Abstract
Leaky gut that is acondition reflecting intestinal barrier dysfunction has been attracting attention for its relations with many diseases such as irritable bowel syndrome or Alzheimer dementia. We have recently demonstrated that ghrelin acts in the brain to improve leaky gut via the vagus nerve. In the present study, we tried to clarify the precise central mechanisms by which ghrelin improves intestinal barrier function through the vagus nerve. Colonic permeability was estimated in vivo by quantifying the absorbed Evans blue in colonic tissue in rats. Adenosine receptor antagonist, 1,3-dipropyl-8-cyclopentylxanthine (DPCPX), blocked the intracisternal ghrelininduced improvement of intestinal hyperpermeability while dopamine, cannabinoid or opioid receptor antagonist failed to prevent it. Since DPCPX can block adenosine A1 and adenosine A2B receptors, we examined which subtype is involved in the mechanism. Intracisternal injection of adenosine A2B agonist but not adenosine A1 agonist improved colonic hyperpermeability, while peripheral injection of adenosine A2B agonist failed to improve it. Intracisternal adenosine A2B agonist-induced improvement of colonic hyperpermeability was blocked by vagotomy. Adenosine A2B specific antagonist, alloxazine blocked the ghrelin or central vagal stimulation by 2-deoxy-D-glucose-induced improvement of intestinal hyperpermeability. These results suggest that activation of adenosine A2B receptors in the central nervous system is capable of improving intestinal barrier function through the vagal pathway, and the adenosine A2B receptors may mediate the ghrelin-induced improvement of leaky gut in a vagal dependent fashion. These findings may help us understand the pathophysiology in not only gastrointestinal diseases but also non-gastrointestinal diseases associated with the altered intestinal permeability.
1. Introduction
A disturbed intestinal barrier function, leaky gut, has been described in many human diseases such as irritable bowel syndrome (IBS), inflammatory bowel disease and Alzheimer，s disease (Camilleri, 2019). Improvement of disturbed intestinal barrier function may contribute to control of activity of several gastrointestinal and non-gastrointestinal diseases. Although accumulating evidence has suggested that intestinal barrier function is regulated by a number of peripheral mechanisms such as tight junction proteins in the epithelium, neuroimmune-related molecules and microbiota (Camilleri, 2019; Keita and Soderholm, 2018), the mechanisms how the brain controls intestinal barrier function still remain to be fully elucidated. We have very recently demonstrated that exogenously administered or endogenously released orexin or ghrelin in the brain blocked colonic hyperpermeability in rats, suggesting for the first time that the brain indeed plays a role in regulation of intestinal barrier function (Ishioh et al., 2020; Okumura et al., 2020). The neuronal rapid protective advantage to the host by improving the intestinal barrier function by the neuropeptides may help us understand the brain-gut interaction in stress sensitive gastrointestinal disorders like IBS associated with the altered intestinal permeability.
Ghrelin, a 28-amino acid peptide acts as an endogenous ligand for the growth hormone secretagogue receptor (GHSR), which is expressed throughout the brain and mainly in the pituitary and hypothalamus (Kojima et al., 1999). The main source of ghrelin is the stomach (Kojima and Kangawa, 2010). Although ghrelin receptors are expressed in the brain, the source of ghrelin in the brain remains unclear (Cabral et al., 2017b). We have very recently demonstrated that ghrelin acts centrally in the brain to block leaky gut evoked by lipopolysaccharide (LPS) through the vagal cholinergic pathway in rats (Ishioh et al., 2020). Since it has been established that activation of vagal efferent nerves is capable of inducing the vagal cholinergic anti-inflammatory action (Borovikova et al., 2000), the vagal–dependent anti-inflammatory action is considered to be a vital mechanism by which brain ghrelin improved leaky gut. In the present study, we tried to further clarify central mechanisms of brain ghrelin-induced improvement of intestinal barrier function. Since accumulating evidence has suggested that dopamine, cannabinoid, adenosine or opioid in the central nervous system (CNS) signaling plays an important role in the control of gastrointestinal function such as visceral sensation (Okumura et al., 2015, 2016, 2018a, 2018b), involvement of these chemicals in the ghrelin-induced improvement of intestinal barrier function was examined in the present study.
