Optogenetics and pharmacogenetics: principles and applications

Abstract

Remote and selective spatiotemporal control of the activity of neurons to regulate behavior and physiological functions has been a long-sought goal in system neuroscience. Identification and subsequent bioengineering of light-sensitive ion channels (e.g., channelrhodopsins, halorhodopsin, and archaerhodopsins) from the bacteria have made it possible to use light to artificially modulate neuronal activity, namely optogenetics. Recent advance in genetics has also allowed development of novel pharmacological tools to selectively and remotely control neuronal activity using engineered G protein-coupled receptors, which can be activated by otherwise inert drug-like small molecules such as the designer receptors exclusively activated by designer drug, a form of chemogenetics. The cutting-edge optogenetics and pharmacogenetics are powerful tools in neuroscience that allow selective and bidirectional modulation of the activity of defined populations of neurons with unprecedented specificity. These novel toolboxes are enabling significant advances in deciphering how the nervous system works and its influence on various physiological processes in health and disease. Here, we discuss the fundamental elements of optogenetics and chemogenetics approaches and some of the applications that yielded significant advances in various areas of neuroscience and beyond.

the complexity of the brain’s structure and function has challenged neuroscientists to understand how neurons operate to orchestrate complex but precise control of behavior and physiological functions. Uncovering how the central nervous system (CNS) maintains consciousness, creates thought and emotion, and integrates stimuli from environment to sustain homeostasis of the entire body should help us understand how the brain controls the body and the pathophysiological processes underlying the diseases arising from CNS dysfunctions.

Extreme complexity and high specificity of neural circuits are orchestrated by billions of neurons working in a highly organized and interconnected manner. Each neuron represents a unit of input and output of information and synergizes with other neurons to receive, process, and convey specific signals to targeted neurons of particular brain regions or to an organ. Communications between neurons rely on refined synaptic connections, which is an essential process for how the brain responds to external diverse array of stimuli. Identification of the fundamental processes governing the function of the CNS requires understanding the neuronal activity patterns, the input and output signals, and the pathways involved in neuronal communication and information processing. Although many decades of neuroanatomical and functional studies have uncovered numerous basic elements, phenomenon, and principles that govern neuronal communication, our current knowledge about the CNS is still limited.

Importantly, recent innovative technological developments open up new avenues to study the brain function in health and disease. In particular, optogenetics and pharmacogenetics approaches have emerged as groundbreaking techniques in neuroscience research, allowing remote manipulation of neuronal activity with high spatiotemporal precision. Combined with the genetically engineered animal models, novel neuronal tract-tracing methodologies and advanced neuroimaging techniques, optogenetics, and pharmacogenetics approaches are revolutionizing the functional dissection of the neural circuits by overcoming the short falls of the conventional neuroscience techniques such as electrical stimulation or lesion studies. In this review, we discuss the basic principles of optogenetics and chemogenetics methods and highlight how these approaches have helped to functionally tease apart complicated neuronal connections and to generate new knowledge about the nervous system and its functions.

Optogenetics

Optogenetics is unique because of its high spatiotemporal resolution, providing the possibility to turn on or off specific populations of neurons to control their functions and behaviors. Although optical stimulation of neurons can be traced back to 1971 when laser radiation was used to scan over the surface of the ganglion of the marine mollusk Aplysia californica to map cellular interconnection (40), the modern neuroscience tool termed optogenetics was introduced by Boyden et al. about a decade ago (20). Optogenetics relies on the genetic modification of known microbial opsin genes encoding light-sensitive proteins (Fig. 1). The seven transmembrane type 1 microbial opsin, which reacts rapidly to light to modulate electrical activity of a neuron, is often used as a control switch (61). Many variants of the opsins are used for optogenetics modulation of neuronal activity, including channelrhodopsin, halorhodopsin, and archaerhodopsin (Fig. 1A). Channelrhodopsin-2 (ChR2) is a light-gated ion channel enabling sodium influx to depolarize neurons within 1–2 ms after light illumination (14). Conversely, halorhodopsin and archaerhodopsin silence neurons through light-induced opening of chloride ion pump and proton pump, respectively, leading to hyperpolarization of neurons within 10–15 and 8–10 ms after light illumination, respectively (51). This millisecond-scale precision evokes synaptic events that mimic the patterns of endogenous single-spike action potential (20). These simulation parameters and the resulting synaptic events are appropriate to evoke release of fast neurotransmitters such as glutamate and γ-aminobutyric acid (GABA) but may be insufficient to cause the release of neuropeptides (13). Thus, the contribution of neuropeptide transmission to the effects evoked by stimulation of ChR2 is unclear.

Fig. 1.

Optogenetics approach. The light-sensitive microbial opsins react rapidly to light to modulate the electrical activity of targeted neurons. A: most popular opsins include channelrhodopsin (ChR), halorhodopsin (HR), and archaerhodopsin (BR). ChR2 is a light-gated sodium channel that responds to blue light (~470 nm) to enable sodium ion influx, leading to neuronal depolarization. Conversely, HR and BR respond to yellow (540 nm) and green (566 nm) lights, respectively, to silence neurons by opening chloride ion channels and proton pumps, respectively, resulting in neuronal hyperpolarization. B: typically, the double-floxed inverse open-reading frame (DIO) contains the inverted repeat (ITR), the elongation factor-1α (EF1α) promoter, an eYFP-ChR2 transgene surrounded by loxP sites and lox2722 sites oriented inward, a woodchuck hepatitis virus posttranscriptional regulatory element (WPRE), and a polyadenylation signal (polyA). Presence of Cre recombinase cleaves the loxP sites, resulting in the expression of the eYFP-ChR2 transgene. C: delivery of a light-sensitive opsin packaged in a viral vector in the desired brain region can be made by stereotaxic microinjection. Once infected, the targeted neurons will be equipped with opsin on the plasma membrane. An implanted fiber optic cannula delivers light causing depolarization or hyperpolarization of the targeted neurons.

