study by Tye and colleagues (Tye, Prakash, Kim, Fenno, Grosenick, Zarabi, Thompson,
Gradinaru, Ramakrishnan & Deisseroth 2011a) the somata of the BLA and the projections from the BLA to the CeL were stimulated/repressed with light. In the first set of experiments ChR2 was expressed in glutamatergic neurons in the BLA. Using a specific selective illumination technique (Tye, Prakash, Kim, Fenno, Grosenick, Zarabi, Thompson, Gradinaru, Ramakrishnan & Deisseroth 2011b) BLA somata or BLA projections in the CeL were stimulated during an anxiety related task (open field/elevated plus maze). When BLA somata were stimulated during the task, no difference in anxious behaviour was observed. When BLA projections in the CeL were stimulated during the task, the animals showed a significant reduction in anxious behaviour.
This manipulation revealed that the activation of axonal projections of the BLA to the CeL, rather than of BLA somata themselves, underlies the anxiolytic effect in this mouse model for anxiety. This suggests that excitatory projections from the BLA to the CeL specifically play a role in the neural circuitry of anxiolysis. The inverse of this effect was observed when eNpHR3.0 was expressed and drove inhibition in BLA glutamatergic projections. Selective inhibition of these projections was sufficient to induce anxious behaviour in WT mice. Electrophysiological data confirmed that the stimulation BLA projections evoke spikes in the CeL with high fidelity and that manipulation of these projections alters the activation pattern in the CeL.
To conclude, n this microcircuit BLA projects to both the CeL and the CeM directly, and the CeM projects to CeL. This creates a difference between direct and indirect inputs in the CeM which could form a modulatory circuit. BLA-CeL synapses are thought to be able to block CeM spiking through feed-forward inhibition. To test this hypothesis BLA-CeL expressing ChR2 were stimulated with blue light. After ChR2 stimulation a subsequent reliable inhibition spiking was observed in the CeM revealing significant modulation of signal within the amygdala microcircuit.
Amygdala projections on the nucleus accumbens: controlling reward seeking behaviour
Amygdala projections to the nucleus accumbens (NAc) are known to be strengthened during cue-reward learning (Tye et al. 2008). One projection that is possibly responsible for this type of
behaviour is the BLA – NAc projection. In a study by Stuber and colleagues (Stuber et al. 2011) ChR2 was expressed under an αCaMKII promoter in glutamatergic BLA neurons projecting to the NAc. Optical stimulation of BLA projections in the NAc drove spiking in the NAc and significantly increased motivated responses. Motivated responding was measured in a behavioural experiment
where a nose poke in a certain area of the cage produced optical stimulation to BLA-NAc projections.
In another experiment, bidirectional control over this behaviour was demonstrated by stimulating eNpHR3.0 in the BLA–NAc projections during a cue-reward learning paradigm. In this task a 5 s tone and a light cue predicted the delivery of sucrose at a receptacle. The number of licks per session was taken as a measurement of cue-related learning. Optical inhibition of the BLA-NAc projections significantly reduced spiking activity and prevented cue-reward learning over multiple conditioning sessions. The added value of the optogenetic approach in this study is the cell type specific approach coupled with the relatively low-invasiveness of the stimulation. This allows for repeated and
reversible optical modulation of a specific type of projection involved in (amygdala) circuitry. Pharmacological investigation of the circuit with the AMPA receptor antagonist CNQX revealed that optically evoked EPSCs in the NAc are mediated by AMPA-R via glutamate release. Selective
dopamine 1 and dopamine 2 receptor (DxR) blocking also revealed that the postsynaptic effect on medium spiny neurons is modulated by the dopamine 1 receptor and not D2R. These experiments show that optogenetic approaches can be compatible with pharmacological manipulations and help to elucidate the role of different chemical pathways in neural circuitry, in a synapse specific manner. Both of these studies utilise the optogenetic opportunities to target specific neuronal populations and synapses. The cell-type specific approach and the site specific optical stimulation combined, offer projection specific (bidirectional) control over electrophysiological and behavioural responses. Stimulation of certain projections is shown to be sufficient to impose changes in activation pattern and behaviour within these animal models. However, apart from showing that a type of activation is able to alter the functioning of a network, this doesn’t necessarily mean that it hold this function in the in vivo situation.
Dissecting emotional memory circuitry
It is widely known that emotional states can significantly alter the strength of a memory in both animal and human models of memory. Emotional memory is modulated by an overwhelming amount of factors but it is widely accepted that a complex interplay of amygdala and hippocampal structures is required for emotional memory formation (Roozendaal & McGaugh 2011). However, the underlying mechanisms and functional connectivity involved in this neural circuitry are not fully understood.