2. Materials and methods
2.1. Ethical considerations
Approval was obtained from the Research and Development and Animal Care committees at Asahikawa Medical University (No. 13030) for all of the experiments conducted in this study.
2.2. Animals
Male Sprague-Dawley rats (Charles River Laboratory,Atsugi, Japan) weighing about 200 g were housed under controlled light/dark conditions (lights on: 07:00– 19:00),and the room temperature was regulated at 23–25 。C. Rats were allowed free access to standard rat chow (solid rat chow; Oriental Yeast Co., Tokyo, Japan) and tap water. All experiments were performed using conscious animals, which had been deprived of food for 24 h but with free access to water until the initiation of the experiments.
2.3. Chemicals
Synthetic human ghrelin and orexin-A were purchased from Peptide Institute Inc., Osaka, Japan and dissolved in normal saline. Lipopolysaccharide (LPS) obtained from Escherichia coli with theserotype O55:B5 or 2-deoxy-D-glucose (2-DG) (Sigma-Aldrich, St. Louis, MO) was dissolved in normal saline. SCH-23390, D1 dopamine receptor antagonist, sulpride, D2 dopamine receptor antagonist, AM251, CB1 receptor antagonist (Wako Chemical, Osaka, Japan), 1,3dipropyl-8-cyclopentylxanthine (DPCPX) (Tocris Bioscience, Bristol, UK), an adenosine A1 and A2B receptor antagonist, or Alloxazine (Sigma-Aldrich),the specific adenosine A2B receptor antagonist were dissolved in 100% dimethyl sulfoxide (DMSO, Sigma-Aldrich). N(6)-cyclopentyladenosine (CPA) (Abcam,Tokyo, Japan), the specific adenosine A1 receptor agonist or BAY-606503 (R&D systems, Inc., Minneapolis, MN), the specific adenosine A2B receptor agonist were dissolved in DMSO.
2.4. Measuring colonic permeability
Colonic permeability measurement was performed according to previous studies (Dai et al., 2012; Kitajima et al., 1999; Lange et al., 1994; Ukena et al., 2007) with minor modification. The permeability was determined 3 hor 4 h after the administration of LPS, respectively as shown in our recent reports (Ishioh et al., 2020; Nozu et al., 2018a,2018b, 2018c; Okumura et al., 2020). The rats anesthetized by intraperitoneal administration of the mixture of medetomidine hydrochloride (Orion Pharma Ltd., Dhaka, Bangladesh, 0.15 mg/kg), midazolam (Sandoz, Tokyo, Japan, 2 mg/kg) and butorphanol tartrate (Meiji Seika Pharma, Tokyo, Japan, 2.5 mg/kg) were placed in a supine position on a heating pad, and laparotomy was performed. The colon was ligated at the junction with the cecum, and the small hole was made by a puncture using 18 G needle at the 1 cm from the ileocecal junction. Then an opentipped catheter (3-Fr, Atom, Tokyo, Japan) was inserted into the proximal colon through the hole and fixed by purse-string sutures. The colon was gently flushed with phosphate buffered saline (PBS, 37 。C) using the catheter until all stools were washed out. Generally, the required volume of PBS was approximately 10 ml and the perfusion rate was 5 ml/min. Then another ligation was added on the colon at approximately 4 cm from the proximal ligation, and 1 ml of 1.5% Evans blue in PBS was instilled into the colon segment between ligations through the catheter. Fifteen min later, the animals were euthanized and the colons were excised. Later they were washed with PBS and 1 ml of 6 mM N-acetylcysteine, and were opened and placed in 2 ml of N, N-dimethylformamide for 12 h. The permeability was calculated by measuring the Evans blue concentration in the supernatant using a spectrophotometer at 610 nm.
