Introduction

Glutamate, a nonessential amino acid, not only is a building block of proteins but has also a variety of other roles in the body. Since its amino moiety is readily exchanged with other amino acids by the corresponding aminotransferase, glutamate metabolism is linked to both the catabolism and anabolism of other amino acids (Sakai et al. 2004). Since deamination of glutamate generates α-ketoglutarate, a TCA cycle metabolite, this process is an important anaplerotic pathway. In addition, exchange of the amino moieties of glutamate and aspartate is an essential component of the malate-aspartate shuttle, which transfers NADH across mitochondrial membranes. Extracellular glutamate also has unique functions. Recently, it was suggested that glutamate is a signalling molecule also in nonneuronal tissues, since glutamate receptors are expressed in various types of cells (Du et al. 2016). To fulfil the function of glutamate as a signalling molecule, extracellular concentrations of glutamate would be maintained at low levels. Indeed, concentration deference of intracellular glutamate in intestinal mucosa from that of plasma are reportedly more than 100 (Adibi and Mercer 1973).

The postprandial concentration of circulating glutamate is also maintained at a low level despite of the dynamic changes in dietary glutamate input. Glutamate is one of the most abundant amino acids in alimentary proteins (Hou et al. 2019; Li et al. 2011); thus, glutamate intake is more than 15 g/person on average, accounting for 20% of the total amino acid intake, according to a national survey in the US (NHANES III) (Food and Nutrition Board 2005). This is not the case for other amino acids, such as branched chain amino acids (BCAAs) and alanine, since their concentrations in the blood are elevated after a meal despite their lower dietary intake than glutamate (Sakai et al. 2004). Dietary glutamate is metabolized during the first pass through the gut, which is why the circulating glutamate level is not elevated after the ingestion of glutamate (Reeds et al. 1996; Riedijk et al. 2007; Nakamura et al. 2013; Hou and Wu 2018). Dietary glutamate carbon (C) is utilized for the synthesis of alanine or lactate or catalysed to CO2 for energy production in the gut, and only a minor proportion of dietary glutamate is taken up by the blood (Reeds et al. 2000). Our study using 15N-labelled glutamate also showed that the majority of dietary glutamate is utilized in the gut as a nitrogen (N) source to synthesize alanine, aspartate, proline, and ornithine (Nakamura et al. 2017).

The gut consists of a variety of organized tissue types, such as smooth muscle, nervous system tissue, and the epithelial mucosa, and cells, such as secretory cells, immune cells, and interstitial cells. Although previous studies suggest that the intestine plays a role in maintaining glutamate homeostasis in the circulation, which intestinal cell types limit direct entry of dietary glutamate into the portal vein has not been elucidated. There are numbers of studies indicating glutamate uptake by primary culture of enterocytes and intestinal epithelial cell lines (Mordrelle et al. 2000; Broer and Fairweather 2018). Intestinal metabolic capacity for glutamate is presumably high, since glutamine is known to be one of major energy source for these cells through its catabolism to glutamate (Fleming et al. 1991; Wu et al. 1995). Thus, it is believed that intestinal epithelial cell layer metabolizes dietary glutamate and limits its direct entry into the circulation (Blachier et al. 2009; Wu 1998). However, direct evidences are limited. Although Oba et al. have shown that enterocytes metabolized glutamate to CO2 and lactate, their study did not quantify either entire fate of glutamate taken up by the cells or its permeability across these cells (Oba et al. 2004). Nicklin et al. have evaluated glutamate transport across intestinal cell layer in in vitro model system using radiolabelled glutamate by measuring radioactivity (Nicklin et al. 1995). However, they evaluated neither glutamate metabolism in the epithelial cells nor its transport itself. Therefore, the present study aimed to quantify glutamate transport across intestinal monolayer epithelial cells in addition to its metabolism during the first pass.

Caco-2 cells, a human intestinal cell line derived from colon cancer, have been shown to differentiate on the membrane filter in transwell culture into cells that exhibit both morphological and biochemical features characteristic of intestinal epithelial cells. The cells develop mucosal villi on the apical membrane and intercellular gap junctions (Hidalgo et al. 1989). An apparent polarity in transport velocity has also been shown in this model (Hidalgo and Borchardt 1990a, b). Thus, Caco-2 cells cultured in transwells are often used to assess intestinal transport and the permeability of the intestine to amino acids (Hidalgo and Borchardt 1990a; Nicklin et al. 1995; Thwaites et al. 1996, 1994; Satsu et al. 2009) as well as to food components (Steensma et al. 2004; Yasuda et al. 2015) and drugs (Artursson and Karlsson 1991). In the present study, we investigated glutamate transport and fates of glutamate-C and N in this intestinal epithelial cell layer model using [U-13C] and [15N]glutamate. Furthermore, we compared glutamate transport across the epithelial cell layer with those of other amino acids. This study demonstrates that monolayer intestinal epithelial cells metabolized almost all the glutamate taken up from the apical medium and limited its entry into the basolateral side.