Retrograde tracing studies have shown that axons originating from the soma of BLA neurons, which are mainly glutamatergic, project mainly to the CA1 region of the hippocampus (Pitkänen et al. 2000). Since the amygdala and hippocampus enter a (complex) interplay during emotional memory formation, this interaction is likely to be dependent on these projections, incidentally this is a common assumption in most emotional memory models (Roozendaal et al. 2009). To test this hypothesis, concerning a specific cell type projecting to a specific brain region, optogenetic methods might allow for an unique opportunity as presented in appendix 1 (Meijer et al. 2012).
Epilepsy
The paroxysmal depolarizing shift (PDS) is a cellular mechanism present in specific neuronal populations that underlies synchronous firing of neurons which in turn can lead to epileptiform activity (Traub & Wong 1982). Suppressing specific types of neuronal firing within identified neuronal circuits could inhibit the formation of PDS and suppress epileptiform activity. In order to do this, it is important to know which populations of neurons drive the epileptiform activity. In an optogenetic study by Tønnensen and colleagues (Tønnesen et al. 2009) principal excitatory cells in the hippocampus (CA1 and CA3) were thought to be involved in the creation of PDSs. If this is the case, optogenetic inhibition of principal excitatory cells should reduce the occurrence of epileptiform activity. To this extent NpHR was expressed under the αCaMKII promoter in the CA1 and CA3 region of the hippocampus. Electrophysiological stimulation train-induced bursting (STIB) in CA1 and CA3 successfully created PDSs. These strength and occurrence of PDSs were highly mitigated during NpHR-mediated inhibition which in turn silenced epileptiform activity in CA3. Optogenetic suppression of primary hippocampal neurons was shown to be able to inhibit focal epileptiform activity within the CA-circuit. The authors claim that their manipulation of these neuronal populations is non-invasive but this might be too optimistic. Whereas no ‘major’ alterations to membrane properties are reported there are differences to the regular excitability (altered I-V relationship) of CA3 pyramidal cells due to NpHR expression (Tønnesen et al. 2009). The observed IPSC’s in these cells were not altered, which according to the authors, could indicate that Cl--pumps can keep up with the newly instated equilibrium. The possible increase in Cl
-conductance in itself would be an adaptation to the physiological situation and reduces the in vivo relevance of this study. This finding warns us that expression of opsins in itself can have an effect on the circuit studied that should be accounted for.
The results from this study are promising in revealing a major player in the hippocampal network involved in epileptiform activity. However, the fact that widespread inhibition of excitatory neurons in the hippocampus ameliorates epileptiform activity is hardly surprising. The real question remains what the exact properties of a network are that make a brain ‘epileptic’, i.e. which cells critically contribute (in what way) to the epileptiform activity?
Global circuits
Global levels consist of multiple functional circuits that communicate with each other to share and relay information on distances spanning the entire brain. Global circuits are studied using
techniques such as electro-encephalography (EEG), magneto-encephalography (MEG), functional magnetic resonance imaging (fMRI), and positron emission tomography (PET). These techniques reveal activity in global circuits throughout the cortex and in subcortical structures. More
specifically they detect a signal on a regional level, combining signals over a range of neuronal populations and a vast number of cells. These detection techniques are by their nature indirect markers of activation, directly limited by the type of signal they detect (electric, magnetic, radiation).
BOLD optogenetics
The Blood Oxygen Level Dependent (BOLD) signal is a widely used marker for activity employed in fMRI. Even though the BOLD signal is widely used as a readout for neuronal activation, there still is considerable debate on whether the BOLD response is directly related to neuronal activity
(Logothetis 2008). The primary assumption in fMRI is that an altered BOLD response, caused by a regional change in the magnetic properties of haemoglobin, marks the use of oxygen in that
particular area. In turn, the use of oxygen is considered a metabolic marker for neuronal activity at this location. However, as the BOLD response is de facto a secondary signal, is has remained unclear whether neuronal activity is truly directly related to the BOLD response. By combining optogenetics and fMRI, a combination now dubbed ofMRI; a new window opens for testing the connection. In a fMRI study performed on a genetically altered rat researchers found that optical excitation of ChR2 expressing cortical pyramidal cells created a BOLD response in the rodent brain (Lee et al. 2010). This BOLD signal was robust, reproducible and linked directly to the duration of the light pulse. The BOLD signal measured resembles the typical BOLD response, marked by an onset delay, an initial dip and an undershoot, all within the expected time frame. Simultaneously recorded signals from the thalamus marked downstream activity as a result of cortical stimulation. This activity arguably reveals activity relayed from the cortex to the thalamus through thalamic projections. A follow up experiment with optical stimulation in the thalamus showed that this connection is reciprocal: thalamic stimulation elicited a BOLD response in the soma in the cortex. These experiments provide a proof of principle that activition in neuronal populations reliably elicits a BOLD response. This provides a compelling empirical evidence for the causal connection between brain activity and the hemodynamic response (Palmer 2010). However, it is not possible to safely attribute the elicited BOLD response to the stimulated neuronal populations (Logothetis 2010). The complicated and extensive nature of neuronal networks urges to consider the possible resulting activation of feedback and –forward processes, activity in nearby (supporting) cells and other local perisynaptic processes that might underlie the measured signal. The measured signal should
therefore still only be considered as the local summation of activity – being inhibitive or exciting – of the heterogeneous cellular population present in each voxel.
