Technical Comments

Comment on “Impaired Respiratory and Body Temperature Control Upon Acute Serotonergic Neuron Inhibition”

See allHide authors and affiliations

Science  10 Aug 2012:
Vol. 337, Issue 6095, pp. 646
DOI: 10.1126/science.1221810


Ray et al. (Reports, 29 July 2011, p. 637) assume that clozapine-N4-oxide (CNO) represents a “biologically inert synthetic ligand” that selectively activates the M4 muscarinic receptor-based DREADD (designer receptor exclusively activated by a designer drug). In contrast, due to the redox cycling of CNO with clozapine and to their cell membrane permeability, CNO is biologically active and its conversion products are capable of undermining DREADD effects.

The DREADD (designer receptor exclusively activated by a designer drug) technology represents a new method with an enormous potential to study cell-type-specific signaling in vivo. The designer receptor here consists of a modified human muscarinic receptor that is activated by clozapine-N4-oxide (CNO) instead of acetylcholine.

Ray et al. (1) assume that CNO represents a “biologically inert synthetic ligand” that selectively activates the inhibitory designer receptor (Di) despite several contrary observations (25). Indeed, inertness and specificity of CNO are essential requirements not only for the DREADD method but also for the applied transgenic mouse model that should have been controlled by Ray and co-workers (1).

This comment addresses various biological effects of the redox partners of CNO and of derived products that may interfere with the results of Ray et al. (1). All points to which we object here were neither controlled nor discussed by the authors.

N-oxidation of tertiary amines like clozapine (CN) and subsequent reduction back to the corresponding base have been established as common routes of biotransformation (6). In mammals, about 40% of the injected CNO, which is not renally eliminated, is rapidly enzymatically as well as nonenzymatically reduced back to CN (3, 4), putatively accompanied by generation of reactive oxygen species (ROS) (Table 1). Various, partly overlapping, tertiary amine N-oxide reduction routes were observed in rodents, including mice (4). Accordingly, mice reduce CNO-like 4-methyl-piperazinyl-4-N-oxides of antipsychotic agents efficiently back to their parent compounds (7).

Table 1

Additional CNO effects based on CNO-CN redox cycling, probably confounding the results of Ray et al. (1). mPFC, medial prefrontal cortex; GSK3β, glycogen synthase kinase 3β; GABA, (gamma)-aminobutyric acid; GAD67, glutamic acid decarboxylase isoform of 67 kD; Kir3, G protein–gated inwardly rectifying K(+) channel 3; PPAR, peroxisome proliferator-activated receptor; SREBP, sterol regulatory element-binding protein; LXR, liver X receptor; AhR, aryl hydrocarbon receptor.

View this table:

In NMRI mice, intravenous injection of either [11C]CN or [11C]CNO revealed that both penetrated the blood-brain barrier within 10 min, but the maximal brain uptake was 4 to 24 times as high for [11C]CN (8). Ray et al. (1) injected CNO as a single dose, which initially should lead to about 40% reduction of CNO to CN (2, 7).

In human liver in vitro, CNO is retroreduced to CN in the presence of the reduced form of nicotinamide adenine dinucleotide phosphate. This conversion is inhibited by ascorbic acid (3). CN is metabolized mainly to N4-desmethylclozapine (NDMC) (3), a discrete drug with a unique receptor-binding profile (9), including wild-type M4mAChR (5). NDMC interacts with the N-methylscopolamine binding site of all M receptor subtypes, albeit with lower affinity than CN, and is the only clozapine-like compound with partial agonism at the M1 receptor (9). Moreover, all “clozapine-like compounds”—CNO, CN, and NDMC—represent efficacious and potent muscarinic M4 DREADD receptor agonists, with concentration-response curve potency (pEC50) and maximal agonist response (Emax) values being highest for CN in agonist-stimulated phosphorylation of extracellular signal–regulated kinase (ERK) 1/2 (5). Ray et al. (1) did not control for exclusive binding of CNO in the presence of CN, NDMC, and derivatives, especially in the brain, where at least the lipophilic CN transiently accumulates (2, 8, 9).

The quantitative reduction of CNO to CN opens the broad field of receptor-mediated, as well as intracellular signaling pathways of CN and of NDMC. CN, as well as NDMC, binds to a huge number of neuroreceptors, thereby influencing not only the muscarinergic but also the dopaminergic, adrenergic, histaminergic, and serotonergic system. Experimental studies with rats revealed a CN-induced hypothermia because of its dopaminergic properties either as D1 receptor agonist (10), D3 receptor agonist (11), or 5-hydroxytryptamine (5-HT) receptor antagonist (12). Hypothermia is also a known side effect in CN-treated schizophrenic patients. Besides CN-induced hypothermia, Monda et al. (12) studied sympathetic nerve activity in rats. Thirty to 60 minutes after CN injection, the sympathetic firing rate decreased to 50 to 60%, exactly the level reached by DREADD technique in Ray et al. (1).

We suppose that the hypothermia of the Di mice is triggered by CN/CNO/NDMC-mediated DREADD effects, as well as by CN/CNO/NDMC-mediated dopamine receptor signaling (10, 11) and 5-HT receptor antagonism (12). In all four trials, a decrease of core body temperature in the control siblings shortly after CNO injection [figure 4A in (1)] may reflect the dopaminergic agonism (10, 11) as well as the serotonergic antagonism of CN (12) and is comparable to the 1 to 1.5°C decrease in body temperature observed previously after CN injection (10, 12). Decisively, the decreased body temperature in the control siblings is induced by CN, not by CNO (10), and thereby confirms CNO retroreduction in the Di mice. Unfortunately, Ray et al. (1) did not report body-temperature profiles of the animals before CNO injection [figure 4 in (1)]. Admittedly, the direct CN effect is not as spectacular as the DREADD effect, but CN/NDMC accumulate in the body and may participate in the unexplained severity adaptation [figure 4A in (1)]. In addition, Ray et al. (1) omitted a detailed consideration about changes in behavior and physiological factors, which are important in thermogenesis.

Besides the extracellular receptor-binding properties of CN and its metabolites, they are all cell membrane permeable (13) and brain penetrant (8, 9). Focusing on Ray et al. (1), proposed targets of the clozapine-like compounds, their effects, and resulting confounders are summarized in Table 1. Among these listed effects, the inhibition of Kir3 by CN is probably most relevant, because neuronal silencing relies on Gi protein-coupled muscarinic subtype 4 designer receptor (hM4Di)–mediated Kir3 activation (14).

Although many intracellular CN effects take hours, and the spontaneous reverse reduction of CNO may be negligible in short-term bioassays lasting only minutes, Houseknecht et al. (15) and others found immediate perturbations of several metabolic pathways such as hyperglycemia and dyslipidemia after a single parenteral administration of CN. Instead, Ray et al. (1) attributed all effects—even the longer-lasting ones (>30 min) or the effects in vivo, which include the hepatic first-pass effect—exclusively to the CNO-Di receptor signaling. They did not control any biological or pharmacological parameters concerning Di receptor–independent CNO/CN effects, although CN itself—without DREADD technology—is able either to modulate or to mimic all the effects observed by the authors, not to mention additional effects triggered by dirty conversion products of CN (3).

In conclusion, scientists working with CNO receptor tools should be on the alert for the redox cycling of CNO as well as for CNO-derived products—CN, NDMC, further metabolites, and ROS/oxidative stress—that are all capable of undermining the designer receptor–based cell and effect specificity.


View Abstract


Navigate This Article