The Pivotal role of Glycogen Synthase Kinase 3 (GSK-3) in vomiting evoked by specific emetogens in the least shrew (Cryptotis parva)

Abstract

Glycogen synthase kinase 3 (GSK-3) is a constitutively active multifunctional serine-threonine kinase which is involved in diverse physiological processes. GSK-3 has been implicated in a wide range of diseases including neurodegeneration, inflammation, diabetes and cancer. GSK-3 is a downstream target for protein kinase B (Akt) which phosphorylates GSK-3 and suppresses its activity. Based upon our preliminary findings, we postulated Akt’s involvement in emesis. The aim of this study was to investigate the participation of GSK-3 and the antiemetic potential of two GSK-3 inhibitors (AR-A014418 and SB216763) in the least shrew model of vomiting against fully-effective emetic doses of diverse emetogens, including the nonselective and/or selective agonists of serotonin type 3 (e.g. 5-HT or 2-Methyl-5-HT)-, neurokinin type 1 receptor (e.g. GR73632), dopamine D₂ (e.g. apomorphine or quinpirole)-, and muscarinic 1 (e.g. pilocarpine or McN-A-343) receptors, as well as the L-type Ca²⁺ channel agonist (FPL64176), the sarco/endoplasmic reticulum Ca²⁺-ATPase inhibitor thapsigargin, and the chemotherapeutic agent, cisplatin. We first determined if these emetogens could regulate the phosphorylation level of GSK-3 in the brainstem emetic loci of least shrews and then investigated whether AR-A014418 and SB216763 could protect against the evoked emesis. Phospho-GSK-3α/β Ser21/9 levels in the brainstem and the enteric nerves of jejunum in the small intestine were upregulated following intraperitoneal (i.p.) administration of all the tested emetogens. Furthermore, administration of AR-A014418 (2.5–20 mg/kg, i.p.) dose-dependently attenuated both the frequency and percentage of shrews vomiting in response to i.p. administration of 5-HT (5 mg/kg), 2-Methyl-5-HT (5 mg/kg), GR73632 (5 mg/kg), apomorphine (2 mg/kg), quinpirole (2 mg/kg), pilocarpine (2 mg/kg), McN-A-343 (2 mg/kg), FPL64176 (10 mg/kg), or thapsigargin (0.5 mg/kg). Relatively lower doses of SB216763 exerted antiemetic efficacy, but both inhibitors barely affected cisplatin (10 mg/kg)-induced vomiting. Collectively, these results support the notion that vomiting is accompanied by a downregulation of GSK-3 activity and pharmacological inhibition of GSK-3 protects against pharmacologically evoked vomiting.

1. Introduction

The emetic nuclei involved in vomiting include the brainstem dorsal vagal complex (DVC) [area postrema (AP), nucleus tractus solitarius (NTS) and dorsal motor nucleus of the vagus (DMNX)], as well peripheral loci such as neurons of the enteric nerves system (ENS) and enterochromaffin cells (EC cells) which are embedded in the lining of the gastrointestinal tract (GIT), as well as vagal afferents which carry input from the GIT to the brainstem DVC (Babic and Browning, 2014; Darmani and Ray, 2009; Ray et al., 2009).

Cisplatin-like cancer chemotherapeutics cause vomiting via release of multiple neurotransmitters [e.g. dopamine, serotonin (5-HT), substance P, etc] from the EC cells and/or the brainstem via a Ca²⁺-dependent process (Darmani et al., 2014). Specific emetogens such as selective or non-selective agonists of serotonin type 3 (5-HT₃R) (e.g. 2-Methyl-5-HT or 5-HT)-, substance P neurokinin type 1 (NK₁R) (e.g. GR73632)-, dopamine D₂ (D₂R) (e.g. quinpirole or apomorphine)-, and muscarinic 1 (M₁R) (McN-A-343 or pilocarpine)-receptors, as well as Ca²⁺ channel regulators comprising the L-type Ca²⁺ channel (LTCC) agonist FPL64176 (Darmani et al., 2014), and the sarco/endoplasmic reticulum Ca²⁺-ATPase (SERCA) inhibitor thapsigargin (Zhong et al., 2016), evoke robust vomiting in vomit-competent species. Based on our Ca²⁺-dependent emesis hypothesis (Zhong et al., 2017), we have demonstrated the broad-spectrum antiemetic nature of two of the selective LTCC agonists, nifedipine and amlodipine, against the above discussed diverse emetogens (Darmani et al., 2014, Zhong et al., 2014a, Zhong et al., 2016).

Cancer chemotherapy is undergoing a paradigm shift from cytotoxic/radiotherapy towards targeted therapy or combination of these modalities. One targeted modality involves antagonism of cancer-promoting receptor tyrosine kinases, and/or inhibition of components of their common downstream intracellular signals, i.e., the phosphoinositide 3-kinase/protein kinase B (PI3K/Akt) pathway, which is a driver system for cancer growth, metastasis and motility (Revathidevi and Munirajan 2019). Recently our laboratory has focused on post-receptor intracellular emetic signaling pathways. Our findings have implicated the Ca²⁺-Ca²⁺/calmodulin kinase II- extracellular signal-regulated protein kinase1/2 (ERK1/2) cascade in the brainstem and gut is one major common mechanism in the regulation of emetic responses elicited by administration of several emetogens including: GR73632, FPL64176, thapsigargin, and 2-Methyl-5-HT (Zhong et al., 2019; 2018; 2016; 2014b), and as well as increased ERK phosphorylation in response to cisplatin administration in our least shrew emesis model (Darmani et al., 2013). In addition, Akt phosphorylation occurs in the shrew brainstem following vomiting evoked by GR73632 (Zhong et al., 2019) or FPL64176 (Zhong et al., 2018).

Glycogen synthase kinase-3 (GSK-3), a multi-functional serine-threonine kinase, is constitutively involved in diverse physiological processes, including metabolism, cell cycle, and gene expression (Luo, 2009; Khan et al., 2017; Matsuda et al., 2019; Saraswati et al., 2018; Walz et al., 2017). GSK-3 is also involved in a wide range of pathologies such as diabetes, inflammation, cancer, neurodegeneration and mental illness (Sahin et al., 2019). GSK-3 is encoded by two known genes, GSK-3α and GSK-3β. GSK-3 is a downstream target protein for Akt signaling pathway. Activation of Akt signaling can be followed by phosphorylation GSK-3α/β at Ser21/9 and its subsequent inactivation (Matsuda et al., 2019). In the present study, we sought to determine: i) if GSK-3 phosphorylation is involved in vomiting evoked by diverse emetogens, and ii) whether pharmacological inhibition of GSK-3 with AR-A014418 (Mazzardo-Martins et al 2012) and SB216763 (Coghlan et al., 2000) has antiemetic potential in the least shrew emesis model. Our results support the hypothesis that pharmacological inhibition of GSK-3 may be useful for the prevention of vomiting.

