Imaging-based chemogenetics for dissecting neural circuits in nonhuman primates

Abstract

Nonhuman primates, particularly macaque and marmoset monkeys, serve as invaluable models for studying complex brain functions and behavior. However, the lack of suitable genetic neuromodulation tools has constrained research at the network level. This review examines the application of a chemogenetic technology, specifically, designer receptors exclusively activated by designer drugs (DREADDs), to nonhuman primates. DREADDs offer a means of reversibly controlling neuronal activity within a specific cell type or neural pathway, effectively targeting multiple brain regions simultaneously. The combination of DREADDs with imaging techniques, such as positron emission tomography and magnetic resonance imaging, has significantly enhanced nonhuman primate research, facilitating the precise visualization and manipulation of specific brain circuits and enabling the detailed monitoring of changes in network activity, which can then be correlated with altered behavior. This review outlines these technological advances and considers their potential for enhancing our understanding of primate brain circuit function and developing novel therapeutic approaches for treating brain diseases.

1. Introduction

Recent advances in neuroscience research have highlighted the importance of nonhuman primates, such as macaques and marmosets, in unraveling the complexities of the human brain.¹⁾ As biological models that more closely resemble humans, these primates offer unparalleled insights into higher cognitive functions and the intricacies of psychiatric and neurological disorders.²⁾ Notably, functional connectivity patterns between brain regions, as revealed through human brain imaging studies, have provided important clues regarding the specific functions and symptoms associated with these disorders.^3),4) However, while research on neural circuits and their functional roles has progressed notably in rodents, primarily through innovative genetic techniques enabling the precise measurement and control of neural activity, there is a significant species gap that prevents direct translation of these findings to humans. Rodents possess a prefrontal cortex and hierarchically organized visual areas, but these regions differ significantly in structure and complexity from those in primates.^5)-7) These differences become particularly evident when studying higher brain functions such as social behavior, decision-making, and understanding visual cognition. A significant gap also persists with respect to pathological conditions and the development of effective treatments.⁸⁾

To bridge these gaps, the application of genetic techniques first developed in rodent models to nonhuman primates has emerged as a critical and attractive research strategy. Such an approach promises to help unravel the complexity of functional brain circuits in primates, which are more relevant to human physiology than is rodent brain circuitry. However, this endeavor is not without its challenges. Achieving optimized gene delivery techniques in the primate brain is an ongoing challenge that needs to be addressed. In addition, techniques such as optogenetics, while revolutionary, face technical limitations in their application to nonhuman primates, in part due to the large size of the primate brain.⁹⁾

Chemogenetics, such as designer receptors exclusively activated by designer drugs (DREADDs), allows the reversible control of neuronal activity through the systemic administration of an agonist that activates designer receptors. This approach is not constrained by spatial limitations associated with optogenetics, making it particularly suitable for use in animals with a large brain such as primates. Moreover, the combination of chemogenetics with in vivo imaging techniques, such as positron emission tomography (PET), now allows the visualization of designer receptor expression, thereby increasing the optimization of gene delivery techniques in nonhuman primates.

This review provides a brief introduction to chemogenetic techniques with a focus on DREADDs, discussing their integration with imaging methods for application in nonhuman primate brains. We then review recent advances reported in state-of-the-art chemogenetic studies of higher-order brain functions, followed by a discussion of how these techniques can be best combined with other methods for measuring neural activity. We also discuss the advantages of these technologies for understanding human brain pathologies and developing therapeutic interventions. Finally, we provide an outline of the future prospects.

2. Imaging-based optimization of chemogenetics in the primate brain

2.1. DREADDs and challenges in their application in nonhuman primates.

DREADDs are a family of engineered receptors created by introducing artificial mutations into human G protein-coupled receptors.¹⁰⁾ Among these, muscarinic-based DREADDs are the most widely used, having been designed specifically to be activated by clozapine-N-oxide (CNO), a metabolite of clozapine (CLZ) known for its limited native activity in the brain. This approach has yielded two primary DREADDs: hM4Di, which is coupled to G_i signaling and includes hyperpolarization via activation of inwardly rectifying potassium channels (GIRKs)¹¹⁾ and synaptic silencing via inhibition of presynaptic neurotransmitter release¹²⁾; and hM3Dq, which is coupled to G_q signaling and leads to depolarization via increasing intracellular calcium levels¹³⁾ (Fig. 1). By genetically introducing these designer receptors into specific neuronal populations, targeted neural activity can be selectively inhibited (hM4Di) or activated (hM3Dq) for several hours through systemic CNO administration without a special device, thus enabling remote control of specific neuronal/circuit functions.¹³⁾ Unlike conventional systemic drug administration, which can affect multiple brain sites and neural systems and result in low specificity, DREADDs offer high specificity by genetically targeting the neurons to be affected, thus providing a more accurate method for remotely controlling specific neural circuits.

