Total of 64 Arc-dVenus mice were used. All Arc-dVenus mice were 7–10 weeks of age at the time of the experiments. The mice were housed at 4–5 animals per cage under controlled temperature (25 ± 1°C) and lighting (12-h/12-h light/dark cycle), and provided with food and water ad libitum. All experimental procedures were conducted in strict accordance with the regulations of the National Institute of Neuroscience (Tokyo, Japan) for animal experiments, and were approved by the Institutional Animal Investigation Committee. All efforts were made to minimize animal suffering and to reduce the number of animals used. Arc-dVenus mice were generated as previously described (Eguchi and Yamaguchi, 2009). The mice were male transgenic heterozygotes, bred and housed at the Animal Center of the National Institute of Neuroscience, National Center of Neurology and Psychiatry (Tokyo, Japan).

L-Arginine hydrochloride (Arg), L-N^G-nitroarginine methyl ester (L-NAME), L-phenylalanine (Phe), and CCK were purchased from Sigma-Aldrich (St. Louis, MO), and dissolved in saline. Arg was used to avoid altering the pH of the solution. Arg, L-NAME, and Phe were dissolved in saline at 150 mM. CCK was dissolved in saline at 0.44 mM. For dose-response experiments, 10 mM Arg and 2.20 mM CCK solutions were also used.

Special attention was paid to minimizing Venus and c-Fos expression artifacts. In particular, as c-Fos and Arc expression are stimulated by exposure to novel environments (Ramirez-Amaya et al., 2005), we devised a method in which the mice were anesthetized in their home cage. Each mouse was transferred from their home cage to a quasi-home cage at 24 h before the experiments. The quasi-home cage was constructed from the same plastic, was the same shape as the home cage, and was filled with wood chips at 1-cm depth, consistent with the home cage. A 4-mm diameter hole was made near the bottom of one wall, and at the time of anesthesia, the outlet of a sevoflurane anesthesia apparatus was connected to the hole. The mice were thus anesthetized in a familiar environment without necessity for transfer to a novel environment. The mice were fasted the night before to avoid the possible variation in the IEG expression by feeding behavior.

Once anesthetized, the mouse was picked up and its head was inserted into a tube through which sevoflurane was continuously administered. Under stable anesthesia, a laparotomy was carried out in the dorsal position and the hepatic portal vein was exposed. Saline, Arg, Arg+L-NAME, Phe, or CCK was injected into the portal vein over 30 s using a 30-gauge needle with a 100-μL Hamilton-syringe. The injected dose of Arg and CCK was 0.61 mmol/kg and 1.75 nmol/kg, respectively, based on previous studies showing enhanced CHBV electrical activity after intraportal administration of these compounds (Niijima and Meguid, 1995; Cox and Randich, 1997; Horn and Friedman, 2004). The injected dose of Phe was 0.61 mmol/kg. In the dose-response experiments, 0.04 mmol/kg Arg and 9.41 nmol/kg CCK were used. After injection, the needle was carefully removed to avoid induction of bleeding, and the abdominal muscles and skin were separately sutured. Any mouse with significant bleeding from the portal vein was removed from the experiment. Following the procedure, each mouse was transferred back to its home cage, and allowed to recover naturally from the anesthesia over several minutes.

At 90 and 180 min after surgery, to allow for expression of Venus (180 min, Eguchi and Yamaguchi, 2009) and c-Fos (90 and 180 min, Young et al., 1991; Hughes and Dragunow, 1995), each mouse was deeply anesthetized with fluothane gas, and transcardially perfused with phosphate-buffered saline (PBS) and 4% paraformaldehyde. The brain was removed from the skull, and post-fixed overnight in 4% paraformaldehyde. Coronal slices (50-μm thickness) were prepared using a vibratome (Sekiguchi et al., 2009). Then, slices were selected according to their bregma level (Paxinos and Franklin, 2001) and were put on the cover slips without sealing and Venus expression was observed directly using a BZ-X710 All-in-One Fluorescence Microscope (Keyence, Tokyo, Japan). A digital file containing a brain atlas (Paxinos and Franklin, 2001) was superimposed over each photomicrograph to specify the brain regions. The number of Venus positive cells was counted manually by an observer blinded to the treatment conditions. The counts from an individual animal (only one slice for each bregma level) are presented for a single data point.