2. Materials and methods

2.1. Animals

A colony of adult least shrews from the Western University of Health Sciences Animal Facilities were housed in groups of 5–10 on a 14:10 light/dark cycle, and were fed and watered ad libitum. The experimental shrews were 45–60 days old and weighed 4–6 g. Animal experiments were conducted in accordance with the principles and procedures of the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All protocols were approved by the Institutional Animal Care and Use Committee of Western University of Health Sciences. All efforts were made to minimize animals suffering and to reduce the number of animals used in the experiments.

2.2. Chemicals

The following drugs were used for the present studies: serotonin HCl (5-HT), 2-Methyl-serotonin maleate salt (2-Methyl-5-HT), apomorphine HCl, quinpirole HCl, pilocarpine, McN-A-343, and the chemotherapeutic agent cisplatin (cis-platinum (II) diamine dichloride (Pt(NH3)2)Cl2) were obtained from Sigma/RBI (St. Louis, MO). The LTCC agonist FPL64176, NK₁ receptor agonist GR73632 and GSK-3 inhibitors AR-A014418 and SB216763 were purchased from Tocris (Minneapolis, MN) and the SERCA inhibitor thapsigargin from Santa Cruz. 5-HT, 2-Methyl-5-HT, apomorphine, quinpirole, McN-A-343, pilocarpine and GR73632 were dissolved in distilled water. AR-A014418, SB216763 and FPL64176 was dissolved in 25% DMSO. Thapsigargin was dissolved in 10% DMSO in water. Cisplatin was dissolved in water by sonication. All drugs were administered intraperitoneally at a volume of 0.1 ml/10 g of body weight.

2.3. Experimental procedures

2.3.1. Animals

On the day of the experiment shrews were brought from the animal facility, separated into individual cages and allowed to adapt for at least two hours (h). Daily food was withheld 2 h prior to the start of the experiment but shrews were given 4 mealworms each prior to emetogen injection, to aid in identifying wet vomits as described previously (Darmani, 1998).

2.3.2. Western blots

To determine the time-dependent profile of GSK-3α/β phosphorylation in response to the above emetogens, different groups of animals (n = 3/group) were sacrificed at 5, 15 and 30 min following administration of each emetogen at their fully effective emetic doses. In the case of cisplatin, we have previously demonstrated that cisplatin (10 mg/kg., i.p.) causes emesis in all tested least shrews over 40 h with respective peak early and delayed vomiting phases occurring at 1–2 and 32–34 h post-injection (Darmani et al., 2017; 2015; 2013; 2009). Thus, shrews were decapitated at the indicated time points (i.e. 5 min, 30 min, 1h, 2h, 3h, 4h, 5h, 24h, 28h, 33h and 40 h) after cisplatin injection (n = 3/group). Shrew brainstems were then collected and homogenized in lysis buffer. Protein extracts from brainstem lysates were subjected to Western blot. The primary antibodies including phospho-GSK-3α/β (Ser21/9) rabbit antibody (1:1000, #9331, Cell Signaling) and GSK-3α/β mouse antibody (1:2000, 44–610, Invitrogen) and secondary antibodies including goat anti-rabbit IRDye 680RD and goat anti-mouse IRDye 800CW (1:10000, LI-COR) were used for Western blots. The blots were visualized correspondingly using Odyssey imaging system. The ratios of phospho-GSK-3α/β (pGSK-3α/β) (~ 50 kD) to GSK-3α/β to GSK-3α/β were calculated. All values were divided by the average value at time point 0 min (vehicle control) for normalization and presented as fold change of control.

2.4. Behavioral emesis studies

To determine the broad-spectrum dose-response antiemetic potential of the GSK-3 inhibitor AR-A014418, varying doses were injected (i.p.) into different groups of shrews at 0 min (n = 6–12/group). Thirty minutes later, different groups of AR-A014418-pretreated shrews were tested for vomiting with a fully efficacious dose of one of the following emetogens (Darmani et al., 2014; Zhong et al., 2016; 2014a): i) a peripherally-acting non-selective 5-HT₃R agonist, 5-HT (5 mg/kg, i.p.), ii) a centrally/peripherally-acting and more selective 5-HT₃R agonist 2-Methyl-5-HT (5 mg/kg, i.p.), iii) a selective NK₁R agonist GR73632 (5 mg/kg, i.p.), iv) a non-selective dopamine D₂R agonist apomorphine (2 mg/kg, i.p.), v) a dopamine D₂ receptor-preferring agonist quinpirole (2 mg/kg, i.p.), vi) a non-selective muscarinic receptor agonist pilocarpine (2 mg/kg, i.p.), vii) a more selective M₁R agonist McN-A-343 (2 mg/kg, i.p.), viii) the LTCC agonist FPL64176 (10 mg/kg, i.p.), xi) the SERCA inhibitor thapsigargin (0.5 mg/kg., i.p.). The antiemetic efficacy of GSK-3 inhibitor SB216763 was also tested against the above emetogens. The vomiting behavior (number of animals vomiting within groups and frequency of vomits) were then observed for 30 min. Each shrew was used once and then euthanized with isoflurane following the termination of each experiment.

2.5. Statistical analysis

The vomit frequency data were analyzed using the Kruskal-Wallis non-parametric one-way analysis of variance (ANOVA) followed by Dunnett’s post hoc test and expressed as the mean ± SEM. The percentage of animals vomiting across groups at different doses was compared using the Chi-square test. Statistical significance for differences between two groups was tested by unpaired t-test. After Western blots, when more than two groups were compared, a one-way ANOVA was used followed by Dunnett’s post hoc test to determine statistical significance between experimental groups and corresponding control. p < 0.05 was considered statistically significant.

3. Results

3.1. Administration of diverse emetogens increases GSK-3α/β phosphorylation in a time-dependent pattern in the least shrew brainstem

In the current study least shrews were treated (i.p.) for various periods (5, 15 and 30 min) with a variety of emetogens. Using the anti-phospho-GSK-3α/β (Ser21/9) and anti-GSK-3α/β antibodies, phosphorylation states of GSK-3α/β in the brainstems were assessed by Western blot. The total levels of GSK-3α/β under all conditions remained similar (Fig. 1A). In response to 5-HT (5 mg/kg) and McN-A-343 (2 mg/kg), phosphorylation of GSK-3α/β rapidly and significantly peaked at 5 min post-injection (p = 0.0431 and 0.0282 vs. 0 min control, respectively) and then quickly returned to basal levels (Fig. 1B). Following 2-Methyl-5-HT (5 mg/kg) administration, increased phosphorylation of GSK-3α/β was first observed at 5 min (p = 0.0018) which persisted up to 15 min (p = 0.0047), but not at 30 min post injection (Fig. 1B). In the case of GR73632 (5 mg/kg), apomorphine (2 mg/kg), quinpirole (2 mg/kg) and thapsigargin (0.5 mg/kg), phosphorylation of GSK-3α/β was triggered at 15 min (p = 0.0305; < 0.0001; 0.0039; 0.0106, respectively) (Fig. 1B). In addition, phosphorylation of GSK-3α/β upregulated at 15 min (p = 0.0246) and then reached a maximal value at 30 min (p = 0.0081) post pilocarpine administration (2 mg/kg). Without declining to unstimulated level (0 min), FPL64176 (10 mg/kg) evoked a stable increase in phosphorylation of GSK-3α/β at all three time points, 5 min (p = 0.0026), 15 min (p = 0.0024) and 30 min (p = 0.0008).

