Technical Comments

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

Science  10 Aug 2012:
Vol. 337, Issue 6095, pp. 646-647
DOI: 10.1126/science.1222519

Abstract

Löffler et al. highlight the important potential of designer receptors exclusively activated by a designer drug (DREADD)–based technologies to study cell type–specific functions but cautions that the triggering DREADD ligand, clozapine-N-oxide (CNO) might, through potential conversion products, have bioactivity outside of the synthetic DREADD receptor system and maintains that Ray et al. did not control for such activity. We recount controls used in our work that indicated no discernible DREADD-independent effects of CNO on the homeostatic assays employed and discuss in this regard murine studies reporting CNO bioneutrality in other assays, the rapid renal clearance of N-oxides like CNO, and evidence of negligible conversion to clozapine (CN) in mice and rats.

In a 2011 study (1), we introduced a new Cre-dependent genetic tool, RC::PDi, and its parental intersectional (Cre- and Flp-dependent) allele, RC::FPDi, for targeting the inhibitory designer receptor exclusively activated by a designer drug (DREADD) (Di) technology (2, 3) to highly resolved cell populations in the mouse. We used this tool to determine involvement of serotonergic neurons in physiological homeostasis. Löffler et al. (4) comment on the important potential of such DREADD-based technologies, yet caution about a potential for CNO to have bioactivity outside the synthetic Di receptor system, in part through potential retroreduction of CNO to clozapine (CN) and/or CN to N4-desmethylclozapine (NDMC). Löffler maintains that Ray et al. did not control for Di-independent effects by CNO and its potential conversion products in the reported experiments. For clarification of this point, we emphasize here the critical CNO controls performed and how these controls indicate negligible Di-independent bioactivity of CNO or its potential derivatives in the assays employed. We also discuss work indicating rapid renal clearance of N-oxides (e.g., CNO) as compared with tertiary amines (e.g., CN) (5, 6) and the negligible back-conversion of CNO to CN in mice and rats (7, 8). Retroconversion may be held in check in mice by high tissue levels of ascorbate, an inhibitor of CNO retroreduction (7, 9) that is endogenously synthesized in mice (10); by contrast, humans and guinea pigs [the subjects of studies (8) cited by Löffler et al. as evidence for conversion of CNO to CN in mammals] have lost the capacity to synthesize ascorbate, rely on dietary intake, and typically have lower tissue levels and thus perhaps greater capacity for retroconversion (11).

Di is an engineered Gi/o protein-coupled receptor (3) that, upon binding of CNO, can trigger cell-autonomous hyperpolarization and diminished cell excitability (2). We developed a Cre-dependent allele RC::PDi and partnered it with the Cre transgenic, Slc6a4-cre (or RC::FDi with Pet1::Flpe) to drive Di expression in serotonergic neurons (1). We established that this system offers specific CNO/Di-dependent suppression of serotonergic neuron action potential firing and found that respiratory and body-temperature control was impaired dramatically in a Di-dependent fashion after a single CNO intraperitoneal injection, but not in CNO-treated control mice (non-Di–expressing single transgenic Slc6a4-cre or RC::PDi siblings).

In each experiment performed by Ray et al. (1), the effects of CNO administration alone were examined as controls for the experimental conditions, in contrast to Löffler et al.’s statement that CNO controls were not performed or discussed. In all experiments, non-Di–expressing single transgenic (RC::PDi or Slc6a4-cre) sibling controls were assessed for the pertinent physiological parameters before and after CNO injection. For example, in assessing the potential for CNO alone or its metabolites to affect the central respiratory CO2 chemoreflex [figure 3C in (1)], we measured minute ventilation in room air and then in 5% CO2 before CNO injection, and then again in the same animals after injection during systemic CNO exposure. Respiratory responses pre-CNO and during CNO exposure were indistinguishable in these control animals, showing a normal doubling of ventilation in response to 5% CO2 as compared with room air. We concluded that in this assay, CNO (or any conversion products) had no appreciable effect on the respiratory CO2 chemoreflex as measured. Further, ventilatory responses of control animals (with and without CNO) were also indistinguishable from the pre-CNO response of the experimental Di-expressing mice (double-transgenic Slc6a4-cre, RC::PDi mice), indicating that Di expression alone (without the triggering ligand CNO) also did not affect the respiratory CO2 chemoreflex. It was only the combination of Di expression and CNO administration that resulted in an attenuated respiratory reflex [figure 3B in (1)].