Materials and methods

Cell culture

Caco-2 cells were purchased from the European Collection of Authenticated Cell Culture (ECACC; Salisbury, UK). The cells were seeded on polycarbonate cell culture inserts (pore size, 0.4 μm; diameter, 24 mm) in a 12-well transwell plate (Corning; NY, USA). The cells were grown and differentiated as described previously (Yasuda et al. 2015) in E-MEM (FUJIFILM Wako, Osaka, Japan) supplemented with 10% heat-inactivated FBS, penicillin (100 U/mL), streptomycin (100 ng/mL), and Gibco™ MEM nonessential amino acid solution (final concentrations of glycine, L-alanine, L-asparagine, L-aspartate, L-glutamate, L-proline and L-serine: 100 μM each) at 37 °C under a humidified atmosphere of 95% air and 5% CO2 for 21 days. Apical and basolateral media (0.5 and 1.5 mL, respectively) were exchanged twice a week. The transepithelial electrical resistance (TEER) value of the monolayers was monitored during culture using a Millicell-ERS instrument (Millipore Corp., Bedford, MA, USA). After confirming epithelial cell differentiation by the TEER value (> 400 Ω cm2), the epithelial layer model was used for experiments.

Experimental procedures

Basolateral and apical media were refreshed, and apical media were supplemented with 300 μM [U-13C] or [15N]glutamate in experiments 1 and 2, respectively. The cells were cultured for 1, 3, 6, 10, and 24 h, and apical and basolateral media were collected. In experiment 3, apical media were supplemented with 1) 400 µM [1-13C]leucine, 400 µM [1-13C]isoleucine, 400 µM [1-13C]valine, 400 µM [1-13C]lysine hydrochloride, 200 µM [1−13C]phenylalanine, 50 µM [1-13C]tryptophan and 2000 µM [U-13C]glutamine; 2) 200 µM [ring-2-13C]histidine hydrochloride monohydrate, 200 µM [1-13C]serine, 200 µM [1-13C]proline, 400 µM [U-13C]threonine, 100 µM [1-13C]methionine and 600 µM [guanido-13C2]arginine hydrochloride; 3) 200 µM [15N]aspartate; 4) 200 µM [15N2]asparagine monohydrate; 5) 200 µM [1-13C]glycine, 200 µM [ring-3,5-D2]tyrosine and 400 µM [U-13C]cysteine; or 6) 200 µM [2,3-13C2]alanine and 200 µM [1-13C]glutamate at the time of medium change. The culture media were harvested after 24 h of culture and stored at –80 °C before analysis. All the amino acids labelled with stable isotopes were purchased from Cambridge Isotope Laboratories (Tewksbury, MA, USA).

Amino acid analysis

The concentrations of amino acids in the media were determined using an amino acid analyser (L-8000, Hitachi, Tokyo, Japan) after deproteinization with 5% trichloroacetic acid. Briefly, amino acids separated by cation-exchange chromatography were detected spectrophotometrically after post-column reaction with ninhydrin reagent.

The isotopic enrichment of amino acids and ammonia in the media was measured by liquid chromatography–tandem mass spectrometry (LC–MS/MS; Agilent 1200 Series LC System and a 6400 Series Triple Quad LC/MS; Agilent Technologies, Santa Clara, CA, USA) after derivatization as reported previously (Nakamura et al. 2015). Briefly, samples of media were deproteinized with methanol, derivatized with 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate (AQC; Waters Corporation, MA, USA), and then injected into a high-performance LC (HPLC) column (L-column ODS; Chemicals Evaluation and Research Institute, Japan). Positive ions generated by electrospray ionization (ESI) were analysed in multiple reaction monitoring (MRM) mode. The following precursor ions (m/z) were used to detect individual amino acids and ammonia: glutamate: 318, glutamine: 317, aspartate: 304, asparagine: 303, alanine: 260, proline: 286, serine: 276, glycine: 246, arginine: 173, ornithine: 473, leucine and isoleucine: 302, valine: 288, phenylalanine: 336, tyrosine: 352, tryptophan: 375, methionine: 320, cystine: 291, threonine: 290, lysine: 487, histidine: 326 and ammonia: 188. To detect their isotopomers, the m/z values were increased by the corresponding numbers. The collision energy was set to 30 eV for ornithine, 10 eV for arginine and 25 eV for other amino acids, and a common ion (m/z 171) generated from AQC derivatives was detected as a product ion. Regarding [U-13C]cysteine, M3 and M6 isotopomers of cystine were measured, since isotope enrichments in cysteine and cystine are considered to be equal due to their rapid exchange in physiological solutions through thiol-disulfide reactions (Lash and Jones 1985).