Fig. 1.

Effects of diverse emetogens on GSK-3 phosphorylation states in the least shrew brainstem. A) Representative Western blots for time-course of GSK-3α/β phosphorylation at Ser21/9 in the least shrew brainstems collected from either vehicle control (0 min) or at the indicated time-points after intraperitoneal (i.p.) administration of fully effective emetic doses of diverse emetogens including : i) selective and/or nonselective agonists of serotonin 5-HT₃ (e.g. 5-HT or 2-Methyl-5-HT, 5 mg/kg)-, )-, neurokinin NK₁ (e.g. GR73632, 5 mg/kg)-, dopamine D2 (e.g. apomorphine or quinpirole, 2 mg/kg)-, muscarinic M₁ (e.g. pilocarpine or McN-A-343, 2 mg/kg)-receptors; and ii) Ca²⁺ channel regulators such as the L-type Ca²⁺ channel agonist FPL64176 (10 mg/kg), and the sarco/endoplasmic reticulum Ca²⁺-ATPase inhibitor thapsigargin (0.5 mg/kg). Phospho-GSK3α/β Ser21/9 (pGSK-3α/β) and GSK-3α/β were detected by Western blot. B) Quantitative analysis of Western blots as shown in A. The ratios of pGSK-3α/β to GSK-3α/β were compared. All ratios were normalized to vehicle-treated control (0 min) values before analysis and expressed as fold change of control. *p < 0.05; **p < 0.01, ***p < 0.001, ****p < 0.0001 vs. 0 min, one-way ANOVA followed by Dunnett’s post hoc test (n = 3). C) Representative Western blots for time-course of GSK-3α/β phosphorylation at Ser21/9 in the least shrew brainstems collected from either vehicle control (0 min) or at the indicated time-points during a 40 h observation period following administration of the chemotherapeutic agent cisplatin (10 mg/kg., i.p.). Phospho-GSK3α/β Ser21/9 (pGSK-3α/β) and GSK-3α/β were detected by Western blot. D) Quantitative analysis of Western blots as shown in C. The ratios of pGSK-3α/β to GSK-3α/β were compared. All ratios were normalized to vehicle-treated control (0 min) values before analysis and expressed as fold change of control. *p < 0.05; **p < 0.01 vs. 0 min, one-way ANOVA followed by Dunnett’s post hoc test (n = 3).

3.2. Cisplatin increases phosphorylation of GSK-3α/β in the least shrew brainstem in a time-dependent manner

We assessed the phosphorylation status of GSK-3α/β at varying time periods (i.e. 5 min, 30 min, 1h, 2h, 3h, 4h, 5h, 24h, 28h, 33h and 40 h) after cisplatin (10 mg/kg) injection. We found that phosphorylation of GSK-3α/β were significantly elevated from the end of cisplatin’s acute (5 h)- to peak delayed phase (33 h) (Darmani et al., 2009) vomiting (p = 0.0494, 5 h vs. 0 min; p = 0.0016, p = 24 h vs. 0 min; p = 0.0048, 28 h vs. 0 min; p = 0.0107, 33 h vs. 0 min), and returned to basal level at 40 h (p = 0.9994, 40 h vs. 0 min) (Fig. 1C–D).

Likewise, at 5 h post cisplatin treatment, increased phospho-GSK3α Ser21 and phospho-GSK-3β Ser9 immunoreactivity occurred in shrew brainstem slices containing the dorsal vagal complex emetic nuclei (area postrema (AP), nucleus tractus solitarius (NTS) and the dorsal motor nucleus of the vagus (DMNX) (Supplemental Fig. S1).

3.3. The broad-spectrum antiemetic potential of AR-A014418 against diverse emetogens

Figure 2 demonstrates the broad-spectrum antiemetic potential of AR-A014418 against vomiting evoked by diverse emetic receptor agonists. Indeed, pretreatment with the selective GSK-3 inhibitor AR-A014418 (0, 5, 10 and 20 mg/kg), partially blocked vomiting caused by the peripherally-acting nonselective 5-HT₃R agonist 5-HT (5 mg/kg, i.p.). In fact, as shown in Fig. 2A, relative to the vehicle-pretreated control group (0 mg/kg), AR-A014418 partially reduced the mean frequency of 5-HT-induced vomiting (KW (3, 24) = 8.973, p = 0.0296) with a significant inhibition (64%) at 20 mg/kg (p = 0.0119), but failed to significantly affect the percentage of shrews vomiting (χ² (3, 24) = 7.093, p = 0.0690). However, AR-A014418 (0, 2.5, 5, 10 and 20 mg/kg,) dose-dependently reduced both the frequency (KW (4, 31) = 26.44, p < 0.0001) and percentage (χ² (4, 31) = 14.46; p = 0.006) of shrews vomiting in response to the administration of the 5-HT₃R-selective agonist, 2-Methyl-5-HT (5 mg/kg, i.p.) (Fig. 2B). A significant reduction in vomit frequency occurred at its 10 (p = 0.0003) and 20 mg/kg doses (p = 0.0001). In addition, the percentage of shrews vomiting was reduced by 50% at its 10 (p = 0.0177) and by 55.6% at 20 mg/kg dose (p = 0.0085).

Fig. 2.

Antiemetic effects of the selective GSK-3 inhibitor AR-A014418 against vomiting caused by diverse emetic receptor agonists. Varying doses of AR-A014418 (i.p.) were injected to different groups of shrews 30 min prior to an injection of a fully effective emetic dose of the peripherally-acting non-selective serotonin 5-HT₃ receptor agonist 5-HT (5 mg/kg, i.p.) (A), the centrally/peripherally-acting 5-HT₃ receptor-selective agonist 2-Methyl-5-HT (5 mg/kg., i.p.) (B), the centrally/peripherally-acting selective neurokinin NK₁ receptor agonist GR73632 (5 mg/kg, i.p.) (C), the non-selective dopamine D₂ receptor agonist apomorphine (2 mg/kg, i.p.) (D), the more selective D₂ receptor agonist quinpirole (2 mg/kg, i.p.) (E), the non-selective muscarinic M₁ receptor agonist pilocarpine (2 mg/kg, i.p.) (F), or the potent and more selective M₁ agonist McN-A-343 (2 mg/kg, i.p.) (G). Emetic parameters were recorded for the next 30 min. Frequency of emesis presented as mean ± SEM: Kruskal–Wallis non-parametric one-way ANOVA followed by Dunnett’s post hoc test. Percentage of shrews vomiting: Chi-square test. *p < 0.05, **p < 0.01 and ***p < 0.001 vs. 0 mg/kg (controls pretreated with vehicle of AR-A014418).

