Intestinal transepithelial permeability of oxytocin into the blood is dependent on the receptor for advanced glycation end products in mice

Plasma oxytocin (OT) originates from secretion from the pituitary gland into the circulation and from absorption of OT in mother’s milk into the blood via intestinal permeability. However, the molecular mechanism underlying the absorption of OT remains unclear. Here, we report that plasma OT concentrations increased within 10 min after oral delivery in postnatal day 1–7 mice. However, in Receptors for Advanced Glycation End Products (RAGE) knockout mice after postnatal day 3, an identical OT increase was not observed. In adult mice, plasma OT was also increased in a RAGE-dependent manner after oral delivery or direct administration into the intestinal tract. Mass spectrometry evaluated that OT was absorbed intact. RAGE was abundant in the intestinal epithelial cells in both suckling pups and adults. These data highlight that OT is transmitted via a receptor-mediated process with RAGE and suggest that oral OT supplementation may be advantageous in OT drug development.


Results
In suckling male and female C57BL/6N mice at PND1, the plasma OT concentration in blood samples collected from the facial vein (Fig. 1a, inset) was higher in 33 pups fed with their mother's milk for 20 min than in pups who were fasted for 20 min (n = 14) (two-tailed Student's t-test, P < 0.05; Fig. 1a). No clear difference in the OT level was observed between male and female pups. This result suggests that milk-borne OT is transmitted from the mother to 1-day-old pups, similar to [ 3 H]OT transfer from rat mothers to neonates 11 . Plasma OT levels after oral delivery in suckling pups. The possibility of transport during the suckling period of development (from PND1 to PND15) was confirmed by comparing plasma OT levels at 10 min after oral delivery with 10 μL of 100 μM synthetic OT solution (1 μg/pup) or the same volume of phosphate buffered saline (PBS) to male and female pups after fasting for 30 min. The time course shown in Fig. 1b shows that plasma OT concentrations collected from the heart (Fig. 1b, inset) were markedly higher in newborn (PND1) and 2-day-old (PND2) pups than in those administered oral saline, similarly to the results of studies in suckling piglets 33 . The plasma OT concentrations after oral OT delivery remained at higher levels at PND3-5 than in the saline controls (two-way analysis of variance (ANOVA), F 8.90 = 3.32, P < 0.002: Bonferroni's post hoc tests for differences between OT and PNDs ( a P < 0.001, b P < 0.01, and c P < 0.05 from the value at PND8) and between OT and PBS ( d P < 0.01). The OT concentration dropped suddenly after PND6, although it remained higher than that in the saline control group. At PND8 and PND15, no concentration difference in plasma OT levels was observed between pups orally administered OT vs. those administered PBS.
The time profile of the plasma OT concentration after oral administration of OT peaked at 10 min in pups at PND6 (n = 5-8, one-way ANOVA, F 5,37 = 3.302, P < 0.02 from the control value; Fig. 1c). The plasma OT level depended on the oral OT dose in pups at PND6 (one-way ANOVA, F 3,24 = 20.35, P < 0.0001; Fig. 1d).
To confirm that the above increase after oral administration was primarily owing to permeability from the intestine, 5 μL of 100 μM OT (0.5 μg/pup) was applied directly into the small intestine, bypassing the mouth, buccal mucosal membrane, and/or upper parts of the digestive tract, including the oesophagus and stomach. The plasma OT in blood samples collected from the neck (Fig. 2a, inset) reached a peak at 10 min and decreased at 60 min and 120 min after intestinal application ( Supplementary Fig. 1). The mean plasma OT concentration in groups of 4-8 male and female pups at PND1-7 was significantly higher in the OT group than the PBS control group (two-way ANOVA, F 3,40 = 3.14, P < 0.05; Bonferroni's post hoc tests for differences between OT and PBS ( # P < 0.001) and OT and PNDs; Fig. 2a, red vs. blue bars). The plasma OT level was higher at PND1 and PND3 than at PND7 in the group injected with OT (***P < 0.001 and **P < 0.01; Fig. 2a).