Fig. 1.

Muscarinic-based DREADDs. A. The inhibitory designer receptor hM4Di (left) and the excitatory designer receptor hM3Dq (right). Both receptors lack responsiveness to any endogenous ligands but are activated by designer agonists such as clozapine-N-oxide (CNO) or deschloroclozapine (DCZ), thereby suppressing or exciting neural activity. B. Chemogenetic neural control via hM4Di activation. (Top) In mouse hippocampal cultured neurons expressing hM4Di, action potentials are inhibited in the presence of CNO. (Bottom) In cultured neurons that do not express hM4Di, CNO does not affect neuronal activity (adapted from Nagai et al., Nat. Commun. 2016).¹⁴⁾

Although DREADDs have become popular tools for manipulating neural activity in rodent models alongside optogenetics, adapting this technology to nonhuman primates presents significant challenges. Effective expression and activation of DREADDs in the larger and more complex primate brain requires careful optimization of gene transfer and agonist delivery, both of which mutually impact chemogenetic controllability. Given the limited availability of monkeys and the need for extensive trial and error, the optimization process that is usually employed in rodents is too time-consuming and too costly, monetarily as well as in terms of animal life. Consequently, novel solutions are required to address the challenges associated with the application of DREADDs in primates.

2.2. Imaging-guided optimization.

To address these challenges, we developed a methodology that integrates chemogenetics with PET imaging: PET imaging-guided chemogenetics (Fig. 2). This technique enables the visualization of DREADD expression in living animals, allowing for longitudinal monitoring of receptor expression levels and distribution. For instance, PET imaging with radiolabeled CLZ ([¹¹C]CLZ) demonstrated that expression levels of hM4Di reached their peak at 60–80 days after vector injection and remained stable for at least 1.5 years.¹⁴⁾ Moreover, by administering the agonist before PET scanning, researchers can measure the occupancy of designer receptors with the resulting PET signal attenuation, thereby establishing the agonist dose-occupancy relationship in vivo.

Fig. 2.

Conceptual illustration of PET imaging-guided chemogenetics. After injection of a viral vector into the brain (left), PET imaging is performed using a DREADD-selective radioactive ligand (center). This process enables the visualization and measurement of the location and level of DREADDs and their stability in the living animal based on the accumulation of the ligand. This is a highly useful screening method before starting a series of behavioral experiments (right). For example, if it is found that the level of DREADD expression is insufficient or that the expression does not cover the targeted brain region, an additional injection of the vector can be performed to achieve ideal expression.

Proof-of-concept studies for PET-guided DREADD manipulation in nonhuman primates provide compelling evidence that this method is highly efficient. One such study used PET to confirm hM4Di expression in the bilateral rostromedial caudate nucleus (rmCD) via injections of adeno-associated virus (AAV) vectors.¹⁴⁾ The study demonstrated that the administration of a sufficient amount of CNO impaired reward sensitivity, a critical function of the rmCD. Additionally, the absence of behavioral effects was found to be correlated with weak/mistargeted hM4Di expression. Other research showed that introducing hM4Di into the orbitofrontal cortex (OFC), a broad and deep region of the primate prefrontal cortex,¹⁵⁾ accompanied by a lesion of the contralateral perirhinal cortex, resulted in deficits in reward-related functions. Similar reversible manipulations can be accomplished through pharmacological techniques or cooling; however, both approaches presented difficulties in achieving reliable reproducibility and in limiting the target regions. Thus, these studies demonstrate the effectiveness of DREADDs in behavioral neuroscience in nonhuman primates and validate the integration of chemogenetics with PET imaging as a useful methodology, thus contributing to the broadening of subsequent nonhuman primate chemogenetic research.