Different opsins have been employed to manipulate neuronal activity, but the most frequently used are ChR2 and halorhodopsin (45). The first generation of halorhodopsin (termed NpHR) tended to accumulate in the endoplasmic reticulum, leading to cellular toxicity. To overcome this issue, an endoplasmic reticulum export motif was fused to the COOH-terminus of NpHR (eNpHR2.0), which improved its export, leading to a reduction in aggregate formation (12, 43, 103, 113). Despite these improvements, eNpHR2.0-expressing hippocampal neurons exhibited considerable intracellular labeling and poor membrane localization even though the protein did not appear trapped in the endoplasmic reticulum (45). Addition of COOH-terminal trafficking signals (eNpHR3.0) dramatically reduced intracellular retention and improved surface membrane expression by ∼20-fold relative to the initial NpHR (45). Notably, multiple opsins (e.g., ChR2 and eNpHR3.0) can be expressed in a single neuron while retaining independent responses to particular wave length of light allowing bidirectional manipulation of the activity (excitation or inhibition) of a neuron in a sequential manner (45, 133).

Tools that complement the microbial opsin strategies are emerging. Airan et al. developed a strategy, termed Opto-XR, that takes advantage of the common structure-function relationships among G protein-coupled receptors (GPCRs) to manipulate with light (500 nm) their downstream signaling pathways (4). For instance, fusing the light-sensitive bovine rhodopsin to the intracellular loops of α₁- or β₂-adrenergic receptors allowed optical control of distinct intracellular signaling events and opposite changes in neuronal activity (4). These constructs can be used to manipulate signaling pathways within neurons of behaving animals by modulating the intracellular second messengers such as cAMP in targeted cells with high spatiotemporal precision (4). A similar approach relying on light-activated receptor recruiting the signaling cascade of a specific serotonin receptor has been reported (88).

Delivery Strategies

Genetic targeting strategies, particularly cell type-specific promoters or modified conditionally active viral vectors, allow delivery of the light-sensitive opsins to specific neuronal populations in the brain of living animals (25, 103, 115). In particular, administration of adeno-associated viral (AAV) vectors carrying double-floxed inverted open-reading frame of microbial opsin genes in mice expressing Cre recombinase under the control of a specific promotor enables precise and spatial control of the expression of the actuator protein (Fig. 1, B and C). To facilitate the detection of the expression of the actuator protein, opsins are typically conjugated with fluorescent proteins such as yellow fluorescent protein (YFP), mCherry, or tdTomado. AAV is the most preferred choice of gene delivery to the brain, but, depending on the length of the insertions, lentiviruses can be useful, especially for those opsins that are driven by lengthy promoters beyond AAV packaging capacity (127).

The stereotaxic microinjections that enable the introduction of opsin genes into neurons of targeted brain regions could be technically challenging, making the availability of experimental animals largely depend on the success rate of the microinjections. Even with successful placement of an injection needle in the targeted brain region, the virus can sometimes spread into surrounding areas to cause potential off-target effects. To overcome this limitation, transgenic mice carrying cell type-specific expression of different opsins have been generated. This includes mice expressing ChR2 under the control of cell type-specific promoter elements of thymocyte-1 (10, 119), vesicular glutamate transporter 2 (50), vesicular GABA transporter, choline acetyltransferase, tryptophan hydroxylase 2 (135), and striatal medium spiny neurons (32). Additionally, generation of mouse lines expressing conditional alleles of ChR2 (ChR2-tdTomato and ChR2-eYFP), halorhodopsin (eNpHR3.0), and archaerhodopsin (Arch-ER2) (76) has allowed more refined cell type-specific control of opsins expression. To date, 27 mouse lines for channnerhodopsin, 4 lines for halorhodopsin, 6 lines for archaerhodopsin, and 18 mouse models for Cre-dependent optogenetics are available at the Jackson laboratory (https://www.jax.org/research-and-faculty/tools/optogenetics-resource).

The availability of mouse lines has greatly facilitated the use of optogenetics, but their systemic evaluation is necessary to ensure specificity of the transgene and Cre recombinase to the cell type or neuronal population considered. Moreover, transgenic mice expressing high copy numbers of the transgene in target cells or carrying strong promoters that drive the expression of optogenetics genes are often used to enhance the effects of optogenenics stimulation, but such strategy can lead to variability in expression pattern of the gene in the targeted neurons, which makes comparison of data between animals difficult (131). Thus, there is a need to generate functionally and more tightly controlled transgenic mice for optogenetics studies.