2.5. Experimental procedures
Initially, to clarify whether adenosine, cannabinoid, dopamine or opioid signaling is involved in ghrelin-induced protective action against colonic hyperpermeability, we examined the effects of subcutaneous injection of DPCPX (1 mg/kg), an adenosine A1 antagonist; AM251 (0.5 mg/rat), cannabinoid 1 receptor antagonist; SCH23390 (0.2 mg/rat), dopamine D1 antagonist; sulpride (40 mg/rat),dopamine D2 antagonist or naloxone (0.5 mg/rat), opioid antagonist on intracisternal ghrelin (2 μg)-induced improvement of colonic hyperpermeability induced by subcutaneously administered LPS (1 mg/kg). To make sure that the effect of adenosine signaling on LPS-induced colonic hyperpermeability, we examined the effects of peripheral or central injection of CPA (4 μgic or 0.2 mgsc),adenosine A1 receptor agonist, or BAY 60–6583 (0, 0.2, 2, 20 or 200 μgic or 200 μg ip),adenosine A2B receptor agonist. To clarify whether vagal cholinergic pathway is involved in adenosine A2Binduced protective action against colonic hyperpermeability, we examined the effect of bilateral subdiaphragmatic vagotomy on the intracisternally administered BAY 60–6583 (2 μg/10 μl)-induced change of colonic hyperpermeability by LPS. The surgical bilateral vagotomy was performed as previously described (Takahashi et al., 1999).Next, to clarify whether adenosine A2B signaling is involved in ghrelin or orexin-induced protective action against colonic hyperpermeability, we examined the effects of intraperitoneal injection of alloxazine (1 mg/kg), an adenosine A2B specific antagonist on intracisternal administered ghrelin (2 μg)-orexin-A (10 μg)-induced improvement of colonic hyperpermeability induced by LPS (1 mg/kg, sc). To examine whether endogenous adenosine A2B in the brain mediates the central vagal activation by 2-DG-induced hyperpermeability, we tested the effect of intracisternal alloxazine, adenosine A2B specific antagonist, on the 2-DG-induced improvement of colonic hyperpermeability by LPS.Intracisternal injection was performed under brief isoflurane anesthesia using a 10-μl Hamilton microsyringe after the rats were mounted in a stereotaxic apparatus (David Kopf Instruments, Tijunga, CA), as reported previously (Okumura et al., 1994).The doses of ghrelin, orexin-A, LPS, DPCPX, AM251, SCH23390, sulpride, naloxone, CPA, alloxazine, BAY-60-6583 and 2-DG were selected according to previous publications (Ishioh et al., 2020; Li et al., 2017; Minor and Hanff, 2015; Okumura et al., 2016, 2018b).
2.6. Statistical analysis
The data were expressed as means ± standard error (SE). The data were compared with Student’s t-test or one-way analysis of variance followed by Tukey’s Multiple Comparison Test. p < 0.05 was considered statistically significant.
3. Results
3.1. Effect of adenosine, cannabinoid, dopamine or opioid signaling on intracisternal ghrelin-induced improvement of colonic hyperpermeability by LPS
We examined the effects of subcutaneous injection of DPCPX, an adenosine receptor antagonist; AM251, cannabinoid 1 receptor antagonist; SCH23390, dopamine D1 antagonist; sulpride, dopamine D2 antagonist and naloxone on intracisternal ghrelin-induced improvement of colonic hyperpermeability induced by LPS. As shown in Fig. 1, subcutaneously administered LPS at a dose of 1 mg/kg significantly increased the colonic permeability and intracisternal injection of ghrelin at a dose of 2 μg significantly improved colonic hyperpermeability as reported previously (Ishioh et al., 2020). DPCPX significantly blocked the ghrelin-induced protective action against colonic hyperpermeability by LPS, suggesting that adenosine mediates ghrelin-induced improvement of colonic hyperpermeability (Fig. 1). On the other hand, AM251, SCH23390, sulpride or naloxone failed to alter colonic permeability induced by LPS (Fig. 2), suggesting dopamine, cannabinoid or opioid is not involved in the ghrelin-induced improvement of colonic hyperpermeability.
3.2. Effect of centrally administered adenosine A1 or A2B receptor agonist on colonic hyperpermeability
Since it has been demonstrated that DPCPX is capable of blocking not only adenosine A1 receptor but also adenosine A2B receptor (Li et al., 2007), we examined the effect of centrally administered adenosine A1 or A2B receptor agonist on colonic hyperpermeability. Intracisternal or subcutaneous injection of CPA, adenosine A1 agonist, failed to alter LPSstimulated colonic hyperpermeability (Fig. 3A, B). On the other hand, intracisternal injection of BAY 60–6583, adenosine A2B agonist, dosedependently inhibited the colonic hyperpermeability (Fig. 4A). The significant effects were observed when BAY 60–6583 at 0.2 μg or more was administered. In contrast, intraperitoneal injection of BAY 60–6583 at 200 μg dose failed to block the stimulated colonic permeability (Fig. 4B), suggesting that adenosine acts on adenosine A2B but not adenosine A1 receptor in the brain to improve colonic barrier function.