Slices to be used in c-Fos immunostaining were selected from the slices used in analysis of Venus expression, according to their bregma level (Paxinos and Franklin, 2001). These slices were incubated in PBS containing 3% normal goat serum (Vector Laboratories, Burlingame, CA) and 0.3% Triton X-100 for 1 h at room temperature, treated with anti-c-Fos antibodies (1:5,000 dilution; Santa Cruz Biotechnology, Dallas, TX) overnight at 4°C in PBS, and washed with PBS. The signals were visualized using the avidin-biotin-peroxidase complex method (Vector Laboratories, Burlingame, CA) as previously described (Odagiri et al., 2012). The number of c-Fos positive cells was counted manually by an observer blinded to the treatment conditions. The counts from an individual animal (only one slice for each bregma level) are presented for a single data point.

Surgery for selective CHBV vagotomy was carried out as previously described (Uno et al., 2006, 2015). Briefly, a laparotomy was performed in the dorsal position as described above. The descending ventral esophagus was revealed by gently retracting the stomach after severing the gastrohepatic ligament. The CHBV, diverging from the ventral subdiaphragmatic vagal trunk immediately caudal to the diaphragm, was then transected near the vagal trunk. The abdominal muscle and skin were separately sutured. The mice were used for portal vein injection at 7 days after surgery. The hepatic branch of the splanchnic nerve was left intact (Uno et al., 2015).

Data were expressed as mean ± standard error of mean (SEM) and analyzed by one-way analysis of variance (ANOVA). Significant ANOVA results were analyzed by a post hoc Bonferroni test. P < 0.05 was considered statistically significant.

Figure 1A shows a photomicrograph of part of a brain slice (bregma −1.06 mm) prepared from an Arc-dVenus mouse intraportally injected with saline. The inset is the corresponding area from the brain atlas of Paxinos and Franklin (2001). The solid white line indicates the area of the insular cortex, and the bright green dots are cells expressing the fluorescent protein Venus. Figures 1B,C show photomicrographs of brain slices from mice injected with Arg (0.61 mmol/kg) and CCK (1.75 nmol/kg), respectively. More Venus positive cells were observed in the insular cortex of the Arg-injected mouse (Figure 1B) than in the insular cortex of the saline-injected mouse (Figure 1A). No such increase was observed in the CCK-injected mouse (Figure 1C). Figure 1D shows a similar photomicrograph of a brain slice from an Arg (0.61 mmol/kg)-injected Arc-dVenus mouse that had undergone CHBV vagotomy 7 days earlier (Arg+vag). The number of Venus positive cells in the vagotomized mouse was similar to that in the saline-injected mouse (Figure 1A), but smaller than that in the Arg-injected non-vagotomized mouse (Figure 1B).

