The experimental procedures were performed in accordance with European Community (EU Directive 2010/63/EU for animal experiments) and National Institutes of Health guidelines for the care and use of laboratory animals. In accordance with the German Protection of Animals Act (TierSchG §4 Abs. 3) procedures were approved by the Animal Welfare Office of University Medical Center Gottingen (file number ID T12/18).

The knockout mouse model for RTT, strain B6.129P2(C)-Mecp2tm1-1Bird (Guy et al., 2001; maintained on a C57BL/6J background) was purchased from The Jackson Laboratory (Bar Harbor, ME, USA). Hemizygous mutant Mecp2^−/y males were generated by crossing heterozygous Mecp2^+/− females with C57BL/6J wild-type males. The genotyping was performed in accordance with The Jackson Laboratory genotyping protocols¹.

The knockout mouse model for the 5-ht_5b receptor, strain 129SvEvBrd (maintained on a 29/SvEv-C57BL/6 background) was obtained from Taconic Europe A/S (Tornbjergvej 40, Ejby, 4623 Lille Skensved, Denmark). Htr5b knockout mice were backcrossed into the C57BL/6 background for at least eight generations. Mecp2^−/y;Htr5b^−/−-double-knockout mice were then generated by crossbreeding female Mecp2^+/−;Htr5b^+/− mice with male Mecp2^+/y;Htr5b^−/− mice.

Tissue samples were incubated in 25 mM NaOH/0.2 mM EDTA for 3 h at 65°C. After neutralization with an equal volume of 40 mM Tris/HCl pH 5.5, 1 μL was taken as a template for subsequent PCR. The primers used for the identification of Htr5b-genotype were: WT-for (5′-ctctgcagtcggtttgatg-3′), WT-rev (5′-gtagagtcaccacaagcac-3′), KO-for (5′-gcagcgcatcgccttctatc-3′), KO-rev (5′-gtgctgggattagaagtcc-3′). The primers used for the identification of the Mecp2 genotype were: WT-for (5′-gaccccttgggactgaagtt-3′), KO-for (5′-ccatgcgataagcttgatga-3′) and WT-KO-rev (5′-ccaccctccagtttggttta-3′).

Ventilation was measured by unrestrained whole-body-plethysmography (Drorbaugh and Fenn, 1955; Bartlett and Tenney, 1970) in 40-day old mice (P40). Mice were placed in a custom-made acrylic glass chamber (300 ml) that was connected to a differential low-pressure transducer (model DP1 03, Validyne Engineering, Northridge, CA, USA). The second channel of the pressure transducer was connected to a reference chamber (300 ml). The signal from the pressure transducer was fed into a sine wave carrier demodulator (CD-15, Validyne Engineering). Animals could explore the chamber freely. For the analysis, pressure changes were band-pass filtered (1.5–500 Hz) and amplified (four times) before storing on an Apple-PC computer. For digitization (1 kHz sampling rate) an ITC-16 interface (InstruTECH/HEKA, Lambrecht) was used that was controlled by Axograph 4.8 software (Axon Instruments, Foster City, CA, USA). A bias flow of 150 ml/min was introduced using a Normocap^® CO₂-sensor (Datex, Instrumentarium Oy, Helsinki, Finland). Pressure changes were exported and converted to axon binary files and used for analysis. Breaths (Inspiratory flow peaks) from a period of 3 min after 12 min adaptation to the chamber were analyzed by the threshold search peak detection method of Axon Clampfit (Molecular Devices, Sunnyvale, CA, USA). We did not discriminate between respiratory cycles associated to different types of behavior e.g., sniffing. Breathing frequencies were calculated as the reciprocal of the averaged inspiratory peak interval. The number of intervals between inspiratory peaks that were longer than 1 s was determined during the 3 min as a parameter for central apneas (Stettner et al., 2008). To define the regularity of breathing, the coefficient of variation was calculated for the interval (CV = STD/mean) as well as an irregularity score (IS) was calculated: IS = 100*ABS[(Int_n − Int_n-1)/Int_n-1] for each respiratory cycle (Barthe and Clarac, 1997; Telgkamp et al., 2002).