Similar neutrality of CNO alone or Di alone was observed with respect to body-temperature regulation. While housed at an ambient room temperature of ~23°C, control single-transgenic non-Di–expressing siblings had an average body temperature of 37.9° ± 0.2°C before CNO administration [figure 4 and supporting online material for (1)] and an average maximal low of 36.7° ± 0.3°C 2 hours after the CNO injection, indicating an average maximal change of ~1.2°C from the pre-injection baseline. Such ~1°C fluctuations occurred repeatedly in control mice throughout the course of the 20-hour assay and fall within reported normal variation for wild-type mice and healthy humans, including changes associated with handling and subsequent acclimation during temperature measurements (12, 13). Also exhibiting body temperatures within the normal range were double-transgenic Di-expressing experimental animals (Slc6a4-cre, RC::PDi mice) before CNO administration (36.9° ± 0.2°C). By contrast, extreme hypothermia ensued after CNO administration to these Di-expressing experimental mice, with body temperatures plummeting to 27.1° ± 0.9°C after CNO injection, an ~10°C drop, while being maintained at an ambient room temperature of ~23°C. Given that this extreme hypothermia was specific to Di-expressing mice after CNO administration and was not exhibited by sibling controls in response to CNO, we concluded that Di-mediated effects produced the extreme hypothermia.

Although Löffler et al. concur that this extreme body-temperature drop is Di-dependent, he comments that the body-temperature variation of ~1.2°C observed in control animals after handling, CNO injection, and temperature probing is a decisive indication that retroreduction of CNO to CN has occurred. In this regard, Löffler et al. cite the work by Salmi and Ahlenius (14) and Monda et al. (15) that shows a temperature drop of up to ~3°C (38.5 to 35.5°C and 37.5 to 36°C, respectively) after direct CN injection in rats. Yet, Salmi and Ahlenius (14) also showed in the same article that there is no significant drop in temperature in response to CNO (or NDMC) injection, as compared with saline, using doses equivalent to our experiments. Thus, the temperature drop of ~3°C cited by Löffler et al. (14) is elicited only if rats are directly injected with CN but not CNO, consistent with the negligible CNO effect seen in our assays using mice. Moreover, Guettier et al. (7) show no significant retroreduction of CNO to CN in mice over a 2-hour period, as determined by blood plasma liquid chromatography tandem mass spectrometry, and Jann et al. (8) report no detectable CN after CNO administration in rats. Additionally, Guettier et al. (7) show that by 30 min the plasma concentration of CNO is reduced by half, and by 2 hours it is absent, presumably by renal clearance. Indeed, N-oxides (e.g., CNO), due to their marked polarity, are substantially more soluble than tertiary amines (e.g., CN) and are readily excreted (5, 6).

Löffler et al. cite Jann et al. (8) as evidence that retroconversion of CNO to CN occurs in mice and rats, yet Jann et al. (8) state the opposite, that CN from its N-oxide metabolite CNO was not found to occur in rats, but rather was found in guinea pigs and humans. Thus, not all mammals convert CNO equivalently. Löffler et al. note that CNO is retroreduced by human liver microsomal extracts in the presence of the reduced form of nicotinamide adenine dinucleotide phosphate (NADPH) and that this retroconversion is inhibited by ascorbate (7, 9). We suggest that inhibition by ascorbate may offer an explanation for the negligible levels of CN detected after CNO administration in mice and rats. Mice and rats endogenously synthesize ascorbate; humans and guinea pigs have lost this capacity and rely fully on dietary sources. Steady-state liver ascorbate levels are typically 3 to 4 times as high in mice as in humans, varying with age and diet (1618). These higher ascorbate levels in mice may efficiently inhibit retroconversion of CNO to CN.

In summary, we found that attenuation of the respiratory chemoreflex and extreme hypothermia were dependent on Di expression coupled with CNO administration in double-transgenic Slc6a4-cre, RC::PDi mice; these phenotypes were not observed after CNO administration to control Di-negative siblings. CNO neutrality has also been observed in control animals subjected to a host of other physiological and behavioral assays in mice (7, 1921). We appreciate the opportunity here to recount our experimental design details and, like Löffler et al., we stress the necessity of controlling for effects potentially mediated by CNO alone and/or its metabolites.

References and Notes

  1. Acknowledgments: Supported by NIH grants F32HD063257 (R.R.), F31NS073276 (R.B.), R21DA023643 (R.B. and S.D.), R21MH083613 (R.R., J.C.K., and S.D.), P01HD036379 (R.R., A.C., R.B., J.C.K., E.N., G.R., and S.D.), R01HL028066 (A.C., and E.N.), and R01HD052772 (G.R.).
View Abstract

Subjects

Navigate This Article