The designer receptors exclusively activated by designer drugs (DREADD) system can be used to suppress nociceptors, resulting in analgesia to heat, reports a new mouse study. But thanks to a set of careful control experiments, researchers led by Brian Davis and Michael Gold, University of Pittsburgh, US, also find an important caveat: DREADD expression alters endogenous channel activity and signaling pathways in sensory neurons independent of the ligand used to activate the receptor. Though these findings were not unforeseen, it’s the first time research has demonstrated such unintended consequences of DREADD technology, and it illustrates the care that pain researchers and, indeed, any researcher using this technology, must take when interpreting results from experiments using DREADDs.
“It’s not at all surprising that expression of the inhibitory DREADD [used in the study] produced a modest off-target effect. Anytime you do anything to a cell, there are unintended changes that take place. I think the important lesson here is to always perform appropriate control experiments,” said Bryan Roth, University of North Carolina at Chapel Hill, US, who invented the DREADD technology but was not involved in the new study.
The new findings were published October 19 in the Journal of Neuroscience.
Using DREADDs to inhibit pain
Following up on their earlier studies using optogenetics to silence sensory neurons (Baumbauer et al., 2015), the research team decided to use DREADDs as a chemogenetic approach to suppress nociceptor function. DREADDs are engineered G protein-coupled receptors that are activated by presumably inert drug-like small molecules (Roth, 2016). They have been used in over 2,000 published studies to date and are becoming an increasingly popular tool in the pain field (Grace et al., 2016). The particular DREADD used in the new study has been used extensively in the central nervous system and in one prior study to suppress nociceptor function (Iyer et al., 2016).
“As a natural outgrowth of our optogenetic studies, we wanted to have chemical control and the ability to modulate pain thresholds over a longer period of time. To do that we used the DREADD system to make animals in which we could depress primary afferents in order to explore different aspects of pain,” explained Davis.
In the new study, the researchers used a mutated muscarinic acetylcholine receptor (Gi-DREADD, also known as human muscarinic receptor 4 or hM4Di) that initiates an inhibitory G protein signaling cascade to decrease neuronal firing and signaling when activated by the ligand clozapine-N-oxide (CNO). The cascade exerts its inhibitory effect through two main second-messenger mechanisms: 1) the alpha subunit of the Gi protein leads to an inhibition of adenylate cyclase activity, enabling the inhibition of pro-nociceptive signaling, and 2) the beta and gamma subunits activate G protein inwardly rectifying potassium channels and/or inhibit voltage-gated calcium channels to decrease excitability and synaptic transmission.
To express the Gi-DREADD selectively in C fibers, the researchers crossed mice expressing the Cre recombinase enzyme under control of the transient receptor potential vanilloid type 1 (TRPV1) promoter with mice in which Gi-DREADD expression is Cre dependent. They termed the resulting offspring V1Gi-DREADD mice.
To determine the effect of the inhibitory DREADD on heat thresholds, co-first authors Jami Saloman and Nicole Scheff and colleagues used the Hargreaves’ assay for heat sensitivity. Here, a beam of light is directed at the animal’s hindpaw, and the length of time until the paw is withdrawn is measured.
While baseline heat thresholds in both male and female V1Gi-DREADD mice did not differ from control animals (littermates lacking the TRPV1-Cre allele), V1Gi-DREADD mice exhibited a significantly longer paw withdrawal latency in response to radiant heat than controls one hour after intraperitoneal injection of CNO, an effect that lasted for three hours. Together, these results indicated that DREADD inhibition of nociceptors produces an analgesic effect in response to heat. As expected based on previous capsaicin ablation studies, experiments with von Frey filament stimulation showed that V1Gi-DREADD mice had normal mechanical sensitivity.
To explore the mechanism by which C fiber inhibition produced heat analgesia, the researchers next assessed primary afferent excitability in dissociated dorsal root ganglion (DRG) neurons from V1Gi-DREADD mice and controls. Whole-cell current-clamp experiments revealed that CNO application in V1Gi-DREADD mice reduced action potential firing, which was associated with hyperpolarized membrane potential, decreased input resistance, and the activation of a low-threshold potassium channel.