Calculation

The isotopic enrichment (IE: molecular percent excess) of each amino acid was calculated from the tracer/tracee ratio (TTR) as follows:

$${\text{IE}} \left( {{\text{mol}}\% } \right) = {\text{TTR}}(1 + {\text{TTR}})$$

When multiple 13C-labels appeared, the tracer/tracee ratio and the isotopic enrichment of the isotopomer with an additional molecular weight of i (TTRi and IEi, respectively) were calculated as follows:

$${\text{TTR}}_{i} \left( {{\text{mol\% }}} \right) = \left( {M_{i} /M_{0} } \right)_{{{\text{sample}}}} - \left( {M_{i} /M_{0} } \right)_{{{\text{back}}}} - \left( {{\text{TTR}}_{i - 1} \times \left( {M_{1} /M_{0} } \right)_{{{\text{back}}}} - {\text{TTR}}_{i - 2} \times \left( {M_{2} /M_{0} } \right)_{{{\text{back}}}} - \cdots } \right)$$
$${\text{IE}}_{i} ({\text{mol}}\% ) = {\text{TTR}}_{i} /(1 + \sum_{(x = 0)}^{n} {\text{TTR}}_{x} ).$$

Here, Mi is the abundance of the isotopomer with an additional molecular weight of i. Thus, (Mi/M0)sample and (Mi/M0)back are the relative abundance of the amino acid in the sample and background, respectively. n is the number of carbons in the amino acid. TTRx is the tracer/tracee ratio of the isotopomer with an additional molecular weight of x. The average 13C enrichment of all the carbons in the amino acids (AIE) was calculated as reported (Nakamura et al. 2013) using the following equation:

$${\text{AIE}} \left( {{\text{mol\% }}} \right) = \mathop \sum \limits_{i = 1}^{n} ({\text{MPE}}_{i} \times i)/n.$$

Then, the total 13C content (C13C) of the amino acid in each medium was calculated as follows:

$$C_{13C} \left( {{\text{nmol}}} \right) = {\text{AIE}} \times C \times V.$$

Here, C is the amino acid concentration (µM) in the medium, and V is the volume of the medium (mL). The 15N content of each amino acid and ammonia was calculated in a similar manner with the following equation:

$$C_{15N} \left( {{\text{nmol}}} \right) = {\text{IE}} \times C \times V.$$

Here, IE is the 15N-enrichment (mol%) of each amino acid.

Uptake (nmol) of each labelled amino acid by the cells from the apical medium was estimated as the difference in the labelled amino acid content in the apical medium before and after culture. The transport (TA-B, % of Up) of the labelled amino acid was calculated as follows:

$$T_{A - B} \left( \% \right) = {\text{IE}} \times C \times V/Up.$$

In this case, IE and C are isotopic enrichment (mol%) and concentration (μM) of a target amino acid in the basolateral medium. V is volume (mL) of basolateral medium and Up is uptake of the target amino acid from the apical medium (nmol). Fractional distribution (% of Up) of each label (i.e., 13C or 15N) into both apical and basolateral amino acids are calculated similarly.

Statistics

Differences in concentration before and after culture and on the apical and basolateral sides were analysed by Sidak’s multiple comparisons test after 1-way analysis of Variance (ANOVA). Changes from baseline were analysed by Dunnett's multiple comparisons test. Differences between apical and basolateral were analysed by paired student’s t test. Amino acid uptake from the apical medium and transport from apical to basolateral compartment were analysed after logarithmic transformation, since there were more than 100 folds variations among the groups. Then statistical differences from glutamate were analysed by Dunnett’s test after one-way ANOVA. A probability less than 0.05 indicated statistical significance. All the data are expressed as the means ± standard deviations (SDs).