A follow-up study, also proposed in a recent review by the author of the original article (Lee 2011), could delve deeper into the heterogeneity of each voxel by studying the BOLD response elicited after optic inhibition with NpHR. When, in a bidirectional controlled study, excitation and inhibition of the same cellular population would result in a different BOLD response, this would suggest that the BOLD signal could distinguish between basic differences between inhibition and activation. This still cannot definitively indicate the relative contribution of other cell types within the circuit but is does open the possibility to use combinatory approaches (including genetic knockout,
pharmacological intervention, etc.) to dissect the BOLD signal into its relative constituents. An important beneficial property of the ofMRI technique is the causal nature of this approach to global connectivity of the brain. Questions on the level and nature of connectivity have been major issues in the field of fMRI mapping of neural circuitry since its development (Logothetis 2008). Initially MRI and Diffusion Tensor Imaging (DTI) have shown to be adept at dissecting connectivity at an anatomical level, able to distinguish which brain regions have an anatomical connection. In fMRI a functional connection is demonstrated when a significant correlation between activation in different regions of the brain is demonstrated through the use of (elaborate) statistical analyses. Importantly, this functional connectivity still only indicates synchronous activity and cannot verify the origin or direction of the connection.
Effective connectivity in fMRI is concerned with sequential activation of brain regions (Horwitz et al. 2005). When determining effective connectivity one sets out to determine whether the activity in one brain region can reliably predict activity in another region. By its nature, effective connectivity analyses rely on an underlying statistical model which makes it important to validate findings in a truly causal manner (Friston 1994). Optogenetic fMRI (ofMRI) can offer a means to validate findings and theories derived from ‘regular’ fMRI. This would allow the researcher to causally manipulate activation within a supposed global circuit, moving away from the highly statistical nature of fMRI. Whether the use of ofMRI heralds the era of causality in MRI will remain much debated for it should be considered that the BOLD signal itself remains a secondary signal and it has serious physical and biological constraints (Logothetis 2008). The BOLD signal is a temporally precise but slow response to the activity of a large amount of cells and cannot distinguish between different types of neuronal activity. Therefore hypotheses testable with ofMRI will, until significant improvements in spatial resolution are achieved, be limited to the level of global circuitry.
General Discussion
On all of the circuit levels described optogenetic approaches have shown to open new windows of opportunity for neuroscience research. Circuit analysis will benefit from the newly created
possibilities but it is important to judge each technique by its merits. A short review will be
presented here, followed by an overall critical review of the expression of opsins and optodynamics. At the subcellular circuit level optogenetics has been applied to test hypotheses on subcellular organization of synaptic connections (Petreanu et al. 2009) and the molecular mechanisms of
synaptic transmission (Y.-P. Zhang, Holbro, et al. 2008; Y.-P. Zhang & Oertner 2007). To this level of circuit analysis optogenetics provide the level of cell type and site specific stimulation essential to these experiments. The relatively low invasiveness creates a larger time frame in studies on synaptic plasticity.
Most of the optogenetic studies published so far are situated at the level of small functional circuits, ranging from informational analysis of γ-oscillations (Cardin et al. 2009; Sohal, F. Zhang, Yizhar & Deisseroth 2009a) to behavioural control by manipulation of amygdala circuitry (Tye, Prakash, Kim, Fenno, Grosenick, Zarabi, Thompson, Gradinaru, Ramakrishnan & Deisseroth 2011a; Stuber et al. 2011). The bidirectional control of specific neuronal populations and good in vivo applicability make it that optogenetic tools are suitable for studying a wide range of neuroscientific questions.
Optogenetic fMRI has been applied to study functionality in global circuits (Lee et al. 2010), revealing more about the nature of the BOLD signal. In this field optogenetics may form an additional possibility to validate findings and hypotheses the BOLD response itself and/or
functionally define different brain regions on a regional level. However, since the first application of ofMRI there have been no follow-up studies using this technique, revealing some reluctance of neuroscientists to endeavour into this field.