2.3. Development of a new DREADD agonist.

The original DREADD agonist, CNO (Fig. 3A), was believed to have low affinity for endogenous receptors and to exert negligible effects on the central nervous system. However, CNO has been shown to be metabolized via back-transformation to CLZ (i.e., retroconversion) in guinea pigs and humans.¹⁶⁾ A rodent study reported that retroconverted CLZ, rather than CNO, enters the brain and activates DREADDs.¹⁷⁾ This retroconversion is particularly problematic as it raises concerns about the potential activation of endogenous receptors by CLZ, leading to off-target effects. In nonhuman primates, although retroconversion is limited,^14),15),18) a PET study demonstrated that a very low proportion of injected CNO entered the brain (Fig. 3B). Consequently, a significant amount of CNO is required for systemic injection to activate DREADDs, resulting in a high cost and a delay in the onset of action of approximately 1 h.¹⁴⁾ This presents logistical challenges in experimental designs, particularly when combined with electrophysiology. These issues highlight the need to develop an optimal agonist as an alternative to CNO.

Fig. 3.

Comparison of DCZ and CNO by PET: efficient properties for good brain penetrance and selective binding in DREADD-expressing regions. A. Chemical structure of CNO. B. [¹¹C]CNO PET image. Standardized tracer uptake values 30–90 min after administration. Low signals were observed in the brain, indicating that CNO has low brain permeability. C. Chemical structure of DCZ. D. [¹¹C]DCZ PET image. Selective and high accumulation is observed only where DREADDs are expressed (left striatum; hM4Di), whereas accumulation on the right side is minimal. The concentrations of [¹¹C]CNO and [¹¹C]DCZ in the entire brain at 30 min were 0.14% and 5.9% of the injected dose, respectively. SUV: standardized uptake value (adapted from Nagai et al., Nat. Neurosci. 2020).²⁰⁾

The optimal characteristics of a DREADD agonist include binding and activating DREADDs at low concentrations (high affinity and potency), no interaction with other endogenous receptors (high DREADD selectivity), rapid brain penetration (blood–brain barrier permeability), and metabolic stability in the body. Leveraging in vivo PET imaging to screen for these properties, we started with CLZ, previously used as a DREADD PET ligand, searching for compounds that exhibited no affinity for endogenous receptors. One CLZ analog, deschloroclozapine (DCZ) (Fig. 3C), was identified as a candidate because of its low affinity for dopamine and serotonin receptors.¹⁹⁾

First, a radiolabeled form of DCZ, [¹¹C]DCZ, was administered to monkeys that received AAV injections expressing hM4Di in the striatum, and PET scans were subsequently obtained.²⁰⁾ [¹¹C]DCZ quickly penetrated the brain and selectively accumulated in areas expressing hM4Di (Fig. 3D), showing promise as an ideal DREADD ligand that may meet the specified criteria. In vitro assays further confirmed that DCZ strongly activated DREADDs and showed high selectivity without affecting endogenous receptors. In addition, in vivo experiments demonstrated that the amount of DCZ that could quickly activate neurons expressing hM3Dq was only one-hundredth of the amount of CNO that was required. A PET occupancy study showed that DCZ occupies DREADDs at lower doses than CNO or C21, the second-generation DREADD agonist. Moreover, in monkeys expressing hM4Di in the dorsolateral prefrontal cortex (DLPFC), a very low dose (0.1 mg/kg) of DCZ administered intramuscularly impaired spatial working memory, a function associated with this brain region.²⁰⁾

DCZ is now commercially available from multiple companies and has become a standard agonist for use in mice, rats, and monkeys.^21)-25) A series of safety studies has demonstrated that 0.1 mg/kg DCZ does not produce detectable off-target effects on neuronal activity or behavior in both monkeys and rodents.^{23),24),26)-28)} Nevertheless, confirmation of the safe dosage in specific experimental conditions is recommended. Although other agonists such as JHU37160²⁹⁾ and olanzapine³⁰⁾ have been introduced, their optimal dosages and off-target effects need to be clarified in future studies. Recent structural and in vitro analyses have indicated that DCZ binds deeply into the pocket of DREADD molecules and exhibits the highest potency among existing DREADD ligands.³¹⁾

In addition to its use as an agonist, the radiolabeled form of DCZ, [¹¹C]DCZ, also functions as a highly effective PET ligand for DREADDs, offering greater DREADD selectivity than [¹¹C]CLZ, with quantification of the expression level.³²⁾ It is also highly useful for facilitating the evaluation of AAV vectors³³⁾ and DREADD expression-reporter imaging.³⁴⁾ The development and adoption of DCZ as both a new agonist and a PET ligand in DREADD-based research mark a significant advance in chemogenetics, addressing the problems posed by CNO and enhancing the precision, efficacy, and applicability of DREADD technology in neuroscience research in nonhuman primates.