The plasma OT concentration at 10 min after intestinal OT injection decreased by approximately 4.6-fold from PND1 to PND7 (Fig. 2a). However, this decrease may not simply reflect a decrease in the total amount of OT absorption and may also be related to the blood volume increase associated with body weight gain during development. Because it is difficult to measure the total blood volume, we estimated it from the body weight. We observed a 2.7-fold body weight gain, from 1.3 ± 0.03 g at PND1 (n = 26 male and female unique individual neonates; Fig. 2b, blue line) to 3.5 ± 0.08 g at PND7 (n = 57 pups). Considering this body weight gain and the linear blood volume increases with grams in neonates 43 , it is likely that total OT absorption slightly decreased, probably owing to intestinal barrier and/or digestive development. Altogether, these data in wild-type mice suggest that OT passes into the blood across the digestive tract, including the intestinal mucus and epithelial cells, and crosses the intestinal epithelial barrier 34, 37, 44 .
Plasma OT levels in RAGE KO pups. Using RAGE KO mice, we next examined OT increases beyond the basal level in the blood after oral OT administration (100 μM × 10 μL; 1 μg/pup). The plasma OT concentrations in cardiac samples (Fig. 3a, inset) in Ager −/− male and female pups (n = 38) were significantly increased from PND1-3 in the oral OT group than the saline group ( Fig. 3a; two-way ANOVA, F 8,90 = 5.16, P < 0.001). Bonferroni's post hoc tests for differences between OT and PBS (P < 0.001) and between OT and PNDs ( a P < 0.001, b P < 0.01, and c P < 0.05 from PND8, respectively). The OT concentration dropped suddenly at PND4 in RAGE KO pups, and there were no subsequent differences in OT concentrations between pups orally administered OT or PBS.
The elevated levels of OT at PND1-3 were similar between the two genotypes ( Fig. 3b). However, interestingly, from PND4 to PND6, significantly lower plasma OT concentrations were observed in RAGE KO mice with orally administered OT than in wild-type mice (two-way ANOVA, F 1,126 = 4.73, P < 0.05). Bonferroni's post hoc tests male and female pups after fasting for 20 min and with or without feeding with dam's milk for 20 min. N = 14 and 33 mice. *P < 0.05, two-tailed Student's t-test. Inset shows blood collection from the facial vein. (b) Time course of plasma OT levels in pups 10 min after oral administration of OT (squares; 1 μg of OT/mouse; 10 μL of 100 μM OT) or PBS (triangles; 10 μL). Experiments were carried out on the indicated postnatal days. The number of mice was 3-15 for OT and 3 for saline. A two-way ANOVA revealed a significant difference between the drug treatment and saline treatment, F 8,90 = 3.32, P < 0.002. Bonferroni's post hoc tests showed a P < 0.001, b P < 0.01 and c P < 0.05 from the value at PND8; d P < 0.01 for the saline group. Inset shows the scheme of OT administration to the mice and blood collection from the heart after opening the chest. (c) Time course of plasma OT levels of PND6 pups measured at the indicated times after oral administration of 1 μg of OT/mouse, N = 4-5 mice. One-way ANOVA, F 5,37 = 3.302, P < 0.0001; *P < 0.02 from the control value. (d) Plasma OT concentrations in PND6 pups after 10 min of oral administration of OT at the indicated concentrations. N = 3 mice. One-way ANOVA, F 3,24 = 20.4, P < 0.0001. Bonferroni's post hoc tests showed ***P < 0.001 from other values. showed significant differences between wild-type and RAGE KO mice with *P < 0.05 (for PND5), **P < 0.01 (PND4), and ***P < 0.001 (PND6). This RAGE dependency was lost after PND7.
To test gene-dosage effects on OT absorption, we examined plasma OT concentrations in pups of three genotypes in the same litter at PND5, using the same method as in Fig. 3a. The plasma OT level in Ager +/+ pups was as high as that in Ager +/− pups and significantly higher than that in Ager −/− pups, as expected ( Fig. 3c; one-way ANOVA, F 2,22 = 6.851, P < 0.01; Bonferroni's post hoc test for the difference between Ager +/+ and Ager −/− , P < 0.01).