Results

Amino acid concentrations in the apical and basolateral media (Exp. 1)

Concentrations of individual amino acids in the apical and basolateral media changed to different extents over 24 h of culture (Fig. 1a, b; Table 1). In the apical medium, the concentrations of glutamate, aspartate, lysine and arginine decreased significantly, while the alanine and leucine concentrations increased during culture. In addition, changes in the basolateral medium were different from those in the apical medium. For example, there was no decrease in the glutamate, aspartate, lysine or arginine concentrations in the basolateral medium. This caused the apical/basolateral concentration ratios for glutamate, lysine and arginine, leucine and alanine to be significantly different from 1.0 (Fig. 1c), which indicates the polarity of Caco-2 cells in transwell culture in the metabolism of these amino acids.

Fig. 1
figure 1

Amino acid concentrations in the apical and basolateral media (Exp. 1). a and b: Amino acid concentrations in the apical and basolateral media after 24 h of culture. Amino acid concentrations are expressed as percentages of those before culture. c Ratios of amino acid concentrations in the apical and basolateral media after 24 h of culture. All data are expressed as the means ± SDs (N = 4). *: p < 0.01, **: p < 0.001 before vs. after culture in a and b and apical vs. basolateral media in c (Sidak’s test)

Full size image
Table 1 Amino acid concentrations before and during the culture in Exp. 1
Full size table

Fate of glutamate-C (Exp. 1)

The glutamate concentration on the apical side decreased in a time-dependent manner (Fig. 2a; Table 1), which was accompanied by a time-dependent decrease in 13C content of apical glutamate (Fig. 2b). At the end of the culture period, 13C-glutamate in the apical medium was almost depleted (Fig. 2b), indicating that almost all the [U-13C]glutamate added to the apical medium had been taken up by the cells. However, the basolateral concentration of glutamate was maintained during culture (Fig. 2a; Table 1), and the [U-13C]glutamate in the basolateral medium after 24 h of culture was only 1.1 ± 0.7% of the amount added.

Fig. 2
figure 2

Time-dependent changes in the concentrations, U-13C enrichment and 13C content of glutamate in the apical and basolateral media (Exp. 1). Concentrations (a) and total 13C content (b) of glutamate in the apical (closed circle) and basolateral (open circle) media during the 24 h of culture were plotted. The 13C content is expressed as a percentage of the amount added (means ± SDs (N = 4)). *: p < 0.05, **: p < 0.01, ***: p < 0.001 vs. before culture (Dunnett’s test)

Full size image

To trace the fate of 13C originating from [U-13C]glutamate, the 13C contents in other amino acids in the media were measured. Significant 13C increases over the natural abundances were detected in proline, alanine and ornithine, and the contents elevated steadily during culture (Fig. 3a; Table 2). After 24 h of culture, the recovered 13C contents in alanine, proline, ornithine and aspartate were 3.1%, 4.3%, 1.7% and 0.1%, respectively, of the 13C taken up (Fig. 3b). Although 3.9% of the 13C was released into the basolateral medium in the form of glutamate, the fraction released directly (i.e., 13C released in the form of [U-13C]glutamate) was less than 30% of the total 13C in the basolateral glutamate, since other isotopomers such as M1, M2, M3 were also released into the basolateral medium (Table 2). The fate of the majority of the 13C (87%) was not identified, presumably due to its metabolism to CO2 and other metabolites, such as glutamine and nucleic acids.

Fig. 3
figure 3

Fate of glutamate-C (Exp. 1). a The combined total 13C content in proline (open circle), alanine (closed circle) and ornithine (open square) in the apical and basolateral media during the 24 h of culture. The values are expressed as a percentage of the amount added (means ± SDs (N = 4)). *: p < 0.05, ***: p < 0.001 vs. before culture (Dunnett’s test). b Fractional distribution of glutamate-C (% of uptake of glutamate-C by the cells from the apical medium) after 24 h of culture. “Others” indicates the unidentified fraction and “Passed” indicates 13C released into the basolateral medium in the form of glutamate. Glutamine was not analysed in Exp. 1

Full size image
Table 2 Isotopic enrichments of apical and basolateral amino acids before and during the culture in Exp. 1
Full size table

Fate of glutamate-N (Exp. 2)

Glutamate concentrations on the apical side decreased in a time-dependent manner, as was observed in Exp. 1 (Fig. 4a; Table 3), which was accompanied by a time-dependent decrease in 15N-glutamate content (Fig. 4b). After 10 h of culture, the [15N]glutamate in the apical medium was almost depleted (Fig. 4b). However, the basolateral concentration of glutamate did not change during the culture period (Fig. 4a and Table 3), and only a minor proportion of the [15N]glutamate (2.3% of the amount added) was found in the basolateral medium after 24 h of culture.