A previous study showed that the insular cortex in rats can be roughly divided into two areas, an anterior (taste) area and a posterior (visceral) area (Cechetto and Saper, 1987). In the case of mice, it can be roughly estimated that the insular cortex starts near bregma +2.4 mm and ends near bregma −1.2 mm, with the joining of the anterior commissure at approximately bregma +0.14 mm (Paxinos and Franklin, 2001). Taking this into account, we analyzed the effect of Arg (0.61 mmol/kg) on Venus expression at bregma +0.62, −0.34, −0.58, and −0.82 mm as well as at bregma −1.06 mm (Figure 2). The number of animals used in each treatment group was 9 for saline, 8 for Arg, 7 for Arg+vag, 4 for CCK, and 4 for CCK+vag. Bregma +0.62 mm is expected to correspond to the taste area (Cechetto and Saper, 1987). As shown in Figure 2, a significant Arg-induced increase in the Venus positive cell number was observed in the posterior portion of the insular cortex at bregma −0.82 mm (P < 0.01), as well as at bregma −1.06 mm (P < 0.001) [F_{(4, 27)} = 5.73, P = 0.0018 for bregma −0.82 mm; F_{(4, 27)} = 6.89, P = 0.0006 for bregma –1.06 mm; one-way ANOVA]. No significant increase was observed in the more anterior portions, such as bregma +0.62, −0.34, and −0.58 mm. The effect of Arg at bregma −0.82 and −1.06 mm was abolished in vagotomized mice (Figure 2). The action of Arg at bregma +0.62 mm in vagotomized mice was not tested (Figure 2, NT), because Arg did not induce a significant increase in the number of Venus positive cells in non-vagotomized mice. In contrast, intraportal CCK (1.75 nmol/kg) did not increase the number of Venus positive cells at bregma −0.34, −0.58, −0.82, and −1.06 mm. The action of CCK at bregma +0.62 mm was not tested (Figure 2, NT), because Arg did not induce a significant increase in the number of Venus positive cells in non-vagotomized mice. We also examined laterality of the CHBV-dependent action of intraportal Arg in the insular cortex. In the same slices (at bregma −1.06 mm) used in this experiment, the number of Venus positive cells were counted in left and right hemispheres separately. Similar action of intraportal Arg on Venus expression was observed both in left and right hemispheres (Figure 2-inset).

The above results suggest that the number of Venus positive cells is enhanced in the posterior insular cortex in response to intraportal Arg in a CHBV-dependent manner, which is consistent with the view that the posterior insular cortex contains a visceral area.

Figure 3A shows a photomicrograph of part of a brain slice (bregma −1.06 mm) prepared from a mouse intraportally injected with saline. The solid black line indicates the area of the insular cortex, and the black dots are immunoreactive signals for c-Fos. Figures 3B,C show photomicrographs of brain slices from mice injected with Arg (0.61 mmol/kg) and CCK (1.75 nmol/kg), respectively. More c-Fos positive cells were observed in the insular cortex of the Arg-injected mouse (Figure 3B) than in the insular cortex of the saline-injected mouse (Figure 3A). No such increase was observed in the CCK-injected mouse (Figure 3C). Similar experiments were performed in 5 saline-injected, 5 Arg-injected, and 4 CCK-injected mice, and the mean values (+ SEM) obtained in each group are shown in Figure 3D. One-way ANOVA revealed significant variation in the data [F_{(2, 11)} = 9.06, P = 0.005]. A post hoc Bonferroni test revealed a significant difference between saline-injected mice and Arg-injected mice (P < 0.05). Intraportal CCK had no significant effect on the number of c-Fos positive cells (−1.06 mm in Figure 3D). In addition, we performed similar analysis at the bregma level of −0.34 mm, a more rostral position, because intraportal Arg does not increase the number of Venus positive cells at this bregma level. The number of c-fos positive cells was not increased either at this level (−0.34 mm in Figure 3D).

As we mentioned earlier, we chose a time point of 180 min after intraportal administration for the analysis of c-Fos expression because we wanted to analyze the expression of c-Fos in the same slice that was analyzed the expression of Venus (analyzed at 180 min post-administration). However, a number of studies have suggested that peak responses of c-Fos protein expression occur within 60–120 min post-stimuli (Hughes and Dragunow, 1995). Accordingly, we also tested a time point of 90 min after CCK administration (Figure 3D-inset) to check the possibility that the c-Fos levels related to CCK would likely have been on the decline or gone at 180 min after CCK administration. In this case, it is expected that the increase in expression of c-Fos can be observed at 90 min after CCK administration. The number of animals used was 3 for saline and 5 for CCK. As Figure 3D-inset shows, however, there was no increases in the c-Fos level at 90 min after CCK administration. These results suggest that intraportal CCK does not increase the number of c-Fos positive cells in the insular cortex at bregma −0.34 and −1.06 mm.