As shown in Fig. 2C, AR-A014418 (0, 2.5, 5, and 10 mg/kg) also caused a dose-dependent decrease in the mean frequency of vomits induced by the NK₁R selective agonist GR73632 (KW (3, 25) = 17.06, p = 0.0007) with significant reductions occurring at its 5 (p = 0.0189) and 10 mg/kg doses (p = 0.0002). In addition, a significant decrease (χ² (3, 25) = 10.11; p = 0.0176) in the percentage of animals vomiting was observed at its 10 mg/kg dose (70%; p = 0.0039) (Fig. 2C).

The antiemetic efficacy of AR-A014418 was then tested against emesis caused by the activation of dopamine D₂Rs (Figs. 2D and 2E). In fact, AR-A014418 (0, 2.5, 5, 10 and 20 mg/kg) significantly and dose-dependently reduced the frequency of vomits induced by the nonselective dopaminergic agonist apomorphine (2 mg/kg) (KW (4, 24) = 13.27, p = 0.01) (Fig. 2D). A significant reduction was seen at its 10 mg/kg dose (p = 0.0305), and complete protection against the induced vomiting occurred at its 20 mg/kg dose (p = 0.007). Likewise, the percentage of shrews vomiting was reduced in a dose-dependent fashion (χ² (4, 24) = 12.31; p = 0.0152) with a 50% reduction at its 5 mg/kg (p = 0.0455), 66.7% reduction at 10 mg/kg dose (p = 0.0143) and 100% reduction at 20 mg/kg dose (p = 0.0009) (Fig. 2D). Furthermore, AR-A014418 (0, 2.5, 5, 10 and 20 mg/kg) was assessed against vomiting caused by the more potent and selective dopamine D₂R agonist, quinpirole (2 mg/kg, i.p.) (Fig. 2E). The mean frequency of quinpirole-induced emesis (KW (4, 31) = 17.38, p = 0.0016) was significantly reduced at 10 (p = 0.0028) and 20 mg/kg doses of AR-A014418 (p = 0.0011) (Fig. 2E). Significant decreases (χ² (4, 31) = 11.92; p 0.0179) in the percentage of animals vomiting were also noted at its 10 (55.6%; p = 0.0085) and 20 mg/kg doses (83.3%; p = 0.0008) (Fig. 2E).

Next, the antiemetic potential of varying doses of AR-A014418 (0, 2.5, 5, 10 and 20 mg/kg) was tested against vomiting caused by either the non-selective muscarinic receptor agonist, pilocarpine (2 mg/kg, i.p.) (Fig. 2F), or the more selective muscarinic M₁R agonist, McN-A-343 (2 mg/kg, i.p.) (Fig. 2G). AR-A014418 reduced the frequency of pilocarpine-induced emesis in a dose-dependent fashion (KW (4, 24) = 16.62, p = 0.0023) with a significant reduction at its 10 mg/kg dose (p = 0.0079) and a near complete suppression at 20 mg/kg (p = 0.0008) (Fig. 2F). In addition, AR-A014418 dose-dependently protected shrews from vomiting (χ² (4, 24) = 10.03; p = 0.04) with 50% reduction at its 10 (p = 0.0455) and 83.3% reduction at 20 mg/kg dose (p = 0.0034) (Fig. 2F). AR-A014418 pretreatment (0, 2.5, 5 and 10 mg/kg) also significantly reduced the frequency of McN-A-343-induced emesis (KW (3, 20) = 12.29, p = 0.0065) (Fig. 2G) in a dose-dependent manner at 5 and 10 mg/kg doses (p = 0.0165 and 0.0039, respectively). The percentage of shrews vomiting was also reduced in a dose-dependent fashion (χ² (3, 20) = 9.333; p = 0.0252) with significant reductions at 5 (66.7%; P = 0.0455) and 10 mg/kg doses (83.3%; p = 0.0034) (Fig. 2G).

We then investigated the antiemetic potential of AR-A014418 against vomiting evoked by Ca²⁺ channel regulators, such as the selective LTCC agonist FPL64176 (10 mg/kg, i.p.) and the SERCA inhibitor thapsigargin (0.5 mg/kg, i.p.) (Fig. 3). AR-A014418 (0, 2.5, 5, and 10 mg/kg) significantly attenuated the mean frequency of FPL64176-induced vomiting in a dose-dependent manner (KW (3, 20) = 13.59, p = 0.0035) with a significant reduction at 5 (p = 0.024) and complete protection at its 10 mg/kg dose (p = 0.0012) (Fig. 3A). In addition, the percentage of shrews vomiting in response to FPL64176 was also suppressed by AR-A014418 (χ² (3, 20) = 12.59; p = 0.0056) with a 50% reduction at its 2.5 mg/kg (p = 0.0455), 66.7% reduction at its 5 mg/kg (p = 0.0143), and 100% reduction at its 10 mg/kg doses (p = 0.0005) (Fig. 3A). Figure 3B shows that AR-A014418 (0, 2.5, 5, 10 and 20 mg/kg) also suppressed vomiting caused by thapsigargin (0.5 mg/kg, i.p.). The frequency of thapsigargin-induced emesis (KW (4, 31) = 16.76, p = 0.0021) was significantly reduced at 10 (p = 0.013) and 20 mg/kg doses (p = 0.0008). Significant decreases (χ² (4, 31) = 12.68; p = 0.0129) in the percentage of animals vomiting were also noted at its 10 (66.7%; p = 0.0063) and 20 mg/kg doses (75%; p = 0.0019).

Fig. 3.

Broad-spectrum antiemetic effects of the selective GSK-3 inhibitor AR-A014418 against vomiting caused by the Ca^{2 +} channel regulators, such as L-type Ca^{2 +} channel agonist FPL64176 and the sarco/endoplasmic reticulum Ca²⁺-ATPase inhibitor thapsigargin. Varying doses of AR-A014418 (i.p.) were injected to different groups of shrews 30 min prior to an injection of a fully effective emetic dose of FPL64176 (10 mg/kg, i.p.) (A), or thapsigargin (0.5 mg/kg., i.p.) (B). Emetic parameters were recorded for the next 30 min. Frequency of emesis presented as mean ± SEM: Kruskal–Wallis non-parametric one-way ANOVA followed by Dunnett’s post hoc test. Percentage of shrews vomiting: Chi-square test. *p < 0.05, **p < 0.01, ***p < 0.001 vs. 0 mg/kg.