All these results show that the higher OT concentrations in neonates to 3-day-old pups of both genotypes were likely to be owing to non-selective transmission, probably resulting from leakage before closure of the intestinal barrier (as indicated by +++ in Fig. 3d; ref. 34, 37). Then, after PND4, permeability decreased, owing to cessation of non-selective transmission, although selective transmission remained after closure of the epithelium and barrier formation. Alternatively, a small nutritional abnormality in KO mice during PND3-6 may be owing to disturbance in the RAGE-dependent absorption of general nutrients. Thus, OT may permeate the intestine through a RAGE-dependent receptor mechanism (Fig. 3d).
OT transport from isolated intestines of suckling pups. Non-selective and selective transmission of OT from the intestinal mucosa was examined (Fig. 4), using isolated segments of the upper intestine. Intestinal segments (4 cm in length, from 1-2 cm below the stomach) were incubated in 3 ml of saline solution in a chamber at room temperature (Fig. 4c). OT (10 μM × 50 μL; 0.5 μg/intestine) was injected directly into the mucosal lumen at the distal cutting edge, which was isolated from the incubated segments with glycerol. OT concentrations in the incubation medium were monitored for 10 min. Figure 4a shows that high OT levels were achieved soon after the start of incubation, and there were no differences between genotypes in neonates (PND1). In 5-day-old pups (PND5), the OT concentrations gradually increased after smaller instantaneous increases in both wild-type (C57BL6) and Ager −/− (RAGE KO) pups (Fig. 4b). A two-way ANOVA for interaction between Genotypes × Time revealed no significance (F 5,84 = 1.39, P = 0.2346), but a significant difference was found for the Genotype effect (F 1,84 = 31.02, P < 0.0001) and the Time effect (F 5,84 = 7.41, P < 0.0001). Bonferroni's post hoc tests showed significant differences with P < 0.05 between wild-type and RAGE KO mice during 3-10 min (Fig. 4b), with respect to OT concentrations in the incubation medium.
It is of interest to note that transluminal absorption in the wild-type mice at 10 min was not different in vitro at P1 and P5 (Fig. 4) whereas there was a decrease in transluminal absorption sampled at 10 min for each age (Fig. 2a). This is probably owing to the denominator (body volume) being kept constant; during in vivo experiments, the denominator (pup "volume") was not held constant because the pups were larger at older ages, resulting in more dilute recovered OT.
The leakage level of the isolated intestines was examined by fluorescein isothiocyanate (FITC)-dextran permeability 45 . Leakage levels were similar between the two genotypes at the two different PNDs (n = 3, P = 0.67 for PND1 and P = 0.59 for PND6; Table 1). This suggests that the isolated intestines of both genotypes were functional, without much damage, although transepithelial electrical resistance should be measured to determine epithelial barrier formation 45 . Plasma OT levels in adult mice. In adult male mice (8 weeks old), when oral OT was administered at the same dose used for neonates (1 μg/mouse), as in Fig. 1b, no increase in the plasma OT concentration was observed (data not shown). To examine whether permeability exists in the adult intestine after barrier formation, a 10-fold higher volume of OT (100 μM × 100 μL) was used for oral delivery (10 μg/mouse). Blood samples were collected from anesthetized mice using a tail snip procedure, and OT concentrations were measured (Fig. 5a, inset). A significant increase in plasma OT was observed at 10 min after administration in wild-type (C57BL/6) mice whereas the levels in RAGE KO mice were significantly lower ( Fig. 5a; two-way ANOVA for genotype differences, F 1,48 = 4.66, P < 0.05). Bonferroni's post hoc tests showed significant differences in OT concentrations Inset indicates oral OT delivery and blood collection from the heart. The number of experiments was 3-15 for OT and 3 for saline. Two-way ANOVA, F 8,90 = 5.16, P < 0.001. Bonferroni's post hoc tests showed significant differences between OT and PBS ( d P < 0.001) or OT and PNDs, as indicated ( a P < 0.001, b P < 0.01, c P < 0.05 from PND8). (b) Plasma OT levels in wild-type and RAGE KO pups 10 min after oral administration of OT. The data are re-plotted from Figs 1b and 3a. Two-way ANOVA, F 1,126 = 4.73, P < 0.05. Bonferroni's post hoc tests showed significant differences between wild-type and RAGE KO mice at *P < 0.05, **P < 0.01, and ***P < 0.001. (c) Plasma OT concentrations in Ager +/+ (n = 4), Ager +/− (n = 8) and Ager −/− (n = 3) pups of the litter mates of 3 dams at PND5 after oral OT administration as in a. One-way ANOVA, F 2,22 = 6.851, P < 0.01. Bonferroni's post hoc tests revealed a difference between Ager +/+ and Ager −/− mice, P < 0.01. (d) The scheme illustrates the relationship between plasma OT increases and barrier formation, digestion, and RAGEdependent absorption after OT delivery during development. 0, + , + + , + + + indicates from little to no permeability to the highest permeability and resulting plasma OT levels.