3. Use of DREADDs in nonhuman primates

3.1. Manipulation in single or multiple areas.

The implementation of DREADDs in nonhuman primates has facilitated significant advances in neuroscience, particularly when studying the pathogenesis of mental and neurological disorders. For instance, administering agonists after introducing the inhibitory DREADD hM4Di into macaque monkey amygdala neurons resulted in a transient reduction in anxiety-related behaviors.³⁵⁾ Similar manipulations in infant macaques have been reported to influence socioemotional behaviors.³⁶⁾ Conversely, activation of the amygdala with the excitatory DREADD hM3Dq has been shown to induce anxiety-like behaviors.³⁷⁾ Additional research has explored the role of the putamen and nucleus accumbens in alcohol consumption,^38),39) as well as the contribution of the OFC–rmCD connection and the caudate nucleus to motivation control and impulsivity.^40)-42) Moreover, another study used macaque monkeys expressing hM4Di in the dorsomedial prefrontal cortex and discovered that receiving the agonist transiently impaired the ability to “read the minds of others,” providing valuable insights into “theory of mind” ability and autism spectrum disorders, which are associated with deficits in this skill.⁴³⁾ These studies demonstrated the value of DREADDs as tools for inducing transient and reproducible changes in activity in specific neuronal populations associated with the manifestation of symptoms.

3.2. Specific cell type and circuit manipulation.

Beyond application to localized neuronal populations, DREADDs also enable targeted manipulation of specific neuronal cell types or pathways. For example, by injecting an AAV vector with a dopamine-specific promoter (TH) into the substantia nigra, researchers have produced marmosets with the excitatory DREADD hM3Dq expressed specifically in dopamine neurons.⁴⁴⁾ The marmosets rotated in a contralateral direction relative to the activated side after consuming food containing DCZ. Similarly, the use of TH promoters has enabled inhibitory DREADD expression selective to noradrenergic neurons in the macaque locus coeruleus.⁴⁵⁾ DCZ administration resulted in a reduction in the firing rate of locus coeruleus neurons and a decrease in vigilance in these monkeys.

Pathway-specific manipulation has been achieved by the joint use of two specific types of viral vectors that both use recombinant enzymes: a retrograde vector expressing Cre recombinase and an anterograde vector expressing functional proteins such as DREADDs in the presence of Cre. By injecting these vectors into the projection and origin areas of a specific neural pathway, neurons infected by both vectors (i.e., those connecting the origin and projection) selectively express functional proteins.⁴⁶⁾ This method has been successfully employed to selectively express hM4Di only in prefrontal neurons that project to the caudate nucleus, temporarily inhibiting neuronal activity and modifying behavior associated with flexible action selection.⁴⁷⁾ An additional study has shown that silencing the projection from the medial prefrontal cortex to the lateral hypothalamus disrupts the valuation of social rewards.⁴⁸⁾

Another approach for pathway-specific manipulation can be achieved by using local agonist infusion to activate hM4Di that is expressed at axon terminals, leading to the suppression of synaptic transmission.¹²⁾ In one study, hM4Di expression was induced in neurons of the bilateral DLPFC, and PET imaging with [¹¹C]DCZ was used to identify the targets of these neurons’ projections: the dorsal part of the caudate nucleus (dCD) and the lateral part of the thalamic mediodorsal nucleus (MDl).⁴⁹⁾ Inhibition of the DLPFC→MDl pathway via local DCZ injection into the MDl impaired performance on a spatial working memory task, whereas inhibition of the DLPFC→dCD pathway did not (Fig. 4). Conversely, inhibition of the DLPFC→dCD pathway by DCZ injection into one side of the dCD resulted in a significant increase in choice bias toward the same side during a free-choice task, whereas the same treatment for the DLPFC→MDl pathway had no impact. This demonstrated the dissociable roles of primate prefronto-subcortical pathways in cognitive functions. Although conventional pharmacological experiments using muscimol and other drugs may demonstrate functional double dissociation, the DREADD approach offers clear advantages by providing direct evidence of the causal roles of different prefronto-subcortical pathways. The same approach was applied to pathways from the OFC to rmCD and the medial part of MD, demonstrating their distinct roles in value-based adaptive decision-making.⁵⁰⁾ Furthermore, this approach with hM3Dq has also been applied to the activation of separate pathways between the anterior cingulate cortex and subcortical regions in marmosets, revealing dissociable circuits for anhedonia and anxiety.⁵¹⁾