3.4. A second GSK-3 inhibitor SB216763, more potently displays broad-spectrum antiemetic efficacy

Varying doses of SB216763 (0, 0.1, 0.25, 1 and 2.5 mg/kg, i.p.) attenuated the mean vomit frequency evoked by 2-Methyl-5-HT (5 mg/kg) (KW (4, 33) = 20.59, p = 0.0004) with significant reductions at its 0.25 (p = 0.0176), 1 (p = 0.0029), and 2.5 mg/kg doses (p = 0.0046) (Fig. 4A). In addition, SB216763 protected shrew from vomiting (χ² (4, 33) = 11.39; p = 0.0225) with significant reductions in percentage of shrews vomiting at its 1 (50%, p = 0.0209) and 2.5 mg/kg doses (62.5%, p = 0.007) (Fig. 4A).

Fig. 4.

Broad-spectrum antiemetic effects of the selective and potent GSK-3 inhibitor SB216763 against vomiting caused by a variety of emetogens. A) Antiemetic effects of SB216763 against vomiting caused by the centrally/peripherally-acting selective serotonin 5-HT₃ receptor agonist, 2-Methyl-5-HT. Varying doses of SB216763 (i.p.) were injected to different groups of shrews 30 min prior to an administration of a fully effective emetic dose of 2-Methyl-5-HT (5 mg/kg., i.p.). Emetic parameters were recorded for the next 30 min. Frequency of emesis presented as mean ± SEM: Kruskal–Wallis non-parametric one-way ANOVA followed by Dunnett’s post hoc test. Percentage of shrews vomiting: Chi-Square test. *p < 0.05, **p < 0.01 vs. 0 mg/kg (control pretreated with vehicle of SB216763). B) Antiemetic effects of a 0.25 mg/kg (i.p.) dose of SB216763 against vomiting caused by maximally-effective doses of a variety of emtogens, including: the non-selective serotonin 5-HT₃ receptor agonist 5-HT (5 mg/kg, i.p.), NK₁ receptor agonist GR73632 (5 mg/kg, i.p.), the non-selective dopamine D₂ receptor agonist apomorphine (2 mg/kg, i.p.) or the more selective D₂ receptor agonist quinpirole (2 mg/kg, i.p.), the non-selective muscarinic M₁ receptor agonist pilocarpine (2 mg/kg, i.p.) or the more selective M₁ agonist McN-A-343 (2 mg/kg, i.p.), the L-type Ca²⁺ channel agonist FPL64176 (10 mg/kg, i.p.), and the sarco/endoplasmic reticulum Ca²⁺-ATPase inhibitor thapsigargin (0.5 mg/kg, i.p.). SB216763 (0.25 mg/kg., i.p.) was injected to different groups of shrews 30 min prior to an administration of a fully effective emetic dose of the emetogens. Emetic parameters were recorded for the next 30 min. Frequency of emesis presented as mean ± SEM: Unpaired t-test. Percentage of shrews vomiting: Chi-square test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 and ns (non-significant) vs. 0 mg/kg. C-D) Antiemetic effects of varying doses of SB216763 against vomiting caused by quinpirole (C) or pilocarpine (D). Varying doses of SB216763 (i.p.) were injected to different groups of shrews 30 min prior to an administration of a fully effective emetic dose of quinpirole (2 mg/kg, i.p.) or pilocarpine (2 mg/kg, i.p.). Emetic parameters were recorded for the next 30 min. Frequency of emesis presented as mean ± SEM: Kruskal-Wallis non-parametric one-way ANOVA followed by Dunnett’s post hoc test. Percentage of shrews vomiting: Chi-square test. *p < 0.05, **p < 0.01 and ***p < 0.001 vs. 0 mg/kg.

We then evaluated the antiemetic efficacy of SB216763 at 0.25 mg/kg (i.p.) against vomiting caused by maximally effective doses of diverse emetogens (5-HT), GR73632, apomorphine, quinpirole, pilocarpine, McN-A-343, FPL64176 and thapsigargin. As shown in Fig. 4B, a 0.25 mg/kg dose of SB216763 exerted potent and broad antiemetic efficacy against vomiting evoked by the above discussed emetogens except quinpirole and pilocarpine. Indeed, relative to the vehicle-pretreated control group, SB216763 caused significant decreases both in the mean vomit frequency in response to 5-HT (p = 0.0019), GR73632 (p = 0.0018), apomorphine (p = 0.0001), McN-A-343 (p = 0.0102), FPL64176 (p < 0.0001), and thapsigargin (p < 0.0001), as well as percentage of shrews vomiting in response to 5-HT (p = 0.007), GR73632 (p = 0.0063), apomorphine (p = 0.0013), McN-A-343 (p = 0.0013), FPL64176 (p = 0.0013) and thapsigargin (p = 0.0013).

We further tested the antiemetic potential of larger doses of SB216763 against quinpirole- and pilocarpine-induced vomiting. In the case of quinpirole-induced vomiting, SB216763 (0.25, 1, and 2.5 mg/kg, i.p.) significantly reduced both the mean vomit frequency (KW (3, 24) = 12.99, p = 0.0047) and the percentage of shrews vomiting (χ² (3, 24) = 13.15; p = 0.0043) (Fig. 4C). In fact, significant reductions in both vomit frequency (p = 0.0012) and percentage of shrew vomiting (p = 0.0004) occurred at its 2.5 mg/kg dose (Fig. 4C). We then evaluated the dose-response antiemetic potential of SB216763 against pilocarpine, which significantly suppressed both the mean frequency of vomits (KW (3, 22) = 12.86, p = 0.005) and the percentage of shrews vomiting (χ² (3, 22) = 14.39; p = 0.0024) (Fig. 4D). Complete reduction in the frequency of vomiting was observed at its 2.5 mg/kg (p = 0.0014) dose. Furthermore, significant reduction in the percentage of shrews vomiting occurred at its 1 mg/kg (50%, p = 0.0241) and complete protection at its 2.5 mg/kg dose (100%, p = 0.0002).

Unfortunately, both GSK-3 inhibitors failed to significantly affect cisplatin (10 mg/kg., i.p.)-induced acute (0–4 h post administration)- or delayed-phase (27–40 h post administration) vomiting in the least shrews (data not shown).

Discussion

Unlike radiation or cytotoxic cancer chemotherapy, targeted cancer chemotherapeutics specifically affect the lipid-derived intracellular signals in the PI3/Akt pathway to suppress cancer, but they also evoke nausea and vomiting (Aghajanian et al., 2018). Although mechanisms of their cancer suppressive effects are probably fairly well understood, the neurogastrophysiology of emetic signals evoked by targeted chemotherapeutics remain unknown. Moreover, historically, development of antiemetics for the prevention of vomiting caused by cytotoxic cancer chemotherapeutics such as cisplatin has largely been empirical (Darmani, 2014), and only more recently mechanistic-based regimens being employed, which are still only effective in up to 85% of patients receiving high-dose cisplatin-type chemotherapy (Navari and Schwartzberg, 2018).