Fig. 4
figure 4

Time-dependent changes in the concentrations, 15N enrichment and 15N content of glutamate in the apical and basolateral media (Exp. 2). Glutamate concentrations (a) and 15N content (b) in glutamate in the apical (closed circle) and basolateral (open circle) media during the 24 h of culture. The 15N content is expressed as a percentage of the amount added (means ± SDs (N = 4)). *: p < 0.05, ***: p < 0.001 vs. before culture (Dunnett’s test)

Full size image
Table 3 Amino acid concentrations before and during the culture in Exp. 2
Full size table

To trace the fate of glutamate-15N added to the apical medium, 15N-labelled amino acids in the medium were analysed (Table 4). The 15N content in alanine, glutamine, ornithine, proline, BCAAs and ammonia increased in a time-dependent manner until 10 h after [15N] glutamate had been added to the apical medium (Fig. 5a). Then, the 15N content in these amino acids remained almost constant until the end (24 h) of the culture period. After 24 h of culture, the recovered 15N contents in alanine, glutamine, ornithine, proline, BCAAs, asparagine, serine and ammonia were 14.4%, 7.7%, 7.0%, 4.0%, 5.6%, 0.3%, 0.5% and 18.1%, respectively, of the glutamate 15N taken up by the cells, and only a minor part (2.3%) of the 15N was released into the basolateral medium in the form of glutamate (Fig. 5b). The fate of approximately 40% of the 15N was not identified, presumably due to the volatilization of 15N-ammonia and utilization of the 15N for the synthesis of other metabolites.

Table 4 15N-enrichments of apical and basolateral amino acids before and during the culture in Exp. 2
Full size table
Fig. 5
figure 5

Fate of glutamate-N (Exp. 2). a The combined total 15N content in proline (open circle), alanine (closed circle) and ornithine (open square) in the apical and basolateral media during the 24 h of culture. The values are expressed as a percentage of the amount added (means ± SDs (N = 4)). The increase in the content of each 15N-amino acid after culture was significant (p < 0.0001) (Dunnett’s test). b: Fractional distribution of glutamate-N (% of uptake of glutamate-N by the cells from the apical medium) after 24 h of culture. “Others” indicates the unidentified fraction and “Passed” indicates 15N released into the basolateral medium in the form of glutamate

Full size image

Permeability of the cell layer to other amino acids and their metabolism (Exp. 3)

Twenty amino acids labelled with stable isotopes (13C, 15N or D) were added to apical media, and their metabolism was compared with that of glutamate. Amino acid concentrations and enrichments of each tracers supplemented are shown in Tables 5 and 6. The uptake of each amino acid by the Caco-2 cells from the apical medium was estimated from the decrease in the labelled amino acid content in the apical medium. While almost 100% of the added glutamate and aspartate was taken up by the cells, significantly less of the other amino acids was taken up (Fig. 6a). However, virtually no glutamate or aspartate crossed the Caco-2 cell layer into the basolateral medium, while other amino acids taken up by the cells were transferred substantially across the basolateral membrane into the medium (Fig. 6b).

Table 5 Amino acid concentrations before and after the culture in Exp. 3
Full size table
Table 6 Isotopic enrichments of the tracer amino acid before and after the culture in Exp. 3
Full size table
Fig. 6
figure 6

Comparison of uptake and transport of 20 kinds of amino acids (Exp. 3). a Uptake of individual amino acids labelled with stable isotopes (13C, 15N or D) from the apical medium was calculated as the difference between the amount added and the content remaining in the apical medium after 24 h of culture and is expressed as a percent of the amount added. b Apical-to-basolateral transport (A-P transport) of each amino acid from apical to basolateral compartment across the Caco-2 cell layer was estimated from the content of each labelled amino acid in the basolateral medium after 24 h of culture and is expressed as the percent of the uptake (means ± SDs (N = 4)). **: p < 0.001 vs. glutamate (Dunnett’s test)