To investigate the regional selectivity of the action of intraportal Arg, we examined the number of Venus positive cells in other cortices observed at the bregma −1.06 mm level, namely the piriform cortex, insular cortex, secondary somatosensory cortex (S2), primary somatosensory cortex (S1) barrel field (S1BF), S1 composed of dysgranular region (DZ), shoulder region (Sh), and hindlimb region (HL), primary and secondary motor cortices (M1 and M2), and region formed by retrosplenial agranular (RSA) and granular (RSG) cortices. The number of mice used were 9 for saline, 8 for Arg, 7 for Arg+vag, 4 for CCK, and 4 for CCK+vag. Except for the insular cortex, there were no regions in which Arg induced an increase in the number of Venus positive cells in a vagotomy-dependent manner [in the insular cortex, F_{(4, 27)} = 7.64, P = 0.0003 with one-way ANOVA; P < 0.001 between saline and Arg and P < 0.01 between Arg and Arg+vag with post hoc Bonferroni test, Figure 4]. There were significant increases in the number of Venus positive cells after injection of Arg in the S2 [F_{(4, 27)} = 3.69, P = 0.016 with one-way ANOVA; P < 0.05 with post hoc Bonferroni test, Figure 4]. However, vagotomy did not significantly reverse the Arg-induced increases in this region (comparison between Arg and Arg+vag, Figure 4). CCK did not induce any clear changes in the number of Venus positive cells in any of the cortical regions examined. These results suggest that there is region-specificity in the CHBV-dependent action of Arg.

To characterize the action of intraportal Arg in the insular cortex, we examined (1) the dose-dependency of the action of Arg, (2) the influence of an inhibitor for nitric oxide synthesis from Arg (L-NAME), and (3) the actions of intraportal Phe, which was shown to have no effect on CHBV neurofilament electrophysiological activity (Niijima and Meguid, 1995), in Arc-dVenus mouse brain slices at bregma −1.06 mm. The results are summarized in Figure 5, and ANOVA for these data showed significant variation [F_{(6, 32)} = 3.26, P = 0.013]. The number of mice used were 9 for saline, 6 for Arg 0.04 mmol/kg, 8 for Arg 0.61 mmol/kg, 3 for Arg 0.61 mmol/kg + L-NAME, 5 for Phe 0.61 mmol/kg, 4 for CCK 1.75 nmol/kg, and 4 for CCK 9.41 nmol/kg. The values for saline and Arg 0.61 mmol/kg, and CCK 1.75 nmol/kg were reproduced from Figure 2 for comparison.

First, intraportal Arg at the lower dose (0.04 mmol/kg) did not significantly increase the Venus positive cell number in the insular cortex (Figure 5). Second, L-NAME was dissolved in the Arg 0.61 mmol/kg solution and this mixture was injected into the portal vein (Arg + L-NAME). In these mice, the number of Venus positive cells in the insular cortex (28.3 ± 3.2, Figure 5) did not significantly differ from the number elicited by intraportal injection of Arg 0.61 mmol/kg alone (34.6 ± 4.1). Finally, intraportal Phe 0.61 mmol/kg caused no significant change in the number of Venus positive cells in the insular cortex.

As we mentioned earlier, CCK at 1.75 nmol/kg had no effects on insular Venus expression. However, it remained possible that much higher doses were active. To clarify this issue, we injected a higher dose of CCK (9.41 nmol/kg) into the portal vein, but observed no significant effects on the number of Venus positive cells (Figure 5).

Next, we analyzed c-Fos expression in the medial NTS, to which the vagal afferents project in the CNS (Figure 6). The expression of c-Fos was analyzed because we could not detect Venus positive cells at all in the NTS. Figures 6A–E show the left NTS at bregma −7.5 mm (the area surrounded by a dotted line). The number of c-Fos positive cells (small black dots) was decreased in the left medial NTS by intraportal Arg (Figure 6A vs. Figure 6B). The number of c-Fos positive cells was not decreased in this Arg-injected vagotomized mouse (Figure 6C) when compared with a saline-injected mouse (Figure 6A). In contrast, intraportal CCK markedly increased this number (Figure 6A vs. Figure 6D), and the increase was not observed in a vagotomized mouse (Figure 6E).