In light of the current knowledge gap, we are investigating intracellular emetic signals evoked by specific emetogens. Our recent findings have well established that ERK1/2 is a common emetic signal in the mediation of vomiting elicited by intraperitoneal administration of diverse emetogens (Zhong et al., 2019; 2016; 2014b; Darmani et al., 2017; 2015; 2013). Moreover, recently our group demonstrated a time-dependent upregulation of phosphorylation of Akt in the least shrew brainstem following administration of the emetic neurokinin NK₁R agonist GR73632 (Zhong et al., 2019) and the selective LTCC agonist FPL64176 (Zhong et al., 2018). In the present study we extend this observation and demonstrate that a major downstream target of Akt, GSK-3, also plays a role in vomiting since phosphorylation of its subtypes GSK-3α and GSK-3β in the least shrew brainstem displays a time-dependent increase in response to a variety of emetogens, including agonists of serotonergic 5-HT₃ (e.g. 5-HT or 2-Methyl-5-HT)-, substance P neurokinin NK₁ (GR73632)-, dopaminergic D₂ (apomorphine or quinpirole)-, cholinergic M₁ (McN-A343 or pilocarpine)-receptors, as well as the Ca²⁺ channel regulators such as the L-type Ca²⁺ channel agonist FPL64176 (Darmani et al., 2014) and the SERCA inhibitor thapsigargin (Zhong et al., 2016).

GSK-3 is a constitutively active protein and its activity can be inhibited through phosphorylation of Ser21 in GSK-3α and Ser9 in GSK-3β (Mazzardo-Martins et al., 2012). In the current study for the first time we demonstrate increased phosphorylation of both GSK-3 subtypes in the least shrew brainstem in a parallel time-dependent profile following the administration of fully-effective emetic doses of the above discussed diverse emetogens. Moreover, immunostaining confirmed that phosphorylation of both subtypes exhibit increased immunoreactivity in the least shrew brainstem DVC emetic nuclei (AP, NTS and DMNX) in response to cisplatin administration. Interestingly, phospho-GSK-3α/β Ser21/9 immunoreactivity was also increased in the hypoglossal nucleus (XII nucleus) which is located below the DMNX (See supplemental Fig. S1). In fact involvement of the hypoglossal nucleus in vomiting was previously postulated by Umezaki and co-workers (1998). The latter nucleus also exhibits the highest density of muscarinic cholinergic receptors in the human brainstem nuclei (Hyde et al., 1988). More importantly, the tested selective GSK-3 inhibitors, AR-A014418 and SB216763, reduced vomiting in least shrews evoked by the above discussed diverse emetogens. The present data support an interesting hypothesis that following pharmacological induction of vomiting, the evoked phosphorylation (i.e. inhibition) of GSK-3 may exert a unique self-protective effect to avoid further vomiting. Thus, utilization of GSK-3 inhibitors may become a useful strategy for achieving a greater degree of inhibition of GSK-3 activity for suppression of vomiting. Consistent with the report by Beaulieu et al. (2004), that lithium, also a known direct inhibitor of GSK-3, can increase GSK-3α/β phosphorylation at Ser21/9 sites and inhibit its activity, our Western blots performed on the least shrew brainstem lysates showed the GSK-3α/β inhibitor AR-A014418 at 5, 10 and 20 mg/kg dosages (i.p.) can also significantly increase GSK-3α/β phosphorylation at Ser21/9 sites (Data not shown). Moreover, the serotonergic system has been postulated to be associated with the regulation of GSK-3 activity in vivo. In fact, stimulation of serotonin 5-HT_1A receptors can increase phosphorylation GSK-3α and GSK-3β in several areas of the hippocampus (Polter et al., 2012). Additionally, 5-HT_1A receptor agonists possess antiemetic efficacy against vomiting evoked by motion, cisplatin, xylazine, nicotine and copper sulfate (Lucot and Crampton, 1989; Javid and Naylor, 2006; Okada et al., 1994).

In a recent publication we had examined the time-dependency of phosphorylation (activation) of ERK1/2, protein kinase A (PKA) as well as the α and βII isoforms of protein kinase C (PKCα/βII), in the least shrew brainstem at 7 different time-points (1, 2, 8, 16, 24, 33 and 40 h) post-cisplatin administration (Darmani et al., 2013). The results showed that cisplatin significantly increased phosphorylation of ERK1/2 at the 2, 8, and 33 h, as well as phosphorylation of PKA at 33 and 40 h time-points in the least shrew brainstem. Regarding the phosphorylation status of PKCα/βII in the brainstem, a significant increase was observed at the 2 h time-point post-cisplatin injection. In the current study, phospho-GSK-3α/β was upregulated at 5 h post-cisplatin administration which was maintained upregulated up to 33 h post-treatment. Thus, it appears that there is a delay in phospho-GSK3α/β increase compared to ERK1/2 phosphorylation which had significantly increased as early as 2 h post cisplatin administration (Darmani et al2013). This observation appears to support the converging evidence indicating that ERK1/2 may act as an upstream element of the GSK-3 signaling pathway (He et al., 2015; Li et al., 2018; Peng et al., 2018; Qiu et al., 2018).

Effects of GSK-3 inhibitors

Of the large number of available GSK-3 inhibitors, only a few have reached clinical trials for the treatment of different diseases including cancer, Alzheimer’s disease or bipolar disorder (Pandey and DeGrado, 2016; Saraswati et al., 2018). Lithium chloride (LiCl) is an inhibitor of GSK-3 as well as other enzymes up and downstream of the PI3K/Akt/GSK-3 pathway and is used in the clinic for the treatment of mania and bipolar disorder (Nobel et al., 2005). In the laboratory LiCl has been used to induce vomiting (150 – 390 mg/kg, i.p. or intramuscularly) and/or gaping (127 mg/kg, i.p.) in several vomit-competent species as well as in rodents, respectively (Rock et al., 2016; Sticht et al., 2013; Wooldridge and Kangas, 2019). In the clinic oral lithium carbonate is used in the range of 900 – 1500 mg in outpatients and 1200 to 2400 mg per day in hospitalized manic patients in divided doses (Baldessarini and Tarazi, 2006). Vomiting and nausea are reported as dose-dependent side-effects of lithium carbonate in such patients. The discrepancy between LiCl’s emetic effects and antiemetic efficacy of the selective GSK-3 inhibitors AR-A014418 and SB216763 used in our study probably reflects the plethora of effects that large doses of lithium produces on diverse enzymes (Kerr et al., 2018). In fact, other GSK-inhibitors (Tideglusib, LY2090314) undergoing phase 1 clinical trials do not demonstrate dose-dependent emetic effects (Tolosa et al., 2014; Gray et al., 2015).