Fig. 4.

Utilizing DCZ-PET for mapping and silencing the DLPFC-subcortical circuit in monkeys. A. Spatial delayed response task. Monkeys must remember the location of food on either the left or the right side for up to 30 s. B. Local inhibition of the DLPFC. Systemic administration of DCZ caused a delay-time-dependent impairment in working memory. C. Pathway-selective silencing of DLPFC neuronal projections by local infusion of DCZ into hM4Di expression sites, as visualized by PET. Inhibition of the DLPFC→MDl thalamic nucleus pathway resulted in working memory impairment (right). In contrast, inhibition of the DLPFC→dCD pathway did not affect performance (left). Asterisks indicate P ＜ 0.05 for significant main effect of treatment (adopted from Oyama et al., Sci. Adv. 2021).⁴⁹⁾

Overall, employing DREADDs to manipulate specific neural pathways in the primate brain allows researchers to better understand the circuitry underlying higher cognitive functions and offers new possibilities for studying and addressing the pathophysiology of human psychiatric and neurological disorders.

3.3. Combining DREADDs with functional imaging techniques.

DREADDs allow manipulation of neural circuits without the need for implanted devices, thereby avoiding issues such as image distortion caused by the devices in combination with whole-brain functional imaging techniques such as MRI. An example of this application involves using hM4Di in the amygdala of macaque monkeys. Researchers examined changes in functional connectivity associated with amygdala inhibition induced by CNO administration using resting-state functional magnetic resonance imaging (fMRI). Their findings demonstrated that the observed changes reflect anatomical connections.⁵²⁾ A more recent resting-state fMRI study demonstrated that chemogenetic silencing of the amygdala via hM4Di/DCZ enhances functional connectivity between the amygdala and the ventrolateral prefrontal cortex. Subsequent electrophysiological examination indicated that the underlying neuronal signature of this phenomenon was an increase in local field potential coherency between these circuits.⁵³⁾

The combination of DREADDs with functional imaging allows the identification of brain regions involved in specific functions and the investigation of their causal contribution by examining induced changes in network dynamics and behavior. In an example of this approach,⁵⁴⁾ monkeys received injections of an AAV vector expressing hM4Di into the primary somatosensory cortex (SI), as identified by fMRI to respond to somatosensory stimulation of a specific finger of the contralateral hand. Subsequent DCZ administration led to impairment when using this hand. Analysis showed that silencing the SI finger region attenuated the somatosensory response not only in the SI, but also in several brain regions of the downstream grasping network, which receive connections from the SI. In addition, silencing the SI hand region disinhibited SI foot representation, accompanied by behavioral hypersensation.

A more recent chemogenetic functional imaging study sought to elucidate the causal functions of the fronto-temporal network in object memory.⁵⁵⁾ Using [¹⁵O]H₂O-PET, part of the OFC was identified as the site of neural activation associated with memory load during a delayed matching task. Focal chemogenetic silencing of the identified OFC region resulted in downregulation of both local OFC activation and remote activation in the anterior temporal cortex, accompanied by impairment of mnemonic performance in the task. Further electrophysiological investigation revealed that top-down modulation from the OFC enhanced stimulus-selective mnemonic signals in individual anterior temporal neurons.

As demonstrated by these studies, the combination of DREADDs with functional imaging allows the discovery of critical network operations that underpin specific brain functions, regardless of prior knowledge or hypotheses. This approach has the potential to be widely applicable to the elucidation of multiscale network operations that subserve cognitive and emotional processes in primate models, thus aiding the translational understanding of corresponding functions and their disruption in the human brain.

