Antiemetic effects of baclofen in a shrew model of postoperative nausea and vomiting: Whole‐transcriptome analysis in the nucleus of the solitary tract

Abstract Aims The molecular genetic mechanisms underlying postoperative nausea and vomiting (PONV) in the brain have not been fully elucidated. This study aimed to determine the changes in whole transcriptome in the nucleus of the solitary tract (NTS) in an animal model of PONV, to screen a drug candidate and to elucidate the molecular genetic mechanisms of PONV development. Methods Twenty‐one female musk shrews were assigned into three groups: the Surgery group (shrew PONV model, n = 9), the Sham group (n = 6), and the Naïve group (n = 6). In behavioral studies, the main outcome was the number of emetic episodes. In genetic experiments, changes in the transcriptome in the NTS were measured. In a separate study, 12 shrews were used to verify the candidate mechanism underlying PONV. Results A median of six emetic episodes occurred in both the Sham and Surgery groups. Whole‐transcriptome analysis indicated the inhibition of the GABAB receptor‐mediated signaling pathway in the PONV model. Baclofen (GABAB receptor agonist) administration eliminated emetic behaviors in the shrew PONV model. Conclusions Our findings suggest that the GABAB receptor‐mediated signaling pathway is involved in emesis and that baclofen may be a novel therapeutic or prophylactic agent for PONV.


| INTRODUC TI ON
Postoperative nausea and vomiting (PONV) is a serious and frequent complication in patients after they emerge from general anesthesia.
Approximately 30% of all patients experience PONV, and in a subpopulation of high-risk patients, the incidence increases to 80%. 1 The latest guideline suggests that risk factors for PONV in adults are female sex, type of surgery (e.g., laparoscopic or gynecological surgery), volatile anesthesia, a history of motion sickness or PONV during previous surgeries, non-smoking status, long surgery duration, younger age, and postoperative opioids. 2 However, the central nervous system mechanisms that underlie the effects of these emetic stimuli on the development of PONV have not been fully elucidated.
Animal studies are rarely used in emetic research, in part because standard laboratory animals, such as rats and mice, are not able to vomit. 3,4 Although ferrets are most commonly used in emetic research, inhalation anesthetics do not induce nausea and vomiting in the ferrets. 5 Musk shrews, Suncus murinus, are now being used in emetic research at several laboratories. 4,[6][7][8][9] Emetic behavior can be evoked in this small animal by various stimuli, including motion, nicotine, and inhalational anesthetics. 4,[6][7][8][9] Despite recent advances in shrew research, no studies have thoroughly examined how inhalation anesthetics and surgical insults influence the development of PONV in shrew models.
In the human brain, the nucleus of the solitary tract (NTS) in the brainstem converges the primary afferents by emetic stimulation from the vestibular system, abdominal vagal afferents, and the area postrema; sends outputs to the somatosensory/viscerosensory cortex via the parabrachial nucleus and the thalamus for the induction of nausea sensation; and sends outputs to the gastrointestinal system and the respiratory system for induction of retching or vomiting. Previous anatomical and physiological studies have indicated that the NTS plays a central role in processing emetic information. 3 The aims of this study were to determine changes in genome-wide gene expression (i.e., whole-transcriptome analysis) in the NTS in a musk shrew PONV model and elucidate the molecular genetic mechanisms of PONV development.

| Animals
This study was approved by the Tohoku University Institutional Animal Care and Use Committee (#2017-236). No specific inclusion or exclusion criteria were set. In total, 33 female musk shrews were used (Jic:SUN-Her/Kwl strain, aged 7-10 weeks, weighing 30-50 g).
The animals were housed individually in 35 cm × 30 cm × 17 cm plastic cages on soft bedding, maintained at 22 ± 2°C with a 12-h light/ dark cycle, and provided with trout pellets and tap water ad libitum.
Efforts were made to reduce both animal numbers and suffering during the experiments. This research was reported in accordance with the ARRIVE guideline. 10

| Low abdominal surgical procedure
Laparotomy was performed as described previously, with slight modification. 11,12 Briefly, the lower abdomen of the shrew was shaved and disinfected with povidone-iodine, and a 1-cm vertical incision was made under 5% isoflurane anesthesia in 1 L/min of oxygen, administered via a facemask. Intestinal paralysis was induced by manipulating the small intestine vigorously with a cotton swab for 30 s. The abdominal muscle and skin were then closed, applying three sutures to each layer using 3-0 braided nylon sutures (Surgilon; Covidien Ltd., Minneapolis, MN). The entire surgical procedure was completed within 10 min.