3.3. Effect of vagotomy on adenosine A2B receptor agonist-induced improvement of colonic hyperpermeability
Since vagal cholinergic pathway plays a role in improvement of altered colonic permeability by LPS (Ishioh et al., 2020), we examined the effect of vagotomy on adenosine A2B receptor agonist-induced improvement of colonic hyperpermeability. As shown in Fig. 5, surgical bilateral vagotomy significantly blocked the adenosine A2B agonistinduced improvement of colonic hyperpermeability in response to LPS, suggesting that the vagal pathway is involved in the improvement of colonic hyperpermeability by activation of central adenosine A2B receptor.
3.4. Effect of adenosine A2B receptor antagonist on 2-DG-induced improvement of colonic hyperpermeability by LPS
As we recently demonstrated, intravenous 2-DG, a central vagal stimulant (Kadekaro et al., 1975), significantly blocked the stimulated colonic permeability by LPS, and atropine or vagotomy blocked the effect of 2-DG, supporting that the vagal cholinergic stimulation mediates the improvement of colonic hyperpermeability by 2-DG (Okumura et al., 2020). We therefore made a hypothesis that endogenous brain adenosine A2B receptor may be involved in the 2-DG-induced blockade of stimulation of colonic permeability. Alloxazine significantly reversed the 2-DG-induced blockade of stimulation of colonic permeability (Fig. 6), suggesting that endogenous brain adenosine A2B receptor may play a role in the improvement of intestinal barrier function.
3.5. Effects of adenosine A2B receptor antagonist on ghrelin-induced improvement of colonic hyperpermeability by LPS
To confirm the implication of central adenosine A2B receptor in the brain ghrelin-induced improvement of intestinal hyperpermeability, we examined the effects of intraperitoneal injection of alloxazine, adenosine A2B specific antagonist, on intracisternal ghrelin-induced improvement of colonic hyperpermeability induced by LPS. Alloxazine significantly blocked the ghrelin-induced protective action against colonic hyperpermeability by LPS (Fig. 7A) while alloxazine failed to block the central orexin-induced improvement of colonic hyperpermeability (Fig. 7B), suggesting that adenosine A2B receptors specifically mediate the ghrelin-induced improvement of colonic hyperpermeability.
Fig. 1. Effect of DPCPX on ghrelin induced improvement of colonic hyperpermeability by LPS in conscious rat. Each column represents the mean ± S.E. The number of rats is shown in the parentheses. * p < 0.01, when compared with LPS (− ) + saline + DMSO. ** p < 0.01, when compared with LPS (+) + saline + DMSO. *** p < 0.01, when compared with LPS (+) + ghrelin + DMSO. ic, intracisternal; sc, subcutaneous. Fig. 2. Effect of SCH23390 (A),sulpride (B), AM251 (C) or Naloxone (D) on ghrelin induced improvement of colonic hyperpermeability by LPS in conscious rat. Each column represents the mean ± S.E. The number of rats is shown in the parentheses. * p < 0.01, when compared with LPS (− ) + saline + DMSO or saline. ** p < 0.01, when compared with LPS (+) + saline + DMSO or saline. ic, intracisternal; sc, subcutaneous. Fig. 3. p < 0.01 Effect of subcutaneous (A) or intracisternal (B) injection of CPA on the increased colonic permeability by LPS in conscious rat. Each column represents the mean ± S.E. The number of rats is shown in the parentheses. * p < 0.01, when compared with DMSO + LPS (− ). ** p < 0.01, when compared with DMSO + LPS (+). ic, intracisternal; sc, subcutaneous. 4. Discussion Adenosine is an extracellular purine nucleotide signaling molecule responsible for diverse actions in the nervous and gastrointestinal systems (Camici et al., 2018). Adenosine A2B receptors are expressed in not only peripheral tissues but also the CNS (Dixon et al., 1996). Although accumulating evidence has indicated that adenosine A2B receptors in the peripheral organs play a role in regulation of gut secretion, motility or visceral sensation (Asano and Takenaga, 2017), little is known about a role in adenosine A2B receptors in the CNS in gastrointestinal functions. The present study clearly demonstrated that adenosine A2B receptor in the CNS is involved in the regulation of intestinal barrier function, because centrally but not peripherally administered adenosine A2B receptor agonist potently blocked colonic hyperpermeability by LPS. It was also shown that adenosine A2B receptor antagonist blocked the brain ghrelin or central vagal stimulation by 2-DG-induced improvement of intestinal barrier function. Thus, the present study provided a novel evidence that activation of adenosine A2B receptors in the CNS improved colonic hyperpermeability in rats. In addition, vagotomy could block the centrally administered adenosine A2B agonist-induced improvement of intestinal barrier function, suggesting that the vagal pathway mediates the central adenosine A2B receptor activation-induced improvement of colonic barrier function. Since the vagal pathway is capable of