3.4. Development of therapeutics.

The use of DREADDs to treat brain disorders such as epilepsy,^56)-59) neuropathic pain,^60)-62) and Parkinson’s disease (PD)^63)-65) is an attractive concept that is gaining traction. While efficacy validation continues through basic research using rodent models,⁶⁶⁾ pioneering studies in nonhuman primates are showing promising results. For example, a proof-of-concept study demonstrated on-demand treatment of frontal lobe epilepsy in monkeys using DREADDs.⁶⁷⁾ After expressing inhibitory DREADDs in the primary motor cortex (MI) of macaque monkeys, injection of the GABAa receptor inhibitor bicuculline into the central part of the expression region induced disinhibition of the neuronal population in MI, resulting in transient epileptic-like brain waves and behavior, characterized by repeated involuntary hand/body movements. Subsequent intramuscular injection of DCZ significantly attenuated the amplitude of epileptic-like brain waves within approximately 1 min, accompanied by markedly reduced frequency of epileptic-like behavior. This effect was observed not only in epileptic-like brain waves that localized to the MI region expressing hM4Di but also in more severe conditions involving widespread cortical regions, including the contralateral side (secondary generalized seizures) (Fig. 5).

Fig. 5.

An example of chemogenetic seizure control in monkeys. A. Schematic showing the position of the ECoG (electrocorticography) electrode array covering the primary motor cortex and the site at which the epilepsy-inducing agent (bicuculline) was delivered. B. Intensity maps showing the strength of cortical surface potentials recorded from 64 sites on the array electrode (top) and typical waveforms recorded from the site of bicuculline injection (bottom). After bicuculline injection, the waveform pattern shifted from spikes (ii) to multi-spike wave complexes (iii), and then to status epilepticus (iv), with an increase in amplitude and spread over the cortex, accompanied by seizure-like behaviors. When DCZ was administered in this state (red rectangle), the large amplitude of spikes that had spread over the cortex disappeared immediately, and the amplitude near the site of bicuculline injection decreased significantly (v–vii), along with a reduction in the frequency of seizure-like behaviors (adapted from Miyakawa et al., Nat. Commun. 2023).⁶⁷⁾

Another study employed a chemogenetic approach to activate striatal D1 dopamine receptor-expressing medium spiny neurons (D1MSNs) to effectively rescue core motor symptoms of PD.⁶⁸⁾ In that study, monkeys received injections of a retrograde AAV into the substantia nigra to express Gs-DREADD (rM3Ds), an engineered G protein-coupled receptor, specifically in D1MSNs driven by a D1 neuron-specific promoter (G88P7). Subsequently, the monkeys received an injection of a dopaminergic toxin into the substantia nigra to induce symptoms characteristic of PD. These symptoms were consistently rescued by continuous treatment with DCZ for 8 months. This demonstrates the potential of DREADDs to regulate specific neural pathways for therapeutic purposes, paving the way for circuit-specific modulatory approaches to treat a range of brain disorders.

Moreover, recent studies in macaques and marmosets have demonstrated that DCZ can activate DREADDs even when taken orally,^44),69) indicating the potential for chronic manipulation in freely moving monkeys without the need for restraints. For example, in macaques with hM4Di expressed in the DLPFC, repetitive oral DCZ administration impaired spatial working memory for up to 2 weeks. Notably, spatial working memory recovered on the day following the final dose, indicating that over 2 weeks of continuous DREADD is feasible without desensitization.⁶⁹⁾ Because this approach can be applied without causing pain or stress to monkeys, the chemogenetic approach using DCZ should expand the opportunities for acute and chronic manipulation of specific circuits in the primate brain. Thus, it holds great promise for the translational use of DREADD technology, especially in the development of long-term therapeutic interventions for neurological and neuro-psychiatric disorders.

To advance clinical applications, the resolution of certain issues, including the safety of gene delivery and DCZ, will be crucial. The use of techniques such as transient passage of the blood–brain barrier by focused ultrasound following intravenous AAV administration offers a promising avenue for noninvasive local gene delivery in humans.⁷⁰⁾ Given that DREADDs utilize human-derived genes and that PET imaging for gene expression is accessible in humans, the potential for clinical applications in humans is substantial. This innovative approach is generating momentum toward further clinical applications, with the potential to revolutionize the treatment of neurological disorders through gene therapy.