| Experimental groups
To determine the effects of isoflurane and low abdominal surgery on shrews, the animals were randomly assigned to three groups: the Surgery group (shrew PONV model, n = 9), treated by incision and suturing of the lower abdomen under 5% of isoflurane inhalation, as described above; the Sham group (n = 6), treated only by shaving the lower abdomen under 5% isoflurane inhalation for 10 min; and the Naïve group (n = 6), which received no treatment.
In a separate study, 12 shrews were randomly assigned to two groups and treated as follows: the Baclofen group (n = 6), lower abdominal surgery followed by single-dose 5 mg/kg intraperitoneal injection of baclofen (a GABA B receptor agonist); and the Vehicle group (n = 6), surgery followed by intraperitoneal injection of normal saline alone. The doses of drugs were determined in preliminary experiments based on previous reports. 13,14 Baclofen or normal saline was administered at 3 min before the start of the emetic behavioral test.

| Emetic behavioral test protocol
The shrews were housed in the institutional rearing area for more than 7 days, and each shrew was transported from the home cage to an observation chamber (cylinder-shaped, 20 cm in diameter ×30 cm in height, made from clear acrylic), where it was allowed free movement for habituation for 30 min. The shrew was then placed in a transparent induction chamber (box-shaped, 15 cm × 10 cm × 10 cm). Shrews in the Surgery or Sham group were anesthetized by isoflurane in 3 L/min of oxygen, with the concentration of isoflurane increased stepwise from 1% to 4% in increments of 1% during 5 min.
After the withdrawal of reflexes to pinching of the tail and the hind paws, the shrew was placed on a surgical table under 5% isoflurane anesthesia with 1 L/min of oxygen administered via a facemask.
Shrews in the Surgery group then underwent the surgical procedure described above, whereas those in the Sham group received only hair shaving and disinfection of the lower abdomen. In the Naïve group, the shrews were exposed to 3 L/min of oxygen in the induction chamber without isoflurane or surgery.
After the treatment, the shrew was transferred back to the ob-

| NTS isolation followed by total RNA extraction
At 1 h after the start of exposure to oxygen with or without isoflurane, each shrew was euthanized via CO 2 exposure. The shrews were rapidly decapitated with a laboratory guillotine, and the brain stem was immediately removed within 10 min. According to the stereotaxic atlas of the rat brain and shrew, 15 the NTS was dissected from a frozen section of the brain stem, and total RNA was extracted from NTS using NucleoSpin RNA/Protein Kit (Macherey-Nagel, Düren, Germany).

| Whole-transcriptome sequencing (RNA-seq)
The density of extracted RNAs was measured using a Qubit3.0 The unique index sequences were incorporated in the adaptors for multiplexed high-throughput sequencing. The final product was assessed for its size distribution and concentration using BioAnalyzer High Sensitivity DNA Kit (Agilent Technologies). Pooled libraries were diluted to 2 nM in EB buffer (Qiagen, Hidden, Germany) and then denatured using the Illumina protocol. The denatured libraries were diluted to 10 pM by pre-chilled hybridization buffer and loaded onto a TruSeq v2 Rapid flow cell on an Illumina HiSeq 2500 and run for 50 cycles using a paired-read recipe according to the manufacturer's instructions.