Stimulation Parameters

Activation of individual microbial opsins requires different wavelength of light (Fig. 1). ChR2 respond to ~470 nm blue light to activate neurons (20), whereas halorhodopsin and bacteriorhodopsin hyperpolarize neurons in response to green or yellow lights [e.g., archaerhodopsin at 566 nm (31) and bacteriorhodopsin at 540 nm (45)]. Because lights are unable to penetrate the skull and travel very limited distance within the tissues, fiber optic cannula is often used to bring the light in close proximity to the targeted brain region in freely moving animals (2, 9). Light sources, which are either from light-emitting diode (LED) or laser, as well as light exposure duration, frequency, and pattern of illumination depend on the type of targeted tissues. Laser light in specific wavelength displays millisecond temporal accuracy and has a feature of narrow spectral width typically <1 nm and low divergence to be able to focus on a small spot size of 50–400 μm. Diode laser and diode-pumped solid state laser used with power output of ∼100 mV are able to generate pulses that last as little as 1 ms at 473- and 532-nm wavelengths (127). A fiber optic cannula can be stereotaxically inserted in a targeted brain region to deliver the light. On the top of the fiber optic cannula, a mating sleeve is coupled to fiber optic ferrules that connect to the different wavelength laser generator to transmit the desired laser light depending on the opsin that is expressed.

Because of its nature, laser illumination requires relatively high voltage/current, which can generate certain levels of heat in the targeted tissues. In addition, whereas lasers require tethered optical fibers for behavioral studies, the LED illumination is suitable for integration with the wireless telemetry implantation system, making it appropriate for studies that require untethered fiber optic cables in behaving animals. To overcome the limitation related to the bulky size of such LED implants, microfabrication techniques have been used to miniaturize the LED implants and improve spatial resolution.

Multiple Electrode and Wireless Interfaces

Light from the tip of a single optical fiber travels a few hundred micrometers within the targeted tissue (9), which limits the application of optogenetics to study large-scale neural networks distributed within a given nucleus or in different parts of the brain. A solution to this is to use multiple channels/site stimulation such as miniaturized multielectrode implants to achieve a wide range of illumination combined with multichannel electrophysiological readouts (7) and multichannel fiber photometry (48). In addition, it is possible to optically stimulate specific neuronal populations while simultaneously recording the activity of the same neurons or distant neurons in the intact brain in vivo (26). The next challenge is to couple the optogenetics stimulation with real-time neuronal activity recording for long-term studies in freely moving animals.

The fiber-tethered light source restricts animal movements and can cause entanglement in long-term studies. Development of optical implants with battery-powered or battery-free devices for wireless optical neuronal stimulation could provide the solution to this issue. Rossi et al. have developed a rechargeable lithium battery to drive a low-power microcontroller optical stimulator with an implant system of 14 × 17 × 5 mm in dimension and weighing 2.9 g for application in freely moving small animals like mice. Importantly, the battery for the implanted LED can last more than 50 days, which is sufficient for most experimental needs (98). Hashimoto et al. have developed wireless LEDs that can be controlled by an infrared signal generated by designed transmitters. The light intensity can reach up to 6 mW/mm² and 15 meters communication range between the transmitter and receiver. The battery with 14 × 14 × 10 mm in dimension and 2.4 g of weight can drive the LED stimulator for 67 min when generating sequential photopulses (50-ms pulses at 10 Hz for 2 s duration at 5-s intervals) (52). Lee et al. reported a miniature, wireless, and fiber-coupled modular optical stimulator with a built-in rechargeable solid-state battery system (1.6 g with 12 × 7 × 11 mm dimension) to target a deep brain structure in freely moving mice (69). Because in the battery-powered system the capacity of the battery depends on the weight, a compromise between the size of the battery and stimulation intensity and duration is needed, depending on the experimental conditions.

Wireless power telemetries with battery-free optical neural stimulators enable chronic or longitudinal experiments with high power requirement. Wentz et al. developed a head-borne wireless device that relies on a resonant radiofrequency power link and an adaptive supercapacitor circuit storing the energy with an optional radio transceiver module with 2-watt power weighing 1 g and <1 cm³ in size (122). The wireless device was composed of a microscale inorganic photodetector (μ-IPD) with an ultrathin silicon photodiode (1.25 μm thick, 200 × 200 μm²) and an optional polyimide film-based light weight (~0.7 g) power scavenger. The field temperature under light illumination was found to be 37.46°C in light-pulse (10-ms) frequencies at 20 Hz with peak light and output of 17.7 mW/mm² (63). Park et al. have reported a miniaturized device of 2.4 × 3.5 × 8.5 mm in dimension and only 70 mg in weight (93). A wireless optogenetics device weighing 20–50 mg and producing a 10- to 25-mm³ light covering area, depending on the tissue (brain, spinal cord, and peripheral nerve endings) of behaving mice, has been reported (82). McCall et al. described a 20-μm implantable μ-ILED device allowing chronic (tested for up to 6 mo) wireless optogenetics manipulation of neural circuitry in the home cage environment (80). It should be noted that wireless optofluidic systems able to deliver peptide ligands concurrently with photostimulation have also been developed (59).