No CNO … uh-oh
However, control experiments revealed a problematic and, the researchers believe, underappreciated phenomenon: V1Gi-DREADD mice exhibited alterations in several membrane properties even in the absence of CNO; this indicates that expression of the Gi-DREADD alone altered some normal channel properties of the sensory neurons. Specifically, the researchers found a smaller action potential overshoot and longer action potential duration in neurons from V1Gi-DREADD mice compared to those from controls.
“The DREADDs worked just like we thought they would to produce analgesia. Unfortunately, Gi-DREADD expression also had unintended consequences,” Gold told PRF.
The results prompted the research team to embark on a thorough investigation. “When we first got an inkling that Gi-DREADD expression was having ligand-independent effects, we felt it was very important to determine exactly what was going on before investing substantial resources in using this approach to silence sensory neurons,” Davis explained.
Several subsequent experiments in dissociated DRG neurons revealed that Gi-DREADD expression in the absence of CNO suppressed voltage-gated calcium and sodium currents, and increased sodium channel (Nav1.7) expression. The team also found altered excitatory and inhibitory second-messenger signaling, including a suppression of adenylate cyclase signaling.
In a final experiment, the group examined the functional consequences of these alterations in endogenous signaling. Intraplantar injection of prostaglandin E2 produced thermal hypersensitivity, an effect that was blocked by the mu-opioid receptor agonist DAMGO, in control but not V1Gi-DREADD mice. This indicated that Gi-DREADD expression was associated with a reduction in the normal analgesic effect resulting from activation of an endogenous inhibitory G protein-coupled receptor.
Not so unexpected?
“Gi-DREADD expression really produced a lot of changes in these neurons. In retrospect, that shouldn’t have been so surprising, given that proteins in a membrane intricately communicate with a variety of signaling molecules,” said Gold. “If you insert a whole bunch of G protein in the membrane, it’s going to start influencing how all the membrane proteins talk to each other,” he explained.
“What was most striking about these mice is that phenotypically they looked pretty normal, despite all of the changes brought on by the Gi-DREADD expression, which speaks to the fact that [in the face of change] the nervous system does what it can to regain homeostasis,” Gold continued. “It wasn’t until we really pushed the system that we saw all of the attempted compensations. That’s really the problem.”
Roth noted that while the current study is the first report of unintended effects of DREADD expression, the field of optogenetics has seen several reports of similar unintended effects, such as those resulting from heat generated by optical stimulation (Stujenske et al., 2015; Otchy et al., 2015).
Though Roth was unsurprised by the results, he was pleased that such a careful characterization of the CNO-independent effects was conducted. He also suggested a possible solution (Roth, 2016). “By using a different promoter or viral construct to lower levels of [Gi-DREADD] expression, these sorts of effects should be eliminated,” Roth explained.
However, given the relatively small increase in nociceptive threshold observed in response to CNO, Gold predicts that, although a lower level of expression may reduce the level of compensatory changes, it would also be unlikely to produce enough DREADD to create the intended effect.
Based on the new results, the researchers have decided not to continue using the V1Gi-DREADD mice. “They’re not going to be a good tool to answer the kinds of mechanistic questions that we want to ask,” Davis said. “If you are interested in how an insult, injury, or disease affects how a cell functions, DREADDs should not be in your toolbox, unless one is going to do comprehensive, deep phenotyping that few investigators can commit to these days.”
But that doesn’t mean that DREADDs can’t be used to silence sensory neurons, he continued. For example, Davis also studies how sensory neurons influence the development of pancreatic cancer. Since his goal is to look at how pancreatic cells respond in the presence or absence of sensory neuron activity and he’s focusing on cancer progression, he believes DREADDs would be useful to silence sensory neurons in that case.
Allison Marin, PhD, is a neuroscientist-turned-science writer who resides in Pittsburgh, US.
Image credit: Jami Saloman