| Computational bioinformatic analysis
The RNA-seq reads were checked for quality by using FastQC ver. 0.11.4. These reads were mapped to the shrew reference genome (Ous:KAT-227c strain, CDS +UTR sequence, Suncus murinus Genome Project in Japan, unpublished draft sequence) by using Bowtie2 ver. 2.3.4.1. 16 The mapped reads were assembled and annotated by using TIGAR2 ver. 2.1 supplied by Suncus murinus Genome Project protein-coding gene annotation file. 17 Differential expressed transcripts (DEX) were compared between the two selected groups by using edgeR ver. 3.5.0. 18 Significance was defined as results with a q value of less than 0.05 calculated by the Benjamini-Hochberg method to control the false discovery rate (FDR). MA plots were generated by using the plotSmear function of edgeR software.
With respect to the DEX between the Surgery and Naïve groups (see the Results section), gene symbol names of top-100 differentially expressed transcripts with fold changes and p values were analyzed with ingenuity pathway analysis (IPA, version 60467501, release date: Nov 11, 2020, Qiagen). Duplicated gene symbol names were excluded to avoid the co-existence of the upregulated and downregulated transcripts with the same symbol name. Significance was determined using the right-tailed Fisher's exact test (p < 0.01) according to the manufacturer's instructions. 19 To decipher possible functions of the genes, the gene symbol names with statistical significance in IPA were analyzed by web-based gene ontology enrichment analysis (g:GOSt, https://biit.cs.ut.ee/gprof iler). 20 Functional information for humans was used, and significance was determined using the g:SCS algorithm (q < 0.01).

| In the behavioral study, isoflurane alone, or surgical insult under isoflurane anesthesia, induced emesis in shrews
We developed a shrew PONV model in the current study (see  identified as DEX between the Sham and the Surgery groups ( Figure 2B). We identified 19 differentially expressed transcripts, including SDC3, KIF1C, and ITM2C genes, among the Naïve and Surgery groups ( Figure 2C). As shown in Figure 2D, the highest DEX existed between the Surgery and the Naïve groups.
We focused on the combined effects of surgical insult and inhalation anesthetic on PONV, which may be important in the clinical setting because general anesthesia is not applied to patients alone without another procedure, such as surgery. receptor, adenylyl cyclase, and voltage-dependent calcium channel as candidate molecules associated with PONV. Figure 3 summarized GABBR1, ADCY1, and CACNB1 gene expression levels in the pathway from the results of RNA-seq in the shrew NTS ( Figure 3A and B).

| GABA B receptor agonism produced the elimination of emetic behaviors in shrews
We  Figure 4B).
The latency to the first emetic episode was 343 (254) s ( Figure 4C).
The latencies to the start of the walk in the Vehicle and Baclofen groups were 225 (34) and 250 (160) seconds, respectively, and there was no difference between the two groups ( Figure 4D, p = 1.00). All measurement values showed a non-Gaussian distribution. The frequencies of emetic episodes were similar in the Sham and Surgery groups (see the equal medians as shown in Figure 1).