Applications of Optogenetics

Correlating change in behavior or physiological functions evoked by photostimulation of specific neuronal populations in brain regions with the activity of those neurons can be achieved in freely moving animals using electrophysiology by fusing the optical fiber with a metal electrode (44), coaxial integrated multielectrode (134), or silicon probe for multisite recording (101). This can also be accomplished by direct nerve recording (1), functional magnetic resonance imaging (68), calcium imaging [e.g., fura 2 probes (133) or green fluorescent protein-calmodulin probe (GcaMP)] (49, 90), or fiber photometry (28, 47, 62). In particular, combining optogenetics with fiber photometry has been useful in determining the effects of optogenetics stimulation on neuronal activity in real time. Fiber photometry consists of implanting optical fibers in the brain region of interest in which neurons express a calcium indicator such as GCaMP6. Upon light delivery, calcium-dependent neural activity is captured through fluorescence microendoscopy in the neuronal soma (34) or projecting terminals (47), enabling the linkage of activity patterns of neural circuits with behavioral paradigms such as anxiety, depression, feeding, and complex learning (48).

The expression of genetically engineered light-sensitive opsins in the axons and terminal projections from an input brain has emerged as an important advantage of optogenetics to delineate the brain neuronal circuits. This approach was particularly very useful to identify and characterize the connectivity of discrete neural pathways and their functional significance in health and disease states (3, 62, 70, 95). Moreover, injection of viral vectors in peripheral tissues has been used to deliver opsins to target subtypes of sensory and motor neurons through retrograde transport from the axon terminals. This strategy enables the expansion of optogenetics to study circuits of the peripheral nervous system and their influence on various functions such as locomotion and pain sensation (56, 114).

Optogenetics has contributed tremendously in uncovering previously unappreciated but important brain functions ranging from fundamental sensing and homeostatic needs, such as smell, touch, sleep, hunger, and drinking, to higher cognitive function and emotional behaviors, such as fear, reward, motivation, learning, and memory (61). These new discoveries were made possible through photostimulation of specific subsets of neurons in the brain of freely moving animals that have been genetically engineered to express Cre recombinase in neurons with unique molecular identity in defined brain regions, which could not be achieved using traditional pharmacological or electronic stimulation approaches. Below we discuss some examples of how optogenetics was used to address some important questions in neuroscience.

Neurological disorders.

To date, the majority of antipsychotics administrated systemically to treat neuropsychiatric diseases can have side effects caused by actions outside the targeted brain, including peripheral tissues and organs. Alternative approaches to specifically modulate targeted neuronal circuits are highly desired. The Food and Drug Administration has approved deep brain stimulation to directly modulate particular brain circuitries in humans to treat Parkinson’s disease (1997), obsessive compulsive disorder (2009), essential tremor (1997), dystonia (2003), chronic pain, major depression, and posttraumatic stress disorder. Existing classical electrical deep brain stimulation therapies nonspecifically stimulate all neurons and axonal projections passing through the targeted field. One currently considered approach is an optogenetics-based deep brain stimulation to stimulate a specific type of neurons only in a targeted brain area. For example, addiction studies have shown that the locomotor sensitization is mediated by the enhanced glutamatergic transmission in the nucleus accumbens (33). Optogenetics-based deep brain stimulation was found to restore normal synaptic transmission and suppress locomotor sensitization after exposure to cocaine compared with classic electrical deep brain stimulation in which locomotor sensitization was not reversed (116). The optogenetics approach reversed the locomotor sensitization via a glutamatergic metabotropic receptor-dependent mechanism that underlies cocaine-evoked synaptic plasticity of glutamatergic transmission in the mesolimbic dopamine system (74). Dysfunctions in the frontostriatal brain circuits have been implicated in repetitive compulsive behaviors. Optogenetics manipulation alleviates repetitive compulsive behavior in a mouse model of obsessive compulsive disorder, bearing deletion of the synaptic scaffolding (Sapap3) gene through enhancement of the feed-forward inhibition in a striatal circuitry (24). Trauma-related contextual fear memory is a key symptom of posttraumatic stress disorder. Optogenetics-mediated silencing of GABAergic neurons in the medial septum, specifically during rapid eye movement sleep, has been shown to have an important role in mediating memory consolidation (19).

The pathological role of either loss- or gain-of-function of cholinergic neurons in Parkinson’s disease has long been debated before the optogenetics technique provided an answer. By selectively expressing channelrhodopsin and halorhodopsin in the striatal cholinergic neurons to activate or silence the cells, it was demonstrated that turning on cholinergic neurons inhibits the majority of GABAergic medium spiny neurons in the striatum, whereas turning off the cholinergic cells activates medium spiny neurons (79). Either administering haloperidol or lesioning the substantia nigra with 6-hydroxydopamine to turn off the cholinergic neurons improved motor function, whereas activating these cells did not affect the motor deficits. Furthermore, hyperactive cholinergic neurons worsen motor symptoms in animal models of Parkinson’s disease (79). Further details of optogenetics application in neurological and psychiatric diseases have been reviewed elsewhere (36, 55, 74, 86, 107, 116).

Feeding and drinking behaviors.