between the two genotypes after oral OT administration ( # P < 0.01). A two-way ANOVA (F 5,48 = 4.53, P < 0.002) indicated a difference between Time and Genotypes, which was significantly different from the control value with *P < 0.05 at 30 min and **P < 0.01 at 10 min after OT administration in wild-type mice. The basal OT level before administration of OT or after oral saline administration was similarly low in both genotypes (Fig. 5a, blue line), suggesting that anesthetization of the mice did not influence the OT levels.
The OT concentrations were determined in plasma obtained from tails after direct administration of OT into the intestine (100 μM × 100 μL; 10 μg/intestine), which is partially exposed to the abdominal cavity during injection, to avoid contamination from other tissues (Fig. 5b, inset). The OT level significantly increased after 10 min of direct OT delivery into the mucosa of the upper intestine; this level was very low in Ager −/− mice ( Fig. 5b; two-way ANOVA for interaction between Genotypes and Time, F 15,168 = 4.25, P < 0.0001). A two-way ANOVA  Table 1. Intestinal permeability measured by FITC-dextran. To assess intestinal tissue permeability using a FITC-dextran assay, 4-kDa FITC-dextran was injected into the mucosal side of the isolated intestine as in Fig. 4. Samples were collected at the beginning of incubation and after 3 min. The fluorescence of the samples was quantified by a spectrofluorophotometer (Model RF6000, Shimadzu, Kyoto, Japan): the excitation wavelength was 485 nm, and the emission wavelength was 538 nm. The result obtained at 0 min (arbitrary units) was subtracted from the result obtained at 3 min, and this difference was then divided by 3 min. N = 3 for each genotype. Two-tailed Student t-test: P = 0.67 between a and b; P = 0.59 between c and d; P < 0.01 between a and c; P < 0.002 between b and d, respectively.
for the Genotype × Treatment effect revealed significant differences, F 3,168 = 8.58, P < 0.0001 and for the Time effect, F 5,168 = 7.58, P < 0.0001. Bonferroni's post hoc tests show significant differences between saline and OT in wild-type mice with ***P < 0.001 and *P < 0.05, between saline and OT in RAGE KO mice with # P < 0.05, and OT levels between wild-type and KO mice with ## P < 0.001. The plasma OT concentration increased in a linear fashion as a function of the dose applied directly to the intestine ( Supplementary Fig. 2). In addition, the same volume of saline delivered into the intestine did not increase the plasma OT level.

Vasopressin levels in the plasma after oral delivery and transport from isolated intestines.
To study the specificity of RAGE-dependent transport of OT, we investigated the intestinal absorption of another nonapeptide, arginine vasopressin (AVP 46 ) by oral delivery and in the isolated intestines. Transport during the suckling period of development (PND3 and PND5) was confirmed by comparing plasma AVP levels at 10 min after oral delivery of 10 μL of synthetic AVP solution (100 μM; 1 μg/pup) or the same volume of PBS to pups after fasting for 30 min. Plasma AVP concentrations were markedly higher at both PND3 and PND5 in pups administered AVP than in those administered oral saline (Fig. 6a). Two-way ANOVA was used to analyse the interaction between Treatment (PBS/AVP) × Age, F 1,18 = 3.14, P = 0.0933; the Treatment (PBS/AVP) effect, F 1,18 = 5.21, P < 0.05; and the Age effect, F 1,18 = 91.26, P < 0.0001. Bonferroni's post hoc tests were used to determine differences in AVP levels between PND3 and PND5, *P < 0.05. It is worth noting that the pup size denominator problem was also present in the above experiment. However, if we roughly estimate AVP recovery, it is approximately 4% at PND3 (22 ng was recovered from the effective dose of 555 ng (1 μg/1.8 g pup)) and also 4% at PND5 (16 ng was recovered from the effective dose of 400 ng (1 μg/2.5 g pup)).