4. Limitations of imaging-based chemogenetics

Although imaging-based chemogenetics offers significant advantages in primate research, certain limitations need to be considered to fully understand the scope and applicability of these techniques.

4.1. Technical constraints.

PET imaging is a valuable method for verifying gene delivery in vivo, but it is subject to several technical constraints. To accurately confirm DREADD expression using [¹¹C]DCZ, it is essential to consider the effect of baseline ligand binding, which can introduce noise.³²⁾ This can be most effectively managed by creating subtraction images between pre- and post-vector scans.⁷¹⁾ In addition, compared with magnetic resonance imaging (MRI), PET has inferior spatial resolution, with a standard animal scanner (micro-PET) showing a resolution of ～1.5 mm. These technical constraints can limit the sensitivity and precision needed to detect low levels of DREADD expression, particularly when visualizing projections from DREADD-positive neurons. Imaging has been successful in certain cortico-subcortical pathways, but visualizing cortico-cortical projections remains limited to cases where the terminals are dense, such as in contralateral projections. As a result, DCZ-PET cannot achieve the same spatial resolution and sensitivity as anatomical tracing methods for complete in vivo anatomical projections.

4.2. Availability and accessibility.

Globally, the availability of animal PET imaging facilities is limited, constraining the widespread use of this methodology. Apart from research conducted at our institution, studies employing in vivo visualization of chemogenetic gene expression have been confined to a few PET facilities, including those at Wisconsin Madison and the National Institutes of Health (NIH, Bethesda and Baltimore campuses).^{29),35),37),72)} Furthermore, the visualization of DREADDs with [¹¹C]DCZ requires tracer synthesis near a cyclotron due to its short half-life (～20 min), which further limits accessibility. To address these challenges, ligands labeled with ¹⁸F, which has a longer half-life (～110 min), have been proposed.²⁹⁾ This allows the transport of PET ligands to PET facilities without cyclotrons and their use for the visualization DREADDs.⁷³⁾ To further support external research institutions, we have engaged in collaborative studies to visualize DREADDs in monkeys injected with AAVs at such sites. In these collaborations, monkeys are transported to our institute for PET imaging and then returned to their home institute. Such collaborations have increased in number and are expected to continue contributing significantly to the field.

4.3. Comparison with other neuromodulation techniques.

Compared with optogenetics, chemogenetics has a lower temporal resolution for neuronal manipulation. In primate research, optogenetics allows precise manipulation of neural activity at a specific timing of behavioral or cognitive processing, revealing a causal link between the coding of target neurons and the behavioral impact of their activities.^74)-76) Another limitation of chemogenetics is that systemic agonist administration does not allow flexibility to change manipulation targets within the same individual, limiting its utility for mapping the effects of local neural activity manipulation. For such purposes, alternative techniques, such as electrical stimulation, local chemical injection, or ultrasound, may prove more suitable.

Overall, although imaging-based chemogenetics provides unique opportunities and advantages in primate neuroscience, researchers must carefully consider these limitations to effectively leverage the full potential of this technology.

5. Future directions

With ongoing advances in primate gene-delivery techniques and the development of high-performance agonists, the use of DREADDs in primate neuroscience research is expanding. The widespread use of cell type- and neural pathway-specific manipulations with DREADDs is expected to open new avenues for understanding the neural basis of higher brain functions in nonhuman primates, particularly functions and behaviors that rely on complex network circuitry among several brain regions.

Beyond DREADDs, other chemogenetic tools have also been developed, including those based on engineered ion channels, which are known as pharmacologically selective actuator modules/pharmacologically selective effector molecules (PSAMs/PSEMs).⁷²⁾ For example, an excitatory PSAM channel (PSAM4-5HT3) is a chimeric protein made from a modified α7 nicotinic acetylcholine receptor ligand-binding domain that is fused to the cation-permeable ion-pore domain of a serotonin receptor. Recent research has successfully integrated imaging-based chemogenetics using the PET ligand [¹⁸F]ASEM with these tools, enabling the reversible control of neural activity via PSAM4-5HT3 in the monkey brain.⁷⁷⁾ Combining PSAMs/PSEMs with DREADDs will permit the multiplexing of chemogenetics, enabling manipulation of multiple brain circuits in a single monkey.