Optogenetics has been heavily applied in instinctive behaviors such as feeding and drinking to decipher the underlying neural circuits. Feeding is a complex reward-driven motivational behavior affected by both homeostatic and hedonic brain pathways. Hypothalamic agouti-related protein (AgRP) neurons are well-known regulators of feeding, since activation of these neurons dramatically increases food intake. Paradoxically, by combining optogenetics with photometry, activity of AgRP neurons was found to be rapidly reduced upon sensing and/or consumption of food (17, 30). Garfield et al. have shown that leptin receptor-expressing GABAergic neurons in the ventral subdivision of the dorsomedial nucleus of the hypothalamus project to AgRP neurons in the arcuate nucleus of the hypothalamus, and optogenetics-mediated stimulation of this inhibitory projection reduces feeding even in overnight-fasted hungry mice (41). Padilla et al. used optogenetics to demonstrate that AgRP neuron projections to the medial amygdala and bed nucleus of the stria terminalis play a critical role in the suppression of territorial aggression and contextual fear (91). These authors also showed that AgRP neurons seem to mediate foraging and repetitive behaviors in the hungry state through neuropeptide Y receptors. Moreover, activation of AgRP neurons can trigger stereotypic behaviors and decrease levels of anxiety upon starvation (37).

Application of optogenetics to glutamatergic neurons of the lateral hypothalamic area revealed that these neurons send projections to lateral habenula to negatively regulate consumption of palatable food (104). On the other hand, photostimulation of lateral hypothalamic area GABAergic projections to the ventral tegmental area (15) and paraventricular nucleus of the hypothalamus (126) increases feeding. Application of optogenetics has also helped uncover some extrahypothalamic circuits that are involved in the control of feeding behavior. For example, Land et al. have shown that optogenetics-mediated activation and inhibition of medial prefrontal dopamine 1 receptor-expressing neurons increase and decrease, respectively, food intake through their innervation of glutamatergic neurons in the medial basolateral amygdala (67).

Although it has long been known that drinking behavior is regulated by the brain circumventricular organs, including the subfornical organ, detailed neural circuits that either initiate or terminate drinking behavior were largely unknown. Combining optogenetics with transgenic animals, Oka et al. have shown the existence of distinct excitatory and inhibitory neurons in the subfornical organ that trigger and terminate, respectively, drinking behavior in mice (89). Although circumventricular organs were conventionally viewed to function as passive sensors to monitor osmolality of blood, optogenetics coupled with calcium fiber photometry allowed Zimmerman et al. to demonstrate that subfornical organ nitric oxide synthase 1-expressing excitatory neurons are involved in the anticipatory regulation of thirst that may affect systemic fluid homeostasis. These findings may open up a new area of research aimed at better understanding of the neurocircuitry underlying thirst and the behavioral mechanisms that detect, integrate, and respond to the change in blood osmolality to maintain fluid homeostasis (139). Rossi et al. have uncovered that photostimulation of GABAergic projections from the substantia nigra par reticulata to the deep layer of the orofacial region of the lateral tectum transiently inhibits the activity of lick-related tectal neurons during self-initiated drinking (99). Deciphering the neurocircuitry controlling anticipatory behaviors for homeostatic needs, including feeding and drinking, is necessary to understand the underlying causes of homeostatic imbalance leading to important medical conditions such as obesity and cardiovascular diseases.

Cardiovascular and autonomic functions.

Optogenetics was used successfully to study cardiovascular regulation in cultured cardiomyocytes (60), zebrafish (12), transgenic mice (22), and nontransgenic animals (87) as well as for computational simulations (21). As a new approach to deal with limitations of electromechanical pacemakers, blue light-mediated illumination of intramyocardial ChR2 was used to pace the rat heart to synchronize ventricular contraction (87). Combined with electroencephalogram and optical mapping of the heart, the diffuse illumination of the ChR2 in the ventricular sites where the transgene was delivered was associated with electrical synchronization of the heart (87). Optogenetics was also coupled with functional magnetic resonance imaging to assess the contribution of vascular components of blood oxygenation level and cerebral blood volume to the functional magnetic resonance imaging signal (129). This strategy provided a hemodynamic mapping signal through the neurovascular network of the individual arteriole and venule (129).

Masamoto et al. have applied optogenetics in a cell-specific manner to explore the role of astrocytes in the regulation of cerebrovascular tone by monitoring the cerebral blood flow using noninvasive laser speckle flowgraphy (78). Mice bearing simultaneous expression of 3 transgenes (ChR2-YFP, neuron-glial antigen 2-Cre, and smooth muscle actin-mCherry) made it possible to demonstrate the precise regulation of vasomotion during neurovascular coupling, particularly the importance of the microvascular diameter or flow change in smooth muscle cells but not pericyte-covered microvessels (54).

Optogenetics had been used to study peripheral nerves using different strategies, including by crossing tyrosine hydroxylase Cre-expressing mice with a reporter mouse model carrying ChR2 (81). Optogenetics stimulation of inguinal fat pad in these transgenic mice significantly increased the release of norepinephrine and stimulated the phosphorylation of hormone-sensitive lipase, which along the anatomical evidence for the sympathetic nerve-adipose junction represent direct evidence about the critical role of the activity of the sympathetic nervous system in the control of white adipose lipolysis (132).

Using photoactivation of noradrenergic locus coeruleus neurons in mice expressing ChR2 under the control of tyrosine hydroxylase-Cre, Wang et al. demonstrated that locus coeruleus noradrenergic neurons influence brain stem parasympathetic cardiac vagal neuronal activity (120). When ChR2 expression is restricted to catecholaminergic sympathetic neurons, photoactivation of the sympathetic fibers triggered norepinephrine release, leading to an increase in contractile force, heart rate, and cardiac electrical activity (121). Optogenetics stimulation of selective vagal sensory neurons either detecting stretch and nutrients in the digestive system or regulating breathing in the respiratory system allowed the demonstration of the importance of these vagal afferents in homeostatic responses to ingested nutrients or mechanical distension of the stomach and intestine and respiratory control, respectively (27, 124).