Expression of opsins
The genetically driven - population specific expression of opsins and its consequent manipulation can be regarded as providing both opportunities as limitations. Genetically targeted expression can work as a magic bullet to target specific populations of neurons, thereby creating a molecular light switch to control specific components in neural circuitry. The genetic basis also forms the first and final limitation to driving expression in certain populations for it is limited to genetic distinctions. Even though there are ongoing improvements in cell type and even cell structure specific expression (Lewis et al. 2009) of genes using conditional promoters this will not overcome all obstacles.
Targeting neuronal populations on the basis of morphology or specific functional properties might be desirable but it doesn’t always align with the possibilities of the (opto-)genetic toolbox.
The cellular regulation of expression of opsing remains inscrutable in some cases and it could very well be possible that expression of opsins unexpectedly drives other cellular processes. For instance, even when eNpHR is expressed under a strong EF1α promoter in parvalbumin interneurons still only 41% of these cells were found to be expressing eNpHR (Sohal, F. Zhang, Yizhar & Deisseroth
2009a). On a different level, the localization of opsins in a neuron has been shown to differ
significantly between different opsins (Schoenenberger et al. 2010), for example different forms of NpHR are expressed differently at specific sites in the cell, due to the nature of the protein (Zhao et al. 2008). Improvement of (control over) expression patterns is necessary and ongoing (Gradinaru et al. 2010; Deisseroth et al. 2006; Gunaydin et al. 2010; Mancuso et al. 2011), but in this light expression patterns should be sought after to fit the requirements of the experiment.
Optodynamics
Not every experimental set up requires opsins with perfect dynamic properties in terms of
excitation, peak current and range of spiking frequencies. When hypotheses tested essentially need dynamics comparable to the physiological situation, the different properties of opsins could become a deciding factor for in vivo validity and succes. Microbial opsins can simulate certain properties of regular membrane channels and pumps but the reliability of driving, peak and shape of the action potential and excitability some issues remain.
The reliability of light train induced spiking, an important parameter in oscillation experiments, differs over several ChR2 variants (Gunaydin et al. 2010; Andre Berndt et al. 2011). The reliability poses a constraint to such other variants as the use of different frequencies in stimulation. ChR2 mutants such as ChIEF and ChETA (Gunaydin et al. 2010) are a clear improvement over the ChR2 WT (Lin 2011) but each of these holds its own restraints in terms of e.g. current amplitude or
opening rate. Testing the reliability of driving spike trains will remain an important control in these experiments.
In studies focusing on the ChR2 evoked action potential it has been revealed that the optical induced action potential (AP) differs significantly from regular APs on two key aspects. Firstly, ChR2 kinetics are slow as compared to regular AP’s (Bamann et al. 2008; Schoenenberger et al. 2010). The relatively shallow slope of the photocurrent and late maximal polarization outlast the K+ current which results in an after-depolarization – in turn imposing limits on a following excitation. Secondly, calcium imaging experiments have revealed that the calcium transients in the light induced AP are
significantly larger than those of regular APs, which is a possible confounding factor in molecularly focused experiments (Schoenenberger et al. 2010).
Optical stimulation itself also poses problems in terms of possible confounds inherent to light stimulation of cells. Blue light is known to have two problems: a low penetration in tissues, and that it is detrimental to cells after prolonged exposure. Red-shifted and far red options offer possibilities for deeper penetration or for experiments that require a longer time span (Gradinaru et al. 2010; F. Zhang, Prigge, et al. 2008). Additionally, mutations that increase light sensitivity widen the have also been shown to overcome some of the problems but also have negative effects on kinetics (André Berndt et al. 2009). Still, much remains unclear about the reliability of opsins in the scale of hours, as well as the effect of prolonged exposure of tissue to light. Up to this moment the longest reported exposure of eNpHR3.1 transfected cells to red light is 30 min (Goshen et al. 2011), what happens after this is unknown. However, solutions such as the novel chimearic light sensitive, K+-permeable
glutamate receptor HiLighter (Janovjak et al. 2010) might provide an alternative. In the case of HiLighter neuronal inhibition is reliably switched on by a brief blue pulse of light, and turned off by a green pulse of light. This removes the necessity of permanent exposure of tissue to light.
Concluding remarks
In general, optogenetics provides the neuroscientist with useful tools for answering specific
questions in circuit analysis. Besides offering a novel approach to answering certain neurobiological questions optogenetic methods can also, not unimportantly, offer more convenient ways of
controlling manipulation. As with most methods, optogenetics has its specific virtues and limitations. Coupled with pharmaca in cue-reward learning (Stuber et al. 2011), dynamic voltage clamping (Sohal, F. Zhang, Yizhar & Deisseroth 2009b) and substance sensitive dyes (Y.-P. Zhang, Holbro, et al. 2008) the range of possibilities grows significantly.