Transmission of AVP from the intestinal mucosa was examined using the same paradigm as that in Fig. 4a. AVP was injected directly into the mucosal lumen (5 μL of 100 μM AVP), and AVP concentrations in the incubation medium were monitored (Fig. 6b). Relatively high AVP levels were achieved soon after the start of incubation in neonates (PND1). In 5-day-old pups (PND5), the AVP concentration was much lower. Two-way ANOVA was used to analyse differences in interaction between Time × Age, F 4,40 = 0.15, P = 0.9616; the Time effect, F 4,40 = 0.27, P = 0.8951; and the Age effect, F 1,40 = 25.95, P < 0.0001. Bonferroni's post hoc tests showed differences in AVP levels between PND1 and PND5, **P < 0.01 (Fig. 6b).
To estimate the amount of released AVP, we calculated the area under the curve (AUC). The AUC values of AVP were significantly lower than those of OT (Fig. 6c). Two-way ANOVA showed a significant difference between Genotypes (F 1,16 = 6.37, *P < 0.05.) and Treatment effect (F 1,16 = 39.92, P < 0.05). In addition, there was a significant decrease after development for OT in RAGE-KO pups and AVP in wild-type pups for both ( # P < 0.05 for PND1). This suggests that AVP is sensitive to barrier formation, although it is necessary to measure it in KO mice.
Determination of OT by mass spectroscopy. Next, we examined whether increased OT after oral administration was owing to the permeability of exogenous OT across the intestinal barrier in an undigested form. The presence of the intact form of OT in the plasma was measured on the basis of the molecular weight, which was identified using LC-MS/MS 47,48 .
To distinguish exogenous OT from endogenous OT in the plasma, we first synthesised two non-radioactive, stably labelled isoleucine [Ile 13 C, 15  Plasma samples were collected from the heart of anesthetized wild-type PND3 pups at 10 min after oral administration of [Ile 13 C, 15   Note that the AUC values of OT in wild-type pups at PND1 and PND5 and KO mice at PND1 were similar, while the AUC value of OT in PND5 KO pups was significantly lower. A two-way ANOVA showed a significant difference between the Genotypes, F 1,16 = 6.37, # P < 0.05. The AUC values of AVP in wild-type pups at PND1 and 5 were lower than those of OT in wild-type pups. A two-way ANOVA revealed a significant difference for the Treatment effect, F 1,16 = 39.92, *P < 0.05 Scientific RepoRts | 7: 7883 | DOI:10.1038/s41598-017-07949-4 [Ile 13 C, 15  Intestinal staining with anti-RAGE antibody. Expression of RAGE in the intestine has been demonstrated in adult humans and rodents 42 , but not in human infants and suckling mice. Therefore, we performed immunohistochemistry using anti-RAGE antibody (Fig. 8). Figure 8c and e illustrates RAGE immunopositive staining in the epithelial cells of the villi of PND3 pups. No identical staining was detected in Ager −/− pups (Fig. 8d). RAGE expression was also detected in E18.5 embryos (Supplementary Fig. 3). After extensive washing of the intestine, this staining remained unchanged, thereby suggesting that this method is likely to detect the membrane-bound form of RAGE as a component of intestinal epithelial cells. Finally, we confirmed no apparent structural abnormalities in the villi of Ager −/− pups by analysis of hematoxylin and eosin stained intestine tissue (Fig. 8a,b,f and g): two-tailed Student's t-test between wild-type and RAGE KO mice, P = 0.34 for villi length and P = 0.18 for villi width (n = 30 counts from 3 mice).

Discussion
Our results provide the first reported evidence of RAGE-mediated transport of OT from the intestinal tract into the blood circulation across the intestinal barrier in mice. This intestinal permeability showed an apparent difference, on the basis of plasma OT concentrations, between wild-type and RAGE KO pups at PND4-6 after oral administration of OT. This capacity was also demonstrated in adults when we used a 10-fold concentrated OT solution. Although it is known that OT is absorbed in neonates and infants 11 , our results showed that increased levels of plasma OT in both wild-type and RAGE KO mice at approximately 1-3 days after birth were equally high, probably because of nonspecific bulk absorption (Fig. 3d). However, our experiments using incubated segments of intestine provided direct evidence for the transportation of OT from the intestinal lumen into the incubation medium.