Optogenetics application in cell signaling.

Although less widely appreciated and used, optogenetic tools have been developed to modify distinct intracellular signaling cascades by incorporating light-sensitive proteins into different systems regulating cell signaling. As discussed above, the Opto-XR strategy enables the manipulation of distinct signaling pathways in the cell. In addition, reversible control of cellular functions was achieved by modifying signaling molecules to make them fuse with photosensitive plant proteins, including the cryptochromes (23, 72), light oxygen-voltage domains (108, 125), phytochromes (85), and the fluorescent protein Dronpa (136). Photocontrol of cellular functions was achieved by switching between photoactivation and photoinactivation in response to specific wavelength of light, conformational changes between hetero- and homodimerization, and manipulating dimerization to control the active and inactive state of a single molecule. Thus, it is possible to use optogenetics to modify cellular processes in a defined cell type of a living organism (42, 111).

Limitations of Optogenetics

Although optogenetics is a powerful tool to explore basic neuronal and circuit dynamics, there are some limitations that are worth mentioning. One such limitation related to the need for surgical implantation of systems to deliver light to the targeted area. Stimulation with red light, which is less scattered by tissue and less absorbed by blood cells, of the red-activatable ChR (ReaChR) enables transcranial optical activation of neurons in deep brain structures without surgical delivery of light through the skull. ReacChR can be stimulated with red-orange (617 nm), red (627 nm), and even far-red (655 nm) light. Thus, the use of ReaChR allows delivery of photons through the skull in the sites of interest without the optical fibers, eliminating fiber optic cannulation-induced tissue damage (71). However, the feasibility of ReaChR to modulate neural activity in deep brain regions such as the hypothalamus remains to be determined.

As mentioned above, delivering the lights into tissues, especially laser, produces heat during the stimulation. To prevent heat-related tissue damage, the light-induced increase in temperature should be <1°C. In array implants, LED in polyethylene terephthalate (80) and probe made by polycrystalline diamond have been reported to be superior in thermal conductivity (39). However, there are many limitations associated with the LED implants, including the fact that their fabrication is time consuming and requires specialized materials, the lack of recording features, and the potential power loss between the LED and the fiber leading to weak illumination in the targeted brain region.

Interestingly, optogenetics stimulation can be used to trigger antidromic spikes when investigating circuits of reciprocal connectivity (102). However, the consequences of antidromic stimulation, a signal traveling opposite to the usual direction of the axon (68), and the “rebound” phenomenon, observed when the light was turned off, remain unclear (110, 127). Light-induced artifacts such as photoelectric or photochemical effect can be reduced by decreasing the area of metal electrode or the wire transducing the light, separating laser-coupled optical fibers from the recording electrodes, changing the angle between the beam and recording metal probe, and by using special material such as thin tungsten wire electrodes (26) or glass electrodes coated with nonreflective material. Instead of gold or platinum, indium-, tin-, and oxide-coated and graphene-based carbon transparent electrodes provide sufficient conductivity with less Becquerel effect (92, 140).

Despite these limitations, the cell type-specific modulation and high spatiotemporal resolution enabled by optogenetics using stable light delivery and implantable low-power wireless device to evoke short-term or long-term neuronal activity coupled with reliable signal readouts is revolutionizing neuroscience. Hopefully, progress in engineering will allow further optimization of the biomaterials and biocompatibility for future use of optogenetics in translational human studies and its application for therapeutic purposes.

Pharmacogenetics

The pharmacogenetics approach consists of the use of receptor proteins that are engineered to exclusively interact with and respond to otherwise inert small molecules to activate or inhibit key downstream cellular pathways (Fig. 2). This strategy is also referred to as receptors activated solely by synthetic ligands because of the use of mutated receptors that are activated by synthetic ligands but not by their endogenous ligands (97). In addition to membrane receptors, various classes of proteins have been pharmacogenetically engineered, including kinases (18, 35, 73), nonkinase enzymes (64, 109), and ligand-gated ion channels (77, 130).

Fig. 2.

Designer receptors exclusively activated by designer drugs (DREADD) approach. The modified G protein-coupled receptors (GPCRs, mutated human M₃ or M₄ muscarinic receptor) can be activated by exogenously administrated otherwise inert drugs like small-molecule clozapine N-oxide (CNO). A: the excitatory DREADD (hM3Dq) is coupled to G_q and, when activated with CNO, results in calcium influx leading to neuronal depolarization. Conversely, the inhibitory DREADD (hM4Di) is coupled to G_i, and its activation with CNO reduces cAMP levels that in turn decrease neuronal activity. B: although the delivery of DREADDs to desired brain regions still requires skilled stereotaxic microinjection, subsequent activation of infected neurons can be simply achieved by administration of CNO systemically, ip or in the drinking water. C: intravenous administration of CNO caused a rapid decrease in the sympathetic nerve activity (SNA) subserving brown adipose tissue (BAT) in mice expressing DREAAD (hM3Dq) in AgRP neurons but not in control mice (106).