It is important to carefully investigate, with each new application of optogenetics, whether the expression of opsins and the results of stimulation or inhibition are suitable to the experiment. The range of options is ever increasing for uch effort goes into creating reliable and tightly controllable opsins ((Lin 2011) overview), each with unique properties suited to different tasks.
In circuit analysis optogenetics now allows for the investigation of parallel or transcending theories on neural circuitry. Probing groups of neurons with specific and differentiated patterns of activity will teach us more about the relative contributions of parts of a circuit and how the system responds to alterations in excitation and inhibition. Multiple recordings and patterned excitation in small functional circuits will allow the researcher to compare modelled hypotheses to biological circuitry in a controlled system.
A final remark should be made on the in vivo validity of using optogenetic methods to excite or inhibit populations of neurons. For example, simultaneous stimulation of large numbers/areas of (neuronal) cells with light probably isn’t a good approximation of the physiological situation. Also, the mere fact that a certain network state can be induced in a circuit doesn’t mean that this state can be found in a normal network In other words: can overwhelming simultaneous stimulation of large groups of neurons/glia or on the other hand tightly controlled feedback circuits truly resemble the in vivo situation?
All in all, optogenetics clearly move neuroscience forward, but abandoning the temporal precision of electrophysiology or the reliable properties of pharmaca is not issued. Optogenetics will serve within the entire spectrum of possibilities and evoke initiatives to create combinatory approaches using drugs, clamping, multiple recordings and modelling to get closed loop control over neuronal circuitry; ever closing in on the most complex question in biology.
Bibliography
Arenkiel, B.R. et al., 2007. In vivo light-induced activation of neural circuitry in transgenic mice expressing channelrhodopsin-2. Neuron, 54(2), pp.205-18.
Bamann, C. et al., 2008. Spectral characteristics of the photocycle of channelrhodopsin-2 and its implication for channel function. Journal of molecular biology, 375(3), pp.686-94.
Berndt, Andre et al., 2011. High-efficiency channelrhodopsins for fast neuronal stimulation at low light levels. Proceedings of the National Academy of Sciences, 108(18), pp.10-15.
Berndt, André et al., 2009. Bi-stable neural state switches. Nature neuroscience, 12(2), pp.229-34.
Bliss, T.V.P. & Lomo, T., 1973. Long-lasting potentiation of synaptic transmission in the dentate area of the aneasthetized rabbit following stimulation of the perforant path. Journal of Physiology, pp.331-356. Boyden, E.S. et al., 2005. Millisecond-timescale, genetically targeted optical control of neural activity. Nature
neuroscience, 8(9), pp.1263-8.
Cardin, J.A. et al., 2009. Driving fast-spiking cells induces gamma rhythm and controls sensory responses. Nature, 459(7247), pp.663-7.
Chen, H.-xin et al., 2001. Is Persistent Activity of Calcium / Calmodulin-Dependent Kinase Required for the Maintenance of LTP ? Is Persistent Activity of Calcium / Calmodulin-Dependent Kinase Required for the Maintenance of LTP ? Journal of Neurophysiology, (01), pp.1368-1376.
Cory, S.M. & Glavinović, M.I., 2006. Molecular Dynamics Simulations of Glutamate Diff usion in Synaptic Cleft. , 18(1-2), pp.61-69.
Crochet, S. & Petersen, C.C.H., 2006. Correlating whisker behavior with membrane potential in barrel cortex of awake mice. Nature neuroscience, 9(5), pp.608-10.
Deisseroth, K. et al., 2006. Next-generation optical technologies for illuminating genetically targeted brain circuits. The Journal of neuroscience : the official journal of the Society for Neuroscience, 26(41), pp.10380-6. Deisseroth, K., 2011. Optogenetics. Nature methods, 8(1), pp.26-29.
Douglas, R.J. & Martin, K. a C., 2007. Mapping the matrix: the ways of neocortex. Neuron, 56(2), pp.226-38. Fiala, A., Suska, A. & Schlüter, O.M., 2010. Optogenetic approaches in neuroscience. Current biology : CB,
20(20), pp.R897-903.
Friston, K.J., 1994. Functional and Effective Connectivity in Neuroimaging : A Synthesis. Human Brain Mapping, 2, pp.56-78.
Goshen, I. et al., 2011. Dynamics of retrieval strategies for remote memories. Cell, 147(3), pp.678-89. Available at: http://www.ncbi.nlm.nih.gov/pubmed/22019004 [Accessed March 10, 2012].