improving leaky gut through vagal cholinergic anti-inflammatory pathway (Borovikova et al., 2000; Ishioh et al., 2020; Okumura et al., 2020), we would suggest that activation of adenosine A2B receptor in the CNS is capable of activating the vagal efferent anti-inflammatory pathway, thereby reducing intestinal hyperpermeability. We have very recently demonstrated that ghrelin acts centrally in the brain to improve gut barrier function through the vagal pathway in rats (Ishioh et al., 2020). In the present study, we tried to clarify the precise brain mechanism by which ghrelin reduces colonic hyperpermeability. Accumulating evidence has indicated that a functional relationship between visceral sensation and intestinal barrier function. For instance, Creekmore et al. have shown that the level of stress-associated visceral hyperalgesia directly correlates with the magnitude of altered colon epithelial permeability, suggesting a strong correlation between visceral sensation and intestinal permeability (Creekmore et al., 2018). Since recent findings have suggested that dopamine, adenosine, cannabinoid or opioid in the CNS plays a role in the regulation of visceral sensation (Okumura et al., 2015, 2016, 2018a, 2018b), we have examined in this study the effects of several antagonists against dopamine, adenosine, cannabinoid or opioid on the ghrelin-induced improvement of colonic hyperpermeability. As shown in the present study, adenosine receptor antagonist, DPCPX, blocked the ghrelin-induced improvement of intestinal hyperpermeability while dopamine, cannabinoid or opioid receptor antagonist failed to prevent it, clearly suggesting that adenosine in the CNS could specifically play a role in the ghrelin-induce improvement of colonic hyperpermeability by LPS. Alloxazine significantly blocked the ghrelin-induced improvement of colonic hyperpermeability while alloxazine failed to block the central orexin-induced improvement of colonic hyperpermeability, suggesting that adenosine A2B receptors are specifically involved in the ghrelininduced action. These results suggest that adenosine A2B receptor mediates brain ghrelin-induced improvement of intestinal barrier function. We have previously showed that ghrelin acts centrally in the brain to improve intestinal barrier function through the vagal efferent pathway (Ishioh et al., 2020). Based on the present finding that activation of adenosine A2B receptor mediates the ghrelin-induced improvement of colonic hyperpermeability through the vagal pathway, we would speculate that Serine Protease inhibitor ghrelin activates adenosine A2B signaling in the CNS, followed by stimulation of the vagal pathway, thereby inducing the improvement of gut barrier function.
Dixon et al. have demonstrated that adenosine A2B receptor gene expression was detected in every brain area examined in rats (Dixon et al., 1996). Although the present study did not provide evidence regarding to site of action of ghrelinto activate adenosine A2Breceptors, we might be allowed to speculate the following neuronal mechanisms to induce the brain ghrelin-adenosine-induced improvement of intestinal barrier function through the vagus nerve. Previous studies demonstrated that adenosine receptors are located on vagal afferent terminal in the nucleus tract solitaries (NTS) and also found on non-vagal terminals in the NTS, such as terminals of projection form hypothalamic neurons (Krstew et al., 1998; St Lambert et al., 1996). Thus, neurons in the NTS could be influenced by adenosine signaling. It is of interest that adenosine uptake highly occurs in the NTS (Bisserbe et al., 1985), furthermore suggesting that the NTS would be considered to be a possible site of action of adenosine. Peripheral ghrelin signaling travels to the NTS at least via the vagus nerves (Date et al., 2006). Accumulating evidence indicate that the dorsal vagal complex (DVC) including the NTS is a key target of ghrelin action and its neuronal basis as well as the physiological outputs have been already clarified (Bron et al., 2013; Cabral et al., 2017a; Cornejo et al., 2018; Cui et al., 2011; Faulconbridge et al., 2003; Faulconbridge et al., 2008; Li et al., 2006; Lin et al., 2004; Swartz et al., 2014). In fact, Faulconbridge demonstrated that injection of ghrelin into the forth ventricle strongly induced c-fos expression in the NTS in rats (Faulconbridge et al., 2008). Since we injected ghrelin into the fourth ventricle, NTS neurons should be activated in the present study. Neurons in the dorsal motor nucleus, cells of origin innervating the vagal efferent pathway, are closely located and deeply associated with neurons in the NTS (Travagli and Anselmi, 2016). The neuronal activity of the vagal efferent pathway is therefore changed by the activation of adenosine and ghrelin signaling in the NTS. Based on these findings, we would suggest that adenosine signaling in the DVC mediates the ghrelin-induced improvement of intestinal barrier function.