Primarily adult RAGE KO mice have been used to study various metabolic diseases [38][39][40][41][42] . However, KO mice do not usually display any apparent developmental impairments 40 . In the current study, after maturing, the body weight of the two genotypes (wild-type and RAGE KO) was identical. The apparent intestinal structure of the villi at microscopic level of the two types of mice was statistically no different (Fig. 8). Intestinal permeability has been shown to have mouse strain heterogeneity 45 . In this study, we used the same C57BL/6 background in both wild-type and KO mice. The leakage levels were nearly identical between the two types, and the intestinal barrier  Table 1). These data clearly suggest that the Ager gene deletion resulted in modest changes in nutritional malabsorption and abnormal development, but specifically disrupted OT absorption. Therefore, this is the first report of RAGE-dependent intestinal absorption of OT at the molecular level, which we identified by comparing wild-type and Ager knockout mice.
OT absorption through the intestine was more difficult to detect in adults. The reasons for this difference with age may be owing to the progressive digestion of OT and strengthening of the gastrointestinal tract barrier. Similarly, a gradual decline in plasma OT concentrations during development appears to be caused by establishment of the intestinal barrier and intestinal digestion (Ref. 49 and Fig. 3d). Digestion is at a low level at birth and increases progressively during the first week in mice and humans 49 . However, digestion of OT takes more than 30 min and is much longer in mucosal tissues 17,50 . The increase in plasma OT within 10 min of OT administration may indicate that a relatively large amount of OT is resistant to digestion and is instead transmitted across the intestinal barrier. This result suggests that novel delivery methods for oral OT supplementation may be advantageous for the treatment of a range of conditions, such as impaired social development in preterm infants who are at risk of autism spectrum disorder, and obesity or diabetes in adults.

Estimation of the bioavailability of OT through the intestine. An estimation of bioavailability was
performed in adult mice. Fifty ng/mL of OT was detected after oral delivery of 10 μg of OT. Young adults weigh approximately 30 g and have 2.3 mL blood, so approximately 1% of oral OT can be absorbed. We have previously were immunostained with an anti-RAGE antibody (green); the nuclei were stained with DAPI. Bar = 100 μm. Co-staining of RAGE with DAPI indicates that RAGE is present in the intestinal villi in an enlarged image (e). (f and g) Morphological parameters such as villus length and villus width were measured from images with haematoxylin and eosin staining. Note that there were no significant differences in these parameters between the two groups N = 3.
identified OT in the mammary tissues of mouse dams and in the milk curd inside the oesophagi of pups 13 . To determine the OT concentrations in milk, we needed to collect milk from the mothers. One common method is to collect milk after stimulating secretion with external OT. Because such milk contains external OT, we were unable to calculate the OT concentrations in milk. However, we carried out a new collection method without exogenous OT stimulation 15 . We obtained samples at 5 days after parturition, and the concentration of milk OT was 1,200 pg/mL. If we estimate a pup's milk intake to be 100 μL, the pups will ingest 120 pg of OT. In agreement with these calculations, the increase after suckling was approximately 100 pg/mL of OT. If the total blood volume in PND5 pups is 200 μL, as calculated above, the increase is 20 pg of OT out of 120 pg, resulting in 16.6% availability. According to Tanaka et al. 11 , when [ 3 H]OT is administered to lactating rats, the radioactivity of intact OT in the neonatal content is 4.4% of the counts obtained for the maternal plasma. These results suggest that OT in maternal blood can be transferred to milk and then to neonates. In a similar estimation, the bioavailability of 1-deamino-8-arginine vasopressin in adult humans has been examined after cutaneous, intranasal, oral, sublingual, and intrarectal administration; the bioavailability has been found to be 3.4% for intranasal, 0.1% for oral, 0% for sublingual, 0% for intrarectal, and 100% for cutaneous delivery 50 . In comparison with these results, our findings indicated a much higher level of absorption. However, it should also be noted that OT can enter the lymph and intraperitoneal space, and tight junctions and epithelial barriers are underdeveloped in all of these areas in neonates.