In addition, the unique manipulative properties of DREADDs, such as remote control in an unrestrained state or chronic manipulation of specific brain circuits in monkeys, allow entirely new research designs that have not been possible before. For example, a researcher can co-locate individual animals with DREADDs expressed in social brain circuits with normal individuals in a large room to observe effects on social behavior. The key to these new studies will be to capture functional changes upon intervention in specific neural systems, which will require not only hypothesis-testing approaches using existing behavioral tasks but also the use of methods that capture behavioral changes in a hypothesis-free manner. Techniques such as video-based behavioral measurement^78)-80) and data-driven behavioral description algorithms^81),82) may prove effective. Indeed, one such algorithm has quantified and visualized the behavioral effects of chemogenetic neural manipulations in marmosets, synthesizing motion motifs without prior knowledge of when and how these effects occur.⁸²⁾

Our institute is one of the few facilities in Japan capable of applying PET imaging to basic primate research, along with advanced multimodal imaging facilities such as MRI and CT. This unique environment has enabled us to pioneer imaging-guided chemogenetics, facilitating the optimization and application of chemogenetics in nonhuman primates. As pioneers and leading experts in this field, we remain fully committed to the challenge of further enhancing the specificity, reliability, and efficacy of chemogenetics to advance the dissection of the nonhuman primate brain. Recognizing the importance of sharing these technical advancements with the community, we actively promote the widespread application of chemogenetics to primate research through collaborations, technical reports, and workshops.

DREADDs offer great potential for investigating the pathophysiology of psychiatric and neurological disorders by modeling them in nonhuman primates. Being able to record whole-brain activity using fMRI allows researchers to identify the networks that underlie human diseases. By adding to that the ability to manipulate the networks in primate disease models, researchers can establish links between changes in brain information processing and behavior. In addition, the clinical application of chemogenetics for the treatment of human brain diseases is promising. The success of DREADD-mediated on-demand suppression of epilepsy and long-term treatment of PD in monkeys indicates the potential applicability of this approach to the treatment of diverse brain disorders.

Acknowledgments

We thank all the members of our laboratory and collaborators who have been involved in our chemogenetic studies. The chemogenetics work in our laboratory is partially supported by MEXT/JSPS KAKENHI Grant Numbers JP19K08138, JP20H05955, JP23K27098, JP24K02347, and JP24H00069; by JST PRESTO Grant Number JPMJPR22S3; and by AMED Grant Number JP23wm0625001.

Conflict of interest

TM and YN are named as inventors on a patent for using DCZ as a DREADD actuator and a PET reporter probe (WO2019/245047).

Non-standard abbreviation list

AAV

adeno-associated virus

CNO

clozapine-N-oxide

D1MSN

D1 dopamine receptor-expressing medium spiny neuron

dCD

dorsal part of the caudate nucleus

DCZ

deschloroclozapine

DLPFC

dorsolateral prefrontal cortex

DREADD

designer receptors exclusively activated by designer drug

fMRI

functional magnetic resonance imaging

MDl

lateral part of the thalamic mediodorsal nucleus

MI

primary motor cortex

MRI

magnetic resonance imaging

OFC

orbitofrontal cortex

PD

Parkinson's disease

PET

positron emission tomography

PSAM/PSEM

pharmacologically selective actuator modules/pharmacologically selective effector molecule

rmCD

rostromedial caudate nucleus

SI

primary somatosensory cortex

Profile

Takafumi Minamimoto was born in Osaka in 1971. He earned his PhD in Neuroscience from the Osaka University Graduate School of Engineering Science in 2002. Following his doctoral studies, he conducted postdoctoral research at Kyoto Prefectural University of Medicine in the Department of Physiology and later at the National Institutes of Health in the United States. In 2008, he joined the Molecular Imaging Center at the National Institute of Radiological Sciences as a Senior Researcher and was promoted to Team Leader in 2010. In 2016, he was appointed Team Leader of the Systems and Neural Circuits Team at the National Institutes for Quantum Science and Technology (QST), where he advanced to Group Leader in 2019. As of 2024, he serves as Deputy Director of the Advanced Brain Imaging Center at QST. Minamimoto is recognized for his pioneering work in primate chemogenetics and neuroimaging. He has received several prestigious awards, including the Minister of Education, Culture, Sports, Science, and Technology Award in 2021 and the Tsukahara Nakaakira Memorial Award in 2022.