The designer receptors exclusively activated by designer drugs (DREADDs) developed by Armbruster et al. at the University of North Carolina are the most commonly used form of chemogenetics. The usefulness of DREADD in modulating neuronal activity was demonstrated first in vitro (11) and later in vivo (6) before it was widely adopted by the research community. DREADD involve the introduction of genetic modifications to endogenous GPCRs to produce synthetic receptors that lack endogenous ligands. However, these modified GPCRs are sensitive to an otherwise inert synthetic ligand, clozapine N-oxide (CNO) (16). For instance, although CNO, but not the endogenous ligand acetylcholine, can activate the modified human muscarinic acetylcholine receptor (Fig. 2A), CNO is not an active ligand for the endogenous muscarinic acetylcholine receptors.

Analysis of CNO pharmacokinetics showed that, in mice, plasma levels of CNO peak at 15 min following intraperitoneal injection (1 mg/kg dose) and become very low after 2 h (46, 123). Although the plasma half-life of acutely administered CNO may be short, the biological effects evoked by this ligand usually last much longer (up to 24 h) (6). Oral CNO administration via drinking water was shown to cause robust phenotypes in mice expressing DREADD in pancreatic β-cells, indicating that significant amounts of CNO are absorbed from the gastrointestinal tract (57).

Although the mechanisms involved are not entirely clear, CNO cross the blood-brain barrier to reach the CNS (53, 96). The fact that CNO can be converted to psychoactive compounds, particularly clozapine, in several species, including humans (59), needs to be taken into consideration. Careful evaluation of the biological effects of CNO metabolites should always be considered. This can be achieved by testing the effects on the parameters considered of CNO in wild-type animals.

Depending on the α-subunit of the G protein to which it is coupled, DREADD activation triggers distinct downstream second messengers with differential consequences on cellular processes and neuronal activity (Fig. 2A). The two most popular variants of DREADDs are derived from the muscarinic acetylcholine receptors. The G_q-coupled hM3G_q receptor is derived from the human M₃ receptor, and its activation by CNO increases intracellular calcium, leading to burst-like firing of neurons. On the other hand, CNO-mediated stimulation of the G_i-coupled hM4Gi receptor, derived from the human M₄ receptor, cause membrane hyperpolarization through a decrease in cAMP. Arrestin-biased DREADD coupled with G_s/G_q is a novel strategy that allows the examination of non-G protein-dependent signaling pathways (66, 75). Nakajima and Wess introduced a mutation (R165L) to the M3Dq DREADD that rendered it unable to recruit G proteins but still retained the ability to recruit arrestin and induce ERK1/2 phosphorylation in a CNO-dependent manner (84).

Bidirectional pharmacogenetics control of neuronal activity can be achieved by combining hM3Gq DREADD with an alternative human κ-opioid receptor inhibitory DREADD that is selectively activated by salvinorin B (118). To avoid potential backmetabolism of CNO to clozapine and other clozapine metabolites such as N-desmethylclozapine (58) and to prevent clozapine-related side effects like hypotension, sedation, and anticholinergic syndrome, Chen et al. developed a new non-CNO chemical actuator, compound 21, that has minimal off-target activity and excellent selectivity for hM3Dq vs. muscarinic and other GPCRs (29).

Pharmacogenetics studies were greatly facilitated by the production of transgenic mouse lines such as those expressing the hM3Dq DREADD specifically in principal neurons of the forebrain driven by the Ca²⁺/calmodulin-dependent protein kinase II α (CaMKIIα) promoter (6) and those expressing the G_i-coupled hM4Di DREADD also under the control of CaMKIIα promoter (138). Moreover, mice bearing Cre recombinase-dependent expression of hM3Dq and hM4Di have been generated (137). There are currently 18 pharmacogenetic transgenic lines available from the Jackson laboratory (https://www.jax.org/research-and-faculty/tools/optogenetics-resource).

Like optogenetics, pharmacogenetics has been broadly used in the field of neuroscience to study the central and peripheral nervous systems and other tissues. CaMKII, AgRP, proopiomelanocortin, GABAergic, dopaminergic, serotonergic, and orexin neurons as well as pancreatic β-cells have been specifically targeted by DREADD approaches. For example, pharmacogenetic activation of AgRP neurons has been shown to increase food intake (65) and impair insulin sensitivity (106). In addition, we demonstrated that DREADD-mediated activation of AgRP neurons reduces sympathetic nerve traffic to thermogenic brown adipose tissue (Fig. 2C and Ref. 106). More details regarding DREADD applications have been reviewed elsewhere (38, 100, 117). Of note, optogenetics can be combined with pharmacogenetics in the same study depending on the purpose of the experiment (106).

Comparison of Optogenetics and Pharmacogenetics

Despite different concepts and stimulation strategies, optogenetics and pharmacogenetics can achieve similar effects on neuronal activity and behavioral outcomes. For example, stimulation of AgRP neurons by either ChR2 (8) or hM3Dq (65) results in robust hyperphagia. Depending on the experimental design, often time both techniques require stereotaxic injection of viral vectors to express either light-sensitive opsin or designed receptors in a targeted brain region or specific neuronal populations. The activity of the gene promoter driving Cre expression may vary, which influences the level of expression of opsin or DREADD. It is also worth noting that viral infections may not occur in all neurons in a targeted area. Because opsins and DREADDs will be preferentially overexpressed in the cell membrane, it may disrupt the expression or function of other endogenous receptors affecting normal neuronal function and perhaps behaviors. Although comparable expression of opsins and DREADDs can be achieved in a particular brain region, the number of neurons activated may be greater in DREADD than optogenetics, since light cannot reach all of the infected neurons because of limited ability to travel within the targeted area, whereas systemic injection of a chemical actuator such as CNO should be able to reach every DREADD-infected neuron through the circulation. On the other hand, systemic administration of chemical actuator may have potential side effects caused by the metabolites.