Application of a similar dose (10 μg) of OT to adult male mice produced higher plasma levels when administered orally (approximately 38,000 pg/mL) as compared with intestinal administration (8,000 pg/mL), as shown in Fig. 5. These results seem counterintuitive as intestinal injection close to the intestinal absorption site would seem to be more effective. However, if we assume that there are multiple absorption sites, including the buccal and oesophageal mucosa 51-53 , longer processes from oral administration could facilitate transportation under less degradation of OT. Although it has not reported in the oesophagus, OT is thought to cross the mucus-unstirred water layer in the oral cavity and then enter the circulation 54,55 . Such buccal OT administration has been used to induce labour since 1964 56 .
In Fig. 7, the concentration of [Ile 13 C, 15 N]OT calculated was 27.9 ng/mL from the ratio of both peak areas and the conversion coefficient. If we take into account the dilution of the initial sample (x10), the estimated mean concentration in the blood is 346 ± 96 ng/mL (n = 4). However, this mean value is too high, even compared with values measured by mass spectrometry 47 using deuterated OT 48 or after a specific reduction-alkylation procedure 57 . The recovery was approximately 60%, which was calculated from total OT concentrations (600 ng in the body weight of 1. 5 g at PND3) by the injected concentration (1000 ng). This value is not far from the reported extraction efficiency and absolute recovery (above 65.8%; ref. 47). However, this raises a problem in that there is a large discrepancy in recovery between LC-MS/MS and enzyme immunoassy measurements of approximately 1%, as discussed above. Although methods to determine blood OT concentrations have not yet been established, this must be considered for such questions in the future. Nonetheless, the identical MRM peaks were scarcely detected in the plasma of RAGE KO pups at PND4-9 or after PND8 in the wild-type mice in our preliminary experiments. Our mass spectroscopy results suggest permeability as a full length form of [Ile 13 C, 15 N]OT, with no digestion, in wild-type pups after oral administration.
Possible mechanisms of OT permeability through the epithelial barrier. The gastrointestinal epithelium forms the boundary between the body and external environment and enables selective transportation through a permeable barrier. Transport routes through the intestinal epithelium include intracellular pathways and intercellular transport through tight junctions 34, 51, 52 . We hypothesized that RAGE functions through transcellular transportation because of its presence at the surface of villi rather than the epithelial cell body. However, we cannot ignore the possibility of rapid transportation through the paracellular pathway.
Future experiments will require in vitro measurements of RAGE binding to OT, including biochemical evidence of the binding of OT to RAGE with a molecular interaction tool. Such experiments are currently underway, and preliminary results show the binding affinity estimated by surface plasmon resonance to be in the micromolar range, which is lower than that of OT receptors at the subnanomolar range, thus suggesting that the OT-RAGE interaction is not identical to that between OT and its receptors.
In summary, expression of RAGE in the intestine affects OT absorption. This is the first report of a novel physiological role for RAGE beyond those in diabetes and cancer progression. Because many other ligands in addition to OT, such as AVP, can bind to RAGE, it is necessary to examine the specificity of this carrier role for OT. Instead of the inefficient nasal application of OT to healthy humans and autistic patients 58-61 , our results favour the oral administration of OT through enteric coated tablets. If feasible, this method for OT delivery may be used to treat diabetes, obesity, and autism because OT may play a role in fat loss and slowing weight gain 62, 63 .
The resultant Ager knockout (Ager −/− ) mice were then backcrossed to C57BL/6 mice (Charles River Japan) for more than eight generations, as described 65 . The Ager −/− mice were then maintained by crossing Ager −/− × Ager −/− mice for at least an additional 10 generations and then used for experiments. Ager +/− heterozygous mutants were obtained by crossing Ager +/− mice, and three genotypes of the same littermates were used in Fig. 3.
The offspring of these mice were born in our laboratory colony, weaned at 21-28 days of age and housed in same-sex groups of 3-5 animals until pairing 66 . Wild-type (C57BL/6) and Ager −/− mice were maintained under standard cage conditions (24 °C; 12-h light/dark cycle, lights on at 8:45 a.m.) with sawdust bedding and food and water provided ad libitum. Pregnant Ager −/− mice were transferred to clean cages 1 day before delivery. The offspring were either kept with biological or nonbiological mothers.