Fig. 4. Effect of intracisternal (A) or intraperitoneal (B) injection of BAY 60-6583 on the increased colonic permeability by LPS in Homogeneous mediator conscious rat. Each column represents the mean ± S.E. The number of rats is shown in the parentheses. * p < 0.01, when compared with DMSO + LPS (− ). ** p < 0.01, when compared with DMSO + LPS (+). ic, intracisternal; ip, intraperitoneal; sc, subcutaneous. Fig. 5. Effects of vagotomy on the BAY60-6583-induced blockade of increased colonic permeability by LPS. Each column represents the mean ± S.E.M. The number of rats examined is shown in parentheses. * p < 0.01, when compared with LPS (− ) + DMSO + sham. ** p < 0.01, when compared with LPS (+) + DMSO + sham. *** p < 0.01, when compared with LPS (+) + BAY60-6583 + sham. Sham, sham operation. BAY 60-6583, specific adenosine A2B receptor agonist. In addition to the involvement of the NTS, we might be allowed to speculate another possibility as following. Neurons in the hippocampus express ghrelin receptors and adenosine A2B receptors (Zigman et al., 2006), and project to the dorsal DVC in the medulla oblongata (Castle et al., 2005). Cui et al. have demonstrated that electroacupuncture modulates the activity of the hippocampus-NTS-vagus nerve pathway to reduce myocardial ischemic injury (Cui et al., 2018). Thus, hippocampus neurons are capable of modulating vagal tone to prevent organ failure, indicating a functional neuronal pathway from the hippocampus to the vagal efferent pathway exists. Considering the evidence, we would suggest that ghrelin or adenosine acts ghrelin receptors or adenosine A2B receptors in the hippocampus neurons, respectively to modulate the activity of hippocampus neurons, followed by activation of vagal tone, thereby improving intestinal barrier function. Thus, the hippocampusvagal pathway might provide a possible neuronal mechanism by which adenosine A2B receptors in the CNS mediate the central ghrelin induced improvement of leaky gut in a vagal-dependent manner.In conclusion, the present study demonstrated that ghrelin acts centrally to improve leaky gut through modulating adenosine A2B receptors, followed by activation of vagal efferent pathway. It is also suggested that adenosine A2B receptor in the CNS should MSC necrobiology be considered to be a possible target for control for diseases associated with leaky gut.
Fig. 6. Effects of intraperitoneal alloxazine on the 2-deoxy-D-glucose (2-DG) -induced blockade of increased colonic permeability by LPS. Each column represents the mean ± S.E.M. The number of rats examined is shown in parentheses. * p < 0.01, when compared with LPS (− ) + saline + DMSO. ** p < 0.01, when compared with LPS (+) + saline + DMSO. *** p < 0.01, when compared with LPS (+) + 2-DG + DMSO. ip, intraperitoneal; sc, subcutaneous. Fig. 7. Effect of alloxazine on ghrelin (A) or orexin (B)-induced improvement of colonic hyperpermeability by LPS in conscious rat. Each column represents the mean ± S.E. The number of rats is shown in the parentheses. * p < 0.01, when compared with LPS (− ) + saline + DMSO. ** p < 0.01, when compared with LPS (+) + saline + DMSO. *** p < 0.01, when compared with LPS (+) + ghrelin + DMSO. ic, intracisternal; ip, intraperitoneal; sc, subcutaneous.