The cAMP concentration of murine tissue was determined using commercially available ELISA-based DetectX-kit (Arbor Assays) and performed according to manufacturer’s instructions. Absorbance was measured at 450 nm using Infinite 200 Pro reader (TECAN). Whole brains were explanted from 40 days old mice, separated and individual regions snap-frozen in liquid nitrogen.

The cAMP measurements and plethysmography were analyzed with non-parametric Kruskal-Wallis and subsequent Dunn’s post hoc test for multiple comparisons. The survival plots were analyzed using the log-rank (Mantel-Cox) and Gehan-Breslow-Wilcoxon tests. Analyses were performed using GraphPad Prism version 5.0d for Mac OSX (GraphPad Software, San Diego, CA, USA). Differences were considered statistically significant at P < 0.05. Data are presented as the mean ± standard error of the mean (SEM; n = number of experiments).

Mecp2^−/y;Htr5b^−/− mice were generated by crossbreeding female Mecp2^+/−; Htr5b^+/− mice with male Mecp^2+/y; Htr5b^−/− mice. To our knowledge, Htr5b^−/− mice have no observable phenotype (see www.taconic.com and Mouse Genome Database, http://www.informatics.jax.org/marker/MGI:96284). In our hands, the life span of Htr5b^−/− mice was normal and body weight at P40 (19.06 ± 1.12 g) was indistinguishable from WT mice (20.60 ± 0.56 g) of the same age. Unrestrained whole-body plethysmography at postnatal day 40 (P40) revealed that Htr5b^−/− mice were breathing normal, which is in line with the notion that 5-ht_5b-receptors are downregulated and thus dispensable at this age (Figure 1A). Like WT mice (n = 9), Htr5b^−/− mice (n = 7) did not show long (longer than 1 s) breathing arrests (Figure 1B), which are typical for Mecp2^−/y, and both, breathing rate (WT 6.68 s⁻¹ ± 0.41 vs. Htr5b^−/− 6.69 s⁻¹ ± 0.36; n.s.; Figure 1C) and IS (WT 0.23 ± 0.03 vs. Htr5b^−/− 0.28 ± 0.02; n.s.; Figure 1D) were indistinguishable from WT.

When comparing breathing of Mecp2^−/y;Htr5b^−/− mice (n = 11) with Mecp2^−/y mice (n = 10), we found significantly improved breathing parameters, yet no full rescue. Unrestrained whole-body plethysmography at P40 revealed a slow irregular breathing rate in Mecp2^−/y mice with a high number of breathing arrests (5.3 ± 2.6 per 180 s, Figure 1B), low respiratory rate (4.6 ± 0.36 s⁻¹; Figure 1C) and high IS (0.46 ± 0.06; Figure 1D). However, the breathing rhythm of Mecp2^−/y;Htr5b^−/− mice was more stable with intermediate values between Htr5b^−/− (or WT) and Mecp2^−/y mice. Mecp2^−/y;Htr5b^−/− mice showed a significantly lower number of breathing arrests longer than 1 s (0.5 ± 0.4 per 180 s; Figure 1B), and a higher respiratory rate (6.23 ± 0.65 s⁻¹; Figure 1C), and the IS (0.33 ± 0.03; Figure 1D) was indistinguishable from WT and Htr5b^−/− mice. Moreover, body weight of Mecp2^−/y;Htr5b^−/− mice (17.52 ± 0.75 g) was higher than age-matched Mecp2^−/y mice (12.31 g ± 0.8; p < 0.001) but did not reach the level of WT and Htr5b^−/− mice. The life span of Mecp2^−/y;Htr5b^−/− mice was improved compared to Mecp2^−/y mice (median 80 vs. 40 days), however no Mecp2^−/y;Htr5b^−/− survived longer than 130 days (Figure 2).

Since 5-ht_5b receptors are highly expressed in the brainstem of MeCP2-deficient mice at P40 (Vogelgesang et al., 2017) and 5-ht_5b receptors are able to reduce cellular cAMP levels in vitro (Niebert et al., 2017), we measured the cAMP concentration in medullary brainstem lysates. The cAMP concentration of Mecp2^−/y mice at P40 was significantly lower (73.49 ± 7.27%) when compared to WT mice (Figure 3), whose cAMP levels were set as 100%. This reduction is in line with previous reports of low cAMP in the brainstem of Mecp2^−/y mice (Mironov et al., 2011). In agreement with the hypothesis of constitutive 5-ht_5b-receptor signaling we found normal cellular cAMP in Mecp2^−/y; Htr5b^−/− mice (98.46 ± 6.34%; Figure 3). These data are in line with the concept that the constitutive activity of 5-ht_5b is impairing cAMP signaling in Mecp2^−/y mice (Vogelgesang et al., 2017).