The effects of Arg and CCK in the left medial NTS were examined using several mice, and the results were summarized in Figure 6F (4–9 mice for each group at all four bregma levels, see a Figure 6 legend in detail). In the left medial NTS at bregma −7.5 mm, the mean number of c-Fos positive cells was significantly [F_{(4, 25)} = 20.54, P < 0.0001 with one-way ANOVA, P < 0.01 with post hoc Bonferroni's test] decreased by intraportal Arg, which was not observed in vagotomized mice. On the contrary, the mean number of c-Fos positive cells was significantly (P < 0.0001) increased by intraportal CCK, which was not observed in vagotomized mice. At bregma −7.2 and −7.7 mm similar effects of intraportal CCK were observed [F_{(4, 23)} = 8.35, P = 0.0003 with one-way ANOVA, P < 0.001 with post hoc Bonferroni test for bregma −7.2 mm; F_{(4, 24)} = 5.85, P = 0.002 with one-way ANOVA and P < 0.01 with post hoc Bonferroni test for bregma −7.7 mm; Figure 6F] but the effects of Arg could not be observed. At bregma −6.9 mm, neither Arg nor CCK have effects upon the number of c-Fos positive cells. There were no significant changes in the number of c-Fos positive cells in the subnucleus gelatinosus and area postrema at bregma −7.5 mm (data not shown).

In the right NTS (Figure 6G, the number of slices used is the same as the left), on the other hand, the significant decrease in the number of c-Fos positive cells by intraportal Arg was not observed at all in four bregma levels tested. In the case of CCK, the vagotomy-sensitive increase in the number of c-Fos positive cells by intraportal CCK was significant at bregma −7.2 and −7.5 mm, and was not observed at bregma −6.9 and −7.7 mm [F_{(4, 23)} = 9.79, P < 0.0001 with one-way ANOVA and P < 0.01 with post hoc Bonferroni test for bregma −7.2 mm; F_{(4, 25)} = 6.84, P = 0.0007 with one-way ANOVA and P < 0.01 with post hoc Bonferroni test for bregma −7.5 mm].

These results suggest that intraportal Arg induces the vagotomy-sensitive decrease in the number of c-Fos positive cells in a left NTS specific manner, and that intraportal CCK induces the vagotomy-sensitive increase in the number without outstanding laterality in the NTS.

Finally, we analyzed c-Fos expression in the ventral portion of the dorsomedial hypothalamus at bregma level of −1.8 mm, because peripheral injection of CCK is known to induce c-Fos expression in this region (Kobelt et al., 2006). The expression of c-Fos was analyzed because we could not detect Venus positive cells at all in the hypothalamus. The number of c-Fos positive cells was apparently increased by intraportal CCK (1.75 nmol/kg) in the ventral portion of the dorsomedial hypothalamus, but not by intraportal Arg (comparisons among Figures 7A–C). The effect of CCK was absent in a vagotomized mouse (Figure 7D). The mean number of c-Fos positive cells is summarized in Figure 7E (data from 5 saline-injected mice, 7 Arg-injected mice, 4 CCK-injected mice, and 4 CCK-injected vagotomized mice). One-way ANOVA revealed significant variation in these data [F_{(3, 16)} = 11.47, P = 0.0003]. The number of c-Fos positive cells was significantly (P < 0.01) increased by intraportal CCK, and the action of CCK was significantly (P < 0.05) suppressed in vagotomized mice. There was significant difference between Arg-injected and CCK-injected mice (P < 0.0001). Meanwhile, there was no significant difference between saline-injected and Arg-injected mice. There was no significant difference between saline- and Arg-injected Arc-dVenus mice fixed at 90 min after portal vein injection (3 mice for saline and 5 mice for Arg, Figure 7E-inset). Therefore, it is likely that the absence of the increase in the number of c-Fos positive cells by Arg in the dorsomedial hypothalamus is not due to disappearance of the Arg response at 180 min post-administration.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.