Gradinaru, V. et al., 2010. Molecular and cellular approaches for diversifying and extending optogenetics. Cell, 141(1), pp.154-65.
Gradinaru, V., Mogri, M., Thompson, K.R., Henderson, J.M. & Deisseroth, K., 2009a. Optical deconstruction of parkinsonian neural circuitry - Supplementary Material. Science (New York, N.Y.), 324(5925), pp.354-9. Gradinaru, V., Mogri, M., Thompson, K.R., Henderson, J.M. & Deisseroth, K., 2009b. Optical deconstruction of
parkinsonian neural circuitry. Science, 324(5925), pp.354-9.
Gradinaru, V. et al., 2007. Targeting and readout strategies for fast optical neural control in vitro and in vivo. The Journal of neuroscience : the official journal of the Society for Neuroscience, 27(52), pp.14231-8.
Gunaydin, L. a et al., 2010. Ultrafast optogenetic control. Nature neuroscience, 13(3), pp.387-92.
Hasenstaub, A. et al., 2005. Inhibitory postsynaptic potentials carry synchronized frequency information in active cortical networks. Neuron, 47(3), pp.423-35.
Hegemann, P. & Möglich, A., 2011. Channelrhodopsin engineering and exploration of new optogenetic tools. Nature methods, 8(1), pp.39-42.
Horwitz, B. et al., 2005. Investigating the neural basis for functional and effective connectivity . Application to fMRI Investigating the neural basis for functional and effective connectivity . Application to fMRI. Society, 360, pp.1093-1108.
Janovjak, H. et al., 2010. A light-gated, potassium-selective glutamate receptor for the optical inhibition of neuronal firing. Nature neuroscience, 13(8), pp.1027-32. Available at:
http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2915903&tool=pmcentrez&rendertype=abstr act [Accessed March 8, 2012].
Kalin, N.H., Shelton, S.E. & Davidson, R.J., 2004. The role of the central nucleus of the amygdala in mediating fear and anxiety in the primate. The Journal of neuroscience : the official journal of the Society for
Neuroscience, 24(24), pp.5506-15.
Kuner, T., Augustine, G.J. & Carolina, N., 2000. Neurotechnique Indicator for Chloride : Capturing Chloride Transients in Cultured Hippocampal Neurons. Neuron, 27, pp.447-459.
LeDoux, J., 2003. The emotional brain, fear, and the amygdala. Cellular and molecular neurobiology, 23(4-5), pp.727-38.
Lee, J.H. et al., 2010. Global and local fMRI signals driven by neurons defined optogenetically by type and wiring. Nature, 465(7299), pp.788-92.
Lee, J.H., 2011. Tracing activity across the whole brain neural network with optogenetic functional magnetic resonance imaging. Frontiers in neuroinformatics, 5(October), pp.1-7.
Lewis, T.L. et al., 2009. Myosin-dependent targeting of transmembrane proteins to neuronal dendrites. Nature neuroscience, 12(5), pp.568-76.
Lin, J.Y., 2011. A user’s guide to channelrhodopsin variants: features, limitations and future developments. Experimental physiology, 96(1), pp.19-25.
Lledo, P.M. et al., 1995. Calcium/calmodulin-dependent kinase II and long-term potentiation enhance synaptic transmission by the same mechanism. Proceedings of the National Academy of Sciences of the United States of America, 92(24), pp.11175-9.
Logothetis, N.K., 2010. Bold claims for optogenetics. Nature, 468(7323), p.E3-E4.
Logothetis, N.K., 2008. What we can do and what we cannot do with fMRI. Nature, 453(7197), pp.869-78. Mainen, Z.F. & Sejnowski, T.J., 1996. Influence of dendritic structure on firing pattern in model neocortical
neurons. Nature, 382(July), pp.1-4.
Mancuso, J.J. et al., 2011. Optogenetic probing of functional brain circuitry. Experimental physiology, 96(1), pp.26-33.
Meijer, G.T., Van Kempen, J. & De Knecht, S.D., 2012. ConsolidatiON / ConsolidatiOFF. CSCA Summer School, p.2012.
Meyer, A.H. et al., 2002. In vivo labeling of parvalbumin-positive interneurons and analysis of electrical coupling in identified neurons. The Journal of neuroscience : the official journal of the Society for Neuroscience, 22(16), pp.7055-64.