5-ht_5b receptors are not targeted to the plasma membrane but to the endosomal compartment, where they can participate in signaling by binding to G_i proteins (Vogelgesang et al., 2017) and lowering cAMP levels (Niebert et al., 2017). Lower cAMP-levels in Mecp2^−/y mice provide a simplistic rationale for the clinical use of cAMP-elevators like theophylline, as the “first choice for respiratory stimulation” in Rett patients (Julu et al., 2008). In addition, the resulting imbalance of cAMP-dependent second messenger cascades is likely to affect not only the serotonergic system (Manzke et al., 2003) but other neurotransmitters that have been implicated in the modulation of the respiratory network (Fujii et al., 2004; Lalley, 2008; Viemari, 2008; Mellios et al., 2014). In this respect, other effects on cAMP may seem contradictory: application of 5-HT_1A receptor agonists, which reduces cAMP, has been reported to have positive effects on respiration in a Rett mouse model (Abdala et al., 2010, 2014) and patients (Andaku et al., 2005; Ohno et al., 2016). However, interactions between different neuromodulator pathways are manifold and even two G_i-mediated pathways can be antagonistic if restricted to different cellular compartments or located in distinct types of neurons (Manzke et al., 2010). Additionally, 5-HT_1A receptors can modulate neuronal activity independently from cAMP, e.g., by modulating potassium currents (Penington et al., 1993). Although we found a net reduction of cAMP in the medulla, we cannot rule out an additional effect of the 5-ht_5b dependent reduction of 5-HT_1A receptor surface expression (Niebert et al., 2017). However, this mechanism appears rather unlikely to be of major importance, since 5-HT_1A knockout mice have only a mild respiratory phenotype (Barrett et al., 2012). Moreover, 5-HT_1A receptor agonists have diverse effects including an anxiolytic action that, in the light of a potential role of the anxiety level of Mecp2^−/y mice (Ren et al., 2012), might influence respiratory rhythm without directly influencing respiratory neurons.

The fact that the improvement of breathing does not completely rescue the life expectancy of the mice is, although disappointing, in line with earlier observations, as for the restoration of MeCP2 in inhibitory neurons (Hülsmann et al., 2016). In Mecp2^−/y;Htr5b^−/− double-knockout mice, body weight was still reduced compared to WT and Htr5b^−/− mice and, although increased, the lifespan of Mecp2^−/y;Htr5b^−/− double-knockout mice was still significantly shorter than the lifespan of WT mice. Since we did not find a dysregulation of the 5-ht_5b-receptor expression in the hippocampus (Vogelgesang et al., 2017), any hippocampal pathology, which is e.g., involved in the increased propensity to seizures (Boison, 2012), is not cured in Mecp2^−/y;Htr5b^−/− mice and, thus, can manifest later during the development of Mecp2^−/y mice leading to still premature death.

Although the persistent expression of 5-ht_5b contributes to the respiratory phenotype in MeCP2-deficient mice, these findings cannot be immediately translated to human patients. Unlike in mice (Grailhe et al., 2001; Maekawa et al., 2010; Vogelgesang et al., 2017), it is accepted that stop codons in the human HTR5B gene prevent its expression (Grailhe et al., 2001). However, we found that 5-ht_5b in mice is also expressed as a truncated protein (Vogelgesang et al., 2017). As nothing is known about the expression and potential splice variants of 5-ht_5b in humans, further analysis is required to identify potential alterations of 5-ht_5b receptor expression in the brainstem of patients with MECP2 mutations.

MeCP2 deficiency affects several 100 targets (Ben-Shachar et al., 2009) so any effect of 5-ht_5b-receptor and cAMP signaling must be seen in the context of multiple other dysregulated genes. Our data supports the notion that 5-ht_5b-receptor dysregulation is an important but probably not the only factor that contributes to respiratory problems in Mecp2^−/y mice. However, currently no data is available indicating a relevance of the 5-ht_5b receptor expression in the pathology of human RTT patients.