In the present study, AR-A014418 suppressed vomiting caused via the activation of specific key emetic receptors (5-HT₃, NK₁, D₂, M₁) and Ca²⁺ channel regulators. A maximally effective-dose of each emetogen was used in the current study which induced emesis in each tested shrew (Darmani et al., 2014; Zhong et al., 2016). The overall results demonstrate that AR-A014418 possess broad-spectrum antiemetic efficacy against specific but diverse emetogens. In fact, AR-A014418 significantly reduced to varying degrees and in a dose-dependent fashion, both the frequency and percentage of shrews vomiting caused by the:

1. selective serotonin 5-HT₃R agonist, 2-Methyl-5-HT (5 mg/kg, i.p.). Although the mean frequency of evoked emesis was reduced by more than 75% at its 10 and 20 mg/kg doses, these doses AR-A014418 only managed to fully protect shrews by approximately 50%. Likewise, a 2.5 mg/kg dose of the first generation selective serotonin 5-HT₃R antagonist tropisetron, was shown to provide a similar degree of protection against 2-methyl-5-HT (5 mg/kg)-evoked vomiting in the least shrew, but its larger doses failed to provide additional protection (Darmani et al., 2011). On the other hand, the second generation potent and selective 5-HT₃R antagonist palonosetron, reduced the mean vomit frequency of 2-methyl-5-HT-evoked vomiting in least shrews by over 60% at 0.1 mg/kg dose, and only its 10 mg/kg dose could completely prevent the evoked vomiting (Darmani et al., 2014). In the current study AR-A014418 also failed to fully protect shrews from vomiting caused by the nonselective 5-HT₃R agonist serotonin (5-HT; 5 mg/kg, i.p.), but it did partially attenuated (64%, P<0.05) the mean vomit frequency at its 20 mg/kg dose. Thus, AR-A014418 appears not to be as efficacious as 5-HT₃R antagonists in suppressing vomiting evoked by activation of serotonin 5-HT₃Rs.

2. selective substance P neurokinin NK₁R agonist GR73632 (5 mg/kg, i.p.). In fact a 10 mg/kg dose of AR-A014418 reduced the mean vomit frequency by 90% and protected 70% of shrews from vomiting. Such degree of protection from the same dose of GR73632 in the least shrew could not be achieved following pretreatment with 5 – 20 mg/kg doses of selective NK₁R antagonists CP99,994 or L733060 under similar experimental conditions (Darmani et al., 2008). Only the very potent NK₁R antagonist netupitant could completely protect least shrews from GR73632 (5 mg/kg)-evoked vomiting at 10 mg/kg, and its 5 mg/kg dose afforded (Zhong et al., 2019) a comparable degree of protection to 10 mg/kg dose of AR-A014418 observed in the current study.

3. non-selective D₂R agonist apomorphine (2 mg/kg, i.p.). Indeed, increasing doses of AR-A014418 (5, 10 and 20 mg/kg) fully protected shrews from apomorphine-evoked vomiting by 50, 67 and 100%. Under identical experimental conditions, a 2 mg/kg (i.p.) dose of the dopamine D₂R selective and potent antagonist sulpiride, completely prevented apomorphine-induced vomiting in least shrews (Darmani et al., 1999). However, sulpiride (up to 8 mg/kg, i.p.) could not fully protect shrews from vomiting evoked by the more selective and potent D₂R agonist, quinpirole (2 mg/kg, i.p.) (Darmani et al., 1999). Likewise, in the current study AR-A014418 was less efficacious against quinpirole-evoked emesis and significantly reduced the mean vomit frequency by 84 and 92% at its 10 and 20 mg/kg doses. Overall, our results suggest that AR-A014418 can suppress vomiting caused by dopaminergic D2R agonists.

4. more selective cholinergic M₁R agonist McN-A-343, where significant vomit protection was achieved at 5 (67%) and 10 (83%) mg/kg doses of AR-A014418. Moreover, these doses of AR-A014418 respectively reduced the mean vomit frequency by 82% and 91%. Likewise, AR-A014418 exerted similar protective effect against the non-selective cholinergic agonist pilocarpine, albeit at relatively larger doses (10 and 20 mg/kg). Moreover, another class of broad-spectrum antiemetic, the LTCC blockers amlodipine or nifedipine, provide similar emetic protection against McN-A-343- and pilocarpine-evoked emesis (Zhong et al., 2014a).

5. totally abrogated vomiting caused by the LTCC agonist FPL64176 in a dose-dependent manner with total protection at its 10 mg/kg dose. Moreover, AR-A014418 significantly suppressed vomiting caused by the SERCA inhibitor, thapsigargin, albeit a larger dose (20 mg/kg). Our findings imply that GSK-3 inhibitors may limit cytosolic Ca²⁺ mobilization. In fact, GSK-3 inhibition by SB216763 blunts liposaccharide-induced rise in cytosolic Ca²⁺ concentration in dendritic cells (Russo et al., 2013; Schmid et al., 2014), Furthermore, AR-A014418 causes antinociceptive effects via modulation of glutamatergic system through its ionotropic (NMDA) receptors which allows extracellular Ca²⁺ influx (Martins et al., 2011). Moreover, it is recognized that activation of emetic serotonergic 5-HT₃Rs on vagal afferent terminals in the brainstem increases glutamatergic transmission related to the emesis (Browning, 2015).

The above discussed broad-spectrum antiemetic efficacy of AR-A014418 in the least shrew is further supported by our findings that lower doses of another GSK-3 inhibitor SB216763, provides more potent anti-emetic effects. In fact, we tested the antiemetic potential of an 0.25 mg/kg (i.p.) dose of SB216763 against the above discussed emetogens. SB216763 potently suppressed both the frequency and the number of shrews vomiting in response to maximally-effective doses of most of above discussed emetogens except quinpirole and pilocarpine. We then tested the antiemetic potential of larger doses of SB216763 (0, 0.25, 1 and 2.5 mg/kg) against the latter two emetogens. SB216763 protected shrews from the evoked vomiting in a dose-dependent fashion with complete or near complete protection at 2.5 mg/kg. Thus, SB216763 appears to a very potent and broad-spectrum antiemetic having greater antiemetic efficacy against diverse target-specific emetogens when compared to their corresponding potent and selective antagonists/inhibitors. Differences in pharmacological properties of the two tested GSK-3 inhibitors may contribute to their differential antiemetic potential. Indeed, AR-A014418 is a selective ATP-competitive GSK-3β inhibitor (IC₅₀ = 104 nM) (Bhat et al., 2003; Gould et al., 2004; Mazzardo-Martins et al., 2012; Tunçdemir et al., 2013), whereas SB 216763 is considered as a more potent but equally effective inhibitor of both GSK-3α and GSK-3β (IC₅₀ = 34.3 nM) (Coghlan et al., 2000; Wang et al., 2011). Our findings suggest that it is possible to suppress diverse causes of vomiting by targeting a common intracellular signal downstream of their corresponding receptors/effectors.