Nagel, G. et al., 2003. Channelrhodopsin-2, a directly light-gated cation-selective membrane channel. Proceedings of the National Academy of Sciences of the United States of America, 100(24), pp.13940-5. Nagel, G. et al., 2005. Light activation of channelrhodopsin-2 in excitable cells of Caenorhabditis elegans
triggers rapid behavioral responses. Current biology : CB, 15(24), pp.2279-84.
Palmer, H.S., 2010. Optogenetic fMRI sheds light on the neural basis of the BOLD signal. Journal of neurophysiology, 104(4), pp.1838-40.
Petreanu, L. et al., 2009. The subcellular organization of neocortical excitatory connections. Nature, 457(7233), pp.1142-5.
Pitkänen, a et al., 2000. Reciprocal connections between the amygdala and the hippocampal formation, perirhinal cortex, and postrhinal cortex in rat. A review. Annals of the New York Academy of Sciences, 911, pp.369-91. Available at: http://www.ncbi.nlm.nih.gov/pubmed/10911886.
Rein, M.L. & Deussing, J.M., 2012. The optogenetic ( r ) evolution. , pp.95-109.
Roozendaal, B., McEwen, B.S. & Chattarji, S., 2009. Stress, memory and the amygdala. Nature reviews. Neuroscience, 10(6), pp.423-33. Available at: http://www.ncbi.nlm.nih.gov/pubmed/19469026 [Accessed March 3, 2012].
Roozendaal, B. & McGaugh, J.L., 2011. Memory modulation. Behavioral neuroscience, 125(6), pp.797-824. Available at: http://www.ncbi.nlm.nih.gov/pubmed/22122145 [Accessed May 11, 2012].
Scanziani, M. & Häusser, M., 2009. Electrophysiology in the age of light. Nature, 461(7266), pp.930-9.
Schoenenberger, P., Zhang Schärer, Y.-P. & Oertner, T.G., 2010. Channelrhodopsin as a tool to study synaptic transmission and plasticity. Experimental physiology, 96(1), pp.34-39.
Shepherd, G.M.G. et al., 2005. Geometric and functional organization of cortical circuits. Nature neuroscience, 8(6), pp.782-90.
Sohal, V.S., Zhang, F., Yizhar, O. & Deisseroth, K., 2009a. Parvalbumin neurons and gamma rhythms enhance cortical circuit performance - Supplementary Information. Nature, 459(7247), pp.698-702.
Sohal, V.S., Zhang, F., Yizhar, O. & Deisseroth, K., 2009b. Parvalbumin neurons and gamma rhythms enhance cortical circuit performance. Nature, 459(7247), pp.698-702.
Stuber, G.D. et al., 2011. Excitatory transmission from the amygdala to nucleus accumbens facilitates reward seeking. Nature, 0(0), pp.1-6.
Tanaka, J.-I. et al., 2008. Protein synthesis and neurotrophin-dependent structural plasticity of single dendritic spines. Science (New York, N.Y.), 319(5870), pp.1683-7.
Traub, R.D., Jefferys, J.G. & Whittington, M.A., 1997. Simulation of gamma rhythms in networks of interneurons and pyramidal cells. Journal of computational neuroscience, 4(2), pp.141-50.
Traub, R.D. & Wong, R.K.S., 1982. Cellular Mechanism of Neuronal Synchronization in Epilepsy. Advancement Of Science, 216(4547), pp.745-747.
Tye, K.M., Prakash, R., Kim, S.-Y., Fenno, L.E., Grosenick, L., Zarabi, H., Thompson, K.R., Gradinaru, V., Ramakrishnan, C. & Deisseroth, K., 2011a. Amygdala circuitry mediating reversible and bidirectional control of anxiety. Nature, 471(7338), pp.358-362.
Tye, K.M., Prakash, R., Kim, S.-Y., Fenno, L.E., Grosenick, L., Zarabi, H., Thompson, K.R., Gradinaru, V., Ramakrishnan, C. & Deisseroth, K., 2011b. Amygdala circuitry mediating reversible and bidirectional control of anxiety - Supplementary Information. Nature, 471(7338), pp.358-62.
Tye, K.M. et al., 2008. Rapid strengthening of thalamo-amygdala synapses mediates cue-reward learning. Nature, 453(7199), pp.1253-7.
Tønnesen, J. et al., 2009. Optogenetic control of epileptiform activity. Proceedings of the National Academy of Sciences of the United States of America, 106(29), pp.12162-7.
Wallace, D.J. et al., 2008. Single-spike detection in vitro and in vivo with a genetic Ca 2 + sensor. Nature Methods, 5(9), pp.797-804.
Wang, X.J. & Buzsáki, G., 1996. Gamma oscillation by synaptic inhibition in a hippocampal interneuronal network model. The Journal of neuroscience : the official journal of the Society for Neuroscience, 16(20), pp.6402-13.