Virgin males and females were paired at 45 to 55 days. A single male and a single female were continuously housed together in a standard mouse maternity cage from the mating period until the delivery of pups. Administration of OT. Milk-borne OT was fed by mothers for 20 min after 20 min of starvation. A synthetic OT solution (100 µM; 10 μL for pups and 100 μL for adults) was administered orally. The same dose of [ILe 13 C, 15 N]OT was delivered orally to pups. For oral administration, pups or adult mice were held by hand, and the OT solution was dispensed into their mouths using polypropylene tips.
OT (100 μM) was injected through needles (28 G) into the proximal part (1~2 cm from the pylorus) of the small intestine of pups (5 μL) and adults (100 μL) under ketamine anaesthesia (100 μg/kg intraperitoneally), after the abdominal skin was cut to expose the small intestine in the abdominal cavity. The intestine was returned to the cavity or covered with paper and rinsed with saline. For the isolated pup intestines, 30 μL of 10 μM OT was injected at the distal end.
Blood sampling. Approximately 30 μL of blood was collected from the facial vein of neonates after quick cold anesthesia on ice ( Fig. 1a; ref. 67, 68). The similar blood volume was obtained from the heart after opening of the chest of pups (Figs 1b-d, 3a,b, 6a and 7), from the carotid artery in the pup neck (Fig. 2a) and the tail vein of adults each time (Fig. 6) under ketamine anaesthesia (100 μg/kg intraperitoneally).

Isolated intestinal preparation.
Male mice were fasted for 1 h before they were euthanized by decapitation. After the intestinal cavity was opened, the gastro-oesophageal junction and the peritoneal ligaments were cut. The mesenterial tissue was stripped from the first 20 cm of the small intestine, and a 4-cm segment was excised 1~2 cm from the pylorus. The segment isolated from PND1-7 pups was washed and placed in 3 mL of incubation buffer in 24-multiwell plates at room temperature 69 . The two openings of the segment were ligated and mounted on the flat surface of the plate. Eighty percent of the segment was dipped again in a new 3 mL of incubation buffer, and the two ends were insulated by using glycerol solution. Fifty microliters of 10 μΜ OT solution was introduced into the lumen of the segment at the distal end by using a 31 G syringe needle. The incubation medium was mixed periodically by pipetting. Fifteen microliters of incubation medium was removed and stored at −80 °C for the measurement of OT concentrations.
Enzyme immunoassay for OT and AVP. OT or AVP immunoreactivity levels were quantified by using OT and AVP enzyme immunoassay (EIA) kits (Enzo Life Sciences, NY, USA), respectively. We performed assays in tissue preparations without pre-treatment 66,70 to compare different tissue samples, though the fractioned sample was recommended 71,72 . Blood samples collected as mentioned above were stored at −80 C. The blood samples were thawed and five μL of them were diluted 1:20 in assay buffer on ice and assayed. The OT or AVP assay had a sensitivity of 5 or 11 pg/mL, and the inter-and intra-assay coefficients of variation were less than 15%.
Mass spectrometry. Standard (STD) samples of [Ile 13 C, 15 N]OT (100 ng/mL) and [Pro&Leu 13 C, 15 N]OT (100 ng/mL) were prepared by dilution of stock solutions with 0.1% TFA as necessary. The [Pro&Leu 13 C, 15 N] OT (100 ng/mL) STD (10 μL) was spiked into each mouse plasma sample (10 μL) as an internal standard (I-STD), diluted with PBS (30 μL), and then de-proteinated using 50 μL of 4% (v/v) TCA. In this sample preparation, both types of OT were diluted 10-fold. The supernatant (~80 μL) was obtained after centrifugation at 14,000 rpm for 15 min at 4 °C. This sample preparation was applied to all mouse plasma samples. The standard samples were also prepared by using the control mouse plasma. Thus, the recovery of [Ile 13 C, 15 N]OT and [Pro&Leu 13 C, 15 N]OT were estimated be similar.
RAGE expression in the intestine of wild-type mouse embryos (18.5 dpc) (4 μm sections from GenoStaff Inc, Tokyo, Japan) were examined using the polyclonal rabbit anti-RAGE (1:1000), Alexa Fluor 488, and DAPI. Images were taken with the EVOS FL Cell Imaging System (ThermoFisher Scientific, Yokohama, Japan).