SB216763 can inhibit cisplatin-induced apoptosis in auditory cells of mice in a dose-dependent manner (Park et al., 2009). However, in the present study both SB216763 and AR-A014418, failed to protect shrews from cisplatin-induced vomiting (data not shown). This may not be surprising since our Western blot findings indicate that cisplatin does not increase phosphorylation levels of GSK-3α/β in the least shrew brainstem within 4 hours of its administration. This probably reflects lack of antiemetic efficacy of both of these GSK-3 inhibitors against the acute-phase of cisplatin-evoked vomiting which peaks 1–2 hours post-cisplatin administration in the least shrew (Darmani et al., 2009). The major emetic neurotransmitter that contributes towards the acute-phase of cisplatin-evoked emesis is thought to be serotonin (Hesketh et al., 2003). This is in line with our current findings that both GSK-3 inhibitors were only partially effective in protecting shrews from serotonin (or 2-methyl-5-HT)-evoked vomiting. However, cisplatin did increase phospho-GSK-3α/β immunoreactivity during the delayed phase of vomiting which peaks at 33 hour post-cisplatin injection (Darmani et al., 2009). Based upon the latter finding as well as the potency and broad-spectrum efficacy of AR-A014418 and SB213763 against the discussed specific direct-acting emetogens, it is surprising that these GSK-inhibitors failed to suppress the delayed-phase of cisplatin-evoked vomiting. Perhaps combination of a GSK-3 inhibitor with either a 5-HT₃R (e.g. palonosetron) or NK₁R (e.g. netupitant) antagonist, may provide additive antiemetic efficacy as it has been demonstrated with amlodipine or nifedipine plus palonosetron (Zhong et al., 2014a; Darmani et al., 2014), or pranlukast plus palonosetron (Darmani et al., 2017).

Peripheral GSK-3

GSK-3 is probably one of the busiest kinase with over 100 substrates and has a number of cytoplasmic, mitochondrial and nuclear targets (Beurel et al., 2015). Although little is known about mechanisms that differentially regulate the two GSK-3 isoforms, some reports do indicate differential action, regulation and expression (Beurel et al., 2015). In addition to observed changes in expression levels of the two GSK-3 paralogs in the shrew brainstem, we also applied immunostaining to investigate the possible impact of cisplatin-induced vomiting on the phosphorylation states of GSK-3 isoforms in the least shrew jejunum.

Our other findings indicate that besides parallel changes in GSK-3α and GSK-3β phosphorylations observed in the least shrew brainstem following administration of diverse emetogens including cisplatin, an increase in both phospho-GSK-3α Ser21 and phospho-GSK-3β Ser9 was also observed in the jejunal enteric nervous system (ENS) of least shrew at 5 h post cisplatin administration (10 mg/kg., i.p.) (Supplemental Figs. S1–2). Indeed, we confirmed phospho-GSK-3α Ser21 expression in the neurons of the jejunal ENS via co-immunostaining with antibodies of phospho-GSK-3α Ser21 and the neuron marker (NeuN) (Supplemental Fig. S3). Both phosphorylation of GSK-3 subtypes have been confirmed in protein lysates of mice small intestine samples (Hey et al., 2016). However, more often, expression of GSK-3β and its phosphorylated form (Ser9) have been determined with Western blots in cultured enteric neurons prepared from the small intestines of rat (Mwangi et al., 2006); cultured human intestinal muscle cells (Kuemmerle, 2005) and as well as in the murine jejunum (Leung et al., 2015; Salinari et al., 2013).

In addition, we co-stained the least shrew jejunal sections with phospho-GSK-3α Ser21 with a 5-HT antibody (generally-accepted as a marker of the intestinal enterochromaffin (EC) cells). The attained results indicate the significant presence of phospho-GSK-3α Ser21 in EC cells (Supplemental Fig. S4), while phospho-GSK-3β Ser9 was barely detected in these phospho-GSK-3α Ser21 positive EC cells (Supplemental Fig. S5). In this regard, it would be extremely important to design future experiments with more highly selective and potent GSK-3α and GSK-3β paralogs so that their corresponding antiemetic potential and matching signaling pathways could be further assessed both in the brainstem and the intestine.

Conclusions

Taken together, for the first time we provide evidence for the involvement of GSK-3 in vomiting and address the feasibility of targeting GSK-3 inhibition as antiemetic therapeutically. We demonstrate time-dependent increases in GSK-3α/β phosphorylation at Ser21/9 in the brainstem following peripheral administration of diverse emetic receptor selective/nonselective emetogens in least shrews. Phospho-GSK3α/β at Ser21/9 expresses throughout the brainstem emetic nuclei, including the AP, NTS and DMNX. Furthermore, a less potent GSK-3β inhibitor, AR-A014418, exerted significant antiemetic potential against vomiting caused by diverse emetogens. This is consistent with the antiemetic efficacy of a low dose of the more potent GSK-3α/β inhibitor, SB216763. These findings supports our hypothesis that GSK-3 phosphorylation in vivo may protect against excessive vomiting evoked by diverse emetogens (except cisplatin) through inhibition of GSK-3. Future studies are required to delineate: i) precisely the mechanism by which GSK-3 inhibition exerts antiemetic effects, ii) the antiemetic role of phospho-GSK-3α Ser21 in jejunal EC cells in the periphery, iii) the downstream targets of GSK-3 involved in vomiting, and iv) why the tested GSK-3 inhibitors failed to suppress cisplatin-evoked emesis.

Highlights.

• Diverse selective and nonselective emetogens increase phospho-GSK-3α/β Ser21/9 levels in the least shrew brainstem.

• The cancer chemotherapeutic agent cisplatin also increases phospho-GSK-3α/β Ser21/9 immunoreactivity in both the brainstem emetic nuclei and the enteric nerves of jejunum in the small intestine.

• The GSK-3 inhibitor AR-A014418, dose-dependently suppressed both the frequency and percentage of shrews vomiting evoked by specific emetogens.

• Another GSK-3 inhibitor SB216763, is a more potent antiemetic than AR-A014418.

Abbreviations

GSK-3

glycogen synthase kinase 3

DVC

dorsal vagal complex

AP

area postrema

NTS

nucleus tractus solitarius

DMNX

dorsal motor nucleus of the vagus

ENS

enteric nerves system

GIT

gastrointestinal tract

EC cells

enterochromaffin cells

5-HT

serotonin

5-HT₃R

serotonin type 3 receptor

NK₁R

neurokinin type 1 receptor

D₂R

dopamine D₂ receptor

M₁R

muscarinic 1 receptor

LTCC

L-type Ca²⁺ channel

SERCA

sarco/endoplasmic reticulum Ca²⁺-ATPase

i.p.

intraperitoneal

ERK1/2

extracellular signal-regulated protein kinase1/2

Akt

protein kinase B

h

hour

p-GSK-3α/β

phospho-GSK-3α/β

PBS

phosphate-buffered saline

PKA

protein kinase A

PKCα/βII

α and βII isoforms of protein kinase C

LiCl

lithium chloride