Weaver, F.M. et al., 2009. Bilateral deep brain stimulation versus best medical therapy for patients with advanced Parkinson’s disease. Current neurology and neuroscience reports, 9(4), pp.266-7.
Welter, M.-L. et al., 2004. Effects of high-frequency stimulation on subthalamic neuronal activity in parkinsonian patients. Archives of neurology, 61(1), pp.89-96.
Zhang, F., Aravanis, A.M., et al., 2007. Circuit-breakers: optical technologies for probing neural signals and systems. Nature neuroscience, 8(August), pp.577-581.
Zhang, F., Wang, L.-P., et al., 2007. Multimodal fast optical interrogation of neural circuitry. Nature, 446(7136), pp.633-9.
Zhang, F., Prigge, M., et al., 2008. Red-shifted optogenetic excitation: a tool for fast neural control derived from Volvox carteri. Nature neuroscience, 11(6), pp.631-3.
Zhang, Y.-P., Holbro, N. & Oertner, T.G., 2008. Optical induction of plasticity at single synapses reveals input-specific accumulation of alphaCaMKII. Proceedings of the National Academy of Sciences of the United States of America, 105(33), pp.12039-44.
Zhang, Y.-P. & Oertner, T.G., 2007. Optical induction of synaptic plasticity using a light-sensitive channel. Nature Methods, Online pub(DECEMBER), pp.1-3.
Zhao, S. et al., 2008. Improved expression of halorhodopsin for light-induced silencing of neuronal activity. Brain cell biology, 36(1-4), pp.141-54.
Box 1. Major Optogenetic Tools
• • •
Within a relatively short time span the number of instruments in the optogenetic toolbox has grown rapidly (Deisseroth 2011) and different properties of opsins have been improved. Since the first successful in vivo applications of Chlamydomonas
channelrhodopsin receptor 2 (ChR2) (Boyden et al. 2005; Nagel et al. 2005) and Natranomas pharaonis halorhodopsin receptor (NpHR) (F. Zhang, Aravanis, et al. 2007) the ChR2-NpHR system has become a popular (actuator) system for bidirectional control of neurons . Both opsins function well within mammalian cells of the nervous system and have (for as far as we know now) at most a limited effect on regular cell physiology in the absence of a light source (Fiala et al. 2010). These opsins can reliably drive excitation and inhibition of neurons and affiliated cells within separate parts of the light spectrum. Over the past few years researchers learned more about the molecular aspects that influence the properties of these opsins and have made improvements. The descriptions below show the major obstacles that have been overcome in matching their optogenetic tools to their task
ChR2 The WT depolarizing ChR2 (Nagel et al. 2003; Boyden et al. 2005) is expressed well in neuronal populations but its kinetic
properties aren’t suited to allow for reliable control over prolonged periods or with stimulation over 25 Hz. Furthermore blue light stimulation has turned out to be detrimental to particular cell type limiting the time span of experiments and applicability in vivo. Targeted mutation of ChR2 has led to the development of a new channelrhodopsin dubbed ChETA (Gunaydin et al. 2010). Contrary to ChR2, ChETA excitation operates with faster kinetic properties – fast depolarization and elimination of the plateau potential – and shows more reliable spiking under stimulation with 5 - 200 Hz light pulses. Following this improvement other mutants such as ChIEF (Lin 2011) that operates with even stricter kinetic properties The ET/TC double mutant (Andre Berndt et al. 2011) is controllable in the far red spectrum at lower light intensities which increases applicability in light sensitive tissues and has better dynamic properties.
NpHR The first applied NpHR (F. Zhang, L.-P. Wang, et al. 2007) initially coped with accumulation of receptors in the ER and low
photocurrents; major improvements have been made up to this day. The second generation enhanced tool: eNpHR2.0, utilizing a genetically added export motif ridded the NpHR of the accumulation in the ER and improved photocurrent properties
significantly (Gradinaru et al. 2010). Additional tweaking of properties and shortening of the construct eventually led to the third generation eNpHR3.1 with even larger photocurrents and improved expression on the membrane (less ER accumulation).
Full optogenetic approach In the majority of published studies optogenetic actuators are combined with quantitative
measurements with the use of electrophysiological techniques. However, a full optogenetic approach is feasible with the use of optogenetic reporters that respond to intra- and intercellular changes in response to cellular processes. Serious candidates that can be utilized in such a setting are ion indicators such as Cl--indicator Clomeleon that can image synaptic inhibition by (Kuner et al.
2000) and Ca2+-indicators such as D3cpVenus (Wallace et al. 2008) as reviewed by Mancuso and colleagues (Mancuso et al. 2011)