Optogenetics yield high spatiotemporal resolution to regulate the activity of specific neuronal populations, which represent an advantage compared with the traditional approaches or DREADDs. In contrast to optogenetics, pharmacogenetics allows a long-term control of neuronal activity and can target a broader area. Indeed, not all neuronal activities require millisecond time scale reactions. In a relatively chronic manipulation, DREADD may be more appropriate than optogenetics. Another feature of DREADD is the possibility to perform persistent stimulation without the need for constant light delivery, which allows the animal to perform complex behavioral tasks without constrain by an implant attached to its head.

DREADD activity lasts from minutes to hours after CNO administration. Thus, DREADDs can modulate neuronal activity less invasively and more persistently compared with optogenetics. Moreover, CNO can be delivered via intracranial microinjection (83). Another unique feature of DREADDs relates to the fact that receptors can selectively modulate intracellular signaling cascades, and pharmacogenetics has been coupled to ion channels to offer fast kinetic modulation (77). However, with the Opto-XR strategy, optogenetics has been adapted to GPCRs as well, making it possible to manipulate distinct intracellular events with this approach.

Perspectives and Significance

Clearly, both optogenetics and chemogenetics are excellent tools for circuit neuronal mapping and to interrogate the effects of modulating the activity of specific neurons on behavior or physiological functions. However, one has to be cautious in drawing conclusions about the neural control of physiological processes based exclusively on these techniques. This is due to the fact that modulation of neuronal activity with these tools remains artificial and does not necessary mean that similar change in neuronal firing or cellular events occurs in a physiological setting. Thus, there is a need to complement studies using optogenetics and chemogenetics with phytologically relevant approaches to ensure the appropriateness of the conclusions drawn from studies using these techniques.

Optogenetics applications have recently been expanded to nonhuman primates, such as rhesus macaques. By coinfecting midbrain dopaminergic neurons of nontransgenic monkeys with AAVs encoding ChR2 and tyrosine hydroxylase-driven Cre, a selective optogenetic stimulation of dopamine neurons promoted reward-related learning processes (105). Optogenetics has also been adapted to peripheral tissues to manipulate nonexcitable cells (94) and even mitochondrial membrane and functions (112). These developments raise the hope for therapeutic optogenetics in the clinics. Ongoing research at Circuit Therapeutics (http://www.circuittx.com/) is focused at applying optogenetics to treat human neurological disorders, including Parkinson’s disease. RetroSense Therapeutics (http://retrosense.com/) has an ongoing clinical trial to treat retinitis pigmentosa using optogenetics. Another company (GenSight Biologics, http://www.gensight-biologics.com/) is attempting to treat retinitis pigmentosa with an opsin protein that responds to red light, which is less harsh on the eyes than blue light. DREADD-based clinical interventions are also under consideration for Parkinson’s disease and seizures (5).

The main challenge before those therapies become available is the development of safe strategies to deliver opsins or DREADD in adult human neurons in targeted brain regions. Among the different viruses currently used for gene therapy in humans, AAV is reported to be safe for human application, but long-term safety survey of optogenetics/pharmacogenetics viral vectors in humans will still need to be conducted carefully. Optogenetics or pharmacogenetics can be combined with gene therapy to promote the expression of genes, miniaturized array, and dedicated light delivery systems, all of which need to be developed for large-scale and complex neuronal circuit applications. The development of other proven safe chemical actuators for pharmacogenetics is also warranted. All of these technical developments hold promise for the use of optogenetics and pharmacogenetics in the clinic for the treatment of a variety of human diseases affecting the CNS and beyond.

NOTE ADDED IN PROOF

A paper (45a) published after the current review was accepted claims that CNO activate DREADD receptors through its derivative, clozapine. Clozapine displayed high affinity for DREADD receptors and potently induce Ca²⁺, whereas CNO was unable to bind or activate these receptors. In addition, CNO exhibited limited ability to cross the blood-brain barrier and reach the brain. In contrast, a very low dose of clozapine administered systemically triggers DREADD-mediated behavioral responses.

GRANTS

The research is supported by National Institutes of Health Grants (NIH) HL-084207 to K. Rahmouni and HL-127673 and MH-109920 to H. Cui, the American Heart Association (14EIA18860041 to K. Rahmouni), the University of Iowa Fraternal Order of Eagles Diabetes Research Center (FOEDRC), and the University of Iowa Center for Hypertension Research (both to K. Rahmouni). J. Jiang is a research trainee of the FOEDRC and is supported by NIH Grant T32-DK-112751–01.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.J., H.C., and K.R. analyzed data; J.J. prepared figures; J.J., H.C., and K.R. drafted manuscript; J.J., H.C., and K.R. edited and revised manuscript; J.J., H.C., and K.R. approved final version of manuscript.