GO molecular function ID Number of annotated genes q value
Nevertheless, in the Surgery group, the frequencies varied and tended to decreased, as demonstrated by the height of the box in the bar for the Surgery group in Figure 1. Surgical insult may be a preventable factor for emesis of shrews. In humans, type of surgery (e.g., laparoscopic surgery) and duration of surgery are risk factors for PONV in adults. 2 The results in the current behavioral study conflicted with clinical observations. However, the association of surgical insult with PONV remains unknown.
In RNA-seq analysis, changes in the transcriptome after isoflurane anesthesia (Figure 2A) and after isoflurane anesthesia followed by surgery ( Figure 2B) overlapped for a few genes, that is, SDC3, KIF1C, and ITM2C ( Figure 2C). The SDC3 gene encodes syndecan protein, which is involved in the cytoskeleton structure. The KIF1C gene encodes kinesin proteins, which function as a cell microtubuledependent molecular motor. The ITM2C gene encodes an integral F I G U R E 3 (A) Schematic drawing of inhibition of the GABA B receptormediated signaling pathway in the shrew PONV model. In shrews showing emetic behaviors, expression of the ADCY1 gene, which encodes adenyl cyclase (orange), was increased, whereas expression of the GABBR1 and CACNB1 genes, which encode GABAB receptor R1 subunit (green) and voltage-gated calcium channel (blue), was decreased. Administration of pharmacological baclofen (a GABA B receptor agonist) eliminated emetic behaviors in the shrew PONV model. (B) Expression levels of the GABBR1, ADCY1, and CACNB1 genes in wholetranscriptome sequencing. Bar graphs and error bars indicate means and standard errors of the means. The gene expression of GABBR1 and CACNB1 was not observed in the Surgery group. FPKM value: fragments per kilobase of exon per million mapped fragments for quantifying the assembled transcript expression membrane protein that functions as a regulator for amyloidβ protein. Although syndecan protein may be involved in obesity or appetite, the associations of all three genes with emesis are still unknown. 26,27 For DEX between the Naïve and Surgery groups, the other 15 genes, except for SDC3, KIF1C, and ITM2C genes, were highly expressed with statistical significance ( Figure 2C, red dots with red circles). This observation suggested that surgical insult and inhalation anesthetics were not independent factors. As described above, gynecological surgery under general anesthesia, which is modeled by the Surgery group in this study, is a significant risk factor for PONV in clinical situations. 2 Taking into consideration clinical settings, the downstream analysis focused on the changes in the wholetranscriptome in the NTS between the Naïve and Surgery groups.
The following computational analysis using the IPA indicated the associations of the six biological pathways, 10 genes, and three upstream regulators (Table 1). Also, the gene ontology enrichment analysis showed a total of 21 GO terms ( Table 2). We interpreted ubiquitous terms, such as "Protein binding (GO:0005515)" or "Cellular response to stimulus (GO:0051716)," as nonspecific function in the NTS. Therefore, we focused on "GABA receptor signaling" in Table 1 and "G protein-coupled receptor signaling pathway (GO:0007187)," "Plasma membrane region (GO:0098590)," or "Synapse (GO:0045202)" in Table 2  reported that GABA B receptor agonists improve the symptoms of motion sickness. 33 One possible explanation for these previous observations is that baclofen stimulated the GABA B receptor-mediated signaling pathway in the NTS as the vomiting center in the brain and that the efferent outputs from the NTS to induce vomiting were strongly attenuated. Hence, interestingly enough, our findings indicate that baclofen may be useful for prevention and treatment for PONV.
One of the limitations of this study was that we did not measure the nausea sensation of the shrews. Unfortunately, it is difficult to evaluate nausea in animal models owing to our inability to communicate with these animals. However, a recent report suggests that a novel behavior, fruit-flavored water avoidance, reflects nauseaassociated behaviors of mice. 34 This methodology may be used in future experiments in shrews.
Another limitation of the current study is that we did not measure protein expression levels in the NTS, such as GABA B receptors or voltage-dependent calcium channels. However, almost every protein experiment needs a protein-specific antibody. Although many of these commercially available antibodies show cross-reactivity among species, including humans, rats, and mice, the reactivity of available antibodies with shrew brain tissues has not been verified.
In the current study, we performed some western blotting experiments; however, the primary antibodies did not show reactivity in the shrew brain (data not shown). Differences in the amino acid sequences of proteins among species can result in differences in antibody reactivity, making further molecular experiments difficult.

| CON CLUS IONS
We established the shrew model of PONV. The RNA-seq quantified the expression level of 52,381 transcripts at the genome-wide scale. We focused on the relationship between the GABA B receptormediated signaling pathway and PONV. Baclofen, a GABA B receptor agonist, eliminated emetic behaviors in our shrew PONV model.
These new findings suggested that baclofen may be a novel therapeutic or prophylactic agent for PONV, particularly for emesis occurring after gynecological surgery.

ACK N OWLED G EM ENTS
We are deeply grateful to Prof. Hideki Noguchi, Center for Genome

CO N FLI C T S O F I NTE R E S T
The authors declare no competing interests.