Full size image

The fates of glutamate-C ([1-13C]glutamate), aspartate-N ([15N]aspartic acid), asparagine-N ([15N2]asparagine), glutamine-C ([U-13C]glutamine), alanine-C ([2,3-13C2]alanine), proline-C ([1-13C]proline), serine-C ([1-13C]serine) and glycine-C ([1-13C]glycine) added to the apical medium were estimated. The distributions of 13C and 15N of these amino acids differed (Table 7). Although glutamate-C and glutamine-C were distributed to glutamine, ornithine and proline, the transport of glutamine-C from apical to basolateral compartment was much higher than that of glutamate-C (Fig. 7a and b). Both aspartate-N and asparagine-N were distributed to alanine, glutamine, ornithine, BCAAs and ammonia (Fig. 7c and d). However, apical-to-basolateral transport of asparagine-N was much higher than that of aspartate-N. Alanine-C was distributed to glutamine, ornithine and proline, while proline-C was distributed to ornithine (Fig. 7e and f). Serine-C was detected in glycine, while glycine-C was detected in serine (Fig. 7g and h).

Table 7 Distribution of Glu-C, Gln-C, Asp-N, Asn-N, Ala-C, Pro-C, Ser-C and Gly-C into apical and basolateral amino acids in Exp. 3
Full size table
Fig. 7
figure 7

Fates of other amino acids-C and N (Exp. 3). Fractional distributions (% of uptake from the apical media) of glutamate-C (a), glutamine-C (b), aspartate-N (c), asparagine-N (d), alanine-C (e), proline-C (f), serine-C (g) and glycine-C (h) after 24 h of culture were estimated from concentrations and isotopic enrichments of amino acids in apical and basolateral media in Exp. 3 as described in the Materials and Methods section. “Others” indicates unidentified fractions, and “Passed” indicates fractions released into the basolateral medium in the form added

Full size image

Discussion

We studied the metabolism of glutamate in transwell cultures of Caco-2 cells, a model for an intestinal epithelial cell layer, and revealed its barrier function against apical-to-basolateral entry of glutamate. Although almost all glutamate added to the apical medium disappeared and had been incorporated into the epithelial cell layer, it barely appeared in the basolateral medium, independent of kinds of tracers used. The results indicate that virtually all the glutamate taken up from the apical medium was metabolized in the cells. Indeed, labelled C and N atoms, derived from glutamate were found in other amino acids in the medium. Glutamate-C was shown to be incorporated into alanine, glutamine, proline, ornithine and aspartate, while glutamate-N was incorporated into alanine, glutamine, proline, ornithine, BCAAs and ammonia. In vivo studies using [U-13C] and [15N]glutamate have also shown that virtually all dietary glutamate is metabolized in the gut to alanine, proline, ornithine, and ammonia by measuring the portal-arterial balance of 13C and 15N (Reeds et al. 1996, 2000; van der Schoor et al. 2001; Nakamura et al. 2013). However, these in vivo studies did not identify the intestinal cell type that contributes to the metabolism of dietary glutamate, although various cell types exist in the gut. Since the present study investigated the metabolism of glutamate only in epithelial cells, the results strongly suggest the contribution of the epithelial cell monolayer to this unique intestinal metabolism of dietary glutamate.

Although apical and/or dietary glutamate-C and N were metabolized to similar amino acids and metabolites both in epithelial cells in vitro and in the gut in vivo (Reeds et al. 2000; Nakamura et al. 2013), their quantitative distributions differed. In vivo studies indicated that approximately half of dietary glutamate-C was metabolized to CO2, while the present in vitro study did not investigate its metabolism to CO2, since the cells were cultured in a CO2 incubator. However, the following results provide appropriate evidence that substantial glutamate-C was also metabolized to CO2 in the present in vitro study. First, the utilization of glutamate-N for the synthesis of ammonia, alanine, and BCAAs indicates the generation of α-ketoglutarate, a TCA cycle metabolite, suggesting the utilization of glutamate-C to generate CO2 via TCA cycle metabolism. Indeed, multiple 13C-labels of glutamate were generated from [U-13C]glutamate (Table 2), indicating the metabolism of glutamate-C via the TCA cycle. Second, the utilization of glutamate-C for the synthesis of alanine and aspartate also indicates the metabolism of glutamate-derived α-ketoglutarate to oxaloacetic acid via the TCA cycle. Glutamate-N metabolism also differed quantitatively in these studies: a large proportion of glutamate-N was metabolized to ammonia in this in vitro study, while it was utilized mainly for the production of alanine in vivo (Nakamura et al. 2013). This difference might be due to the features of the Caco-2 cells used in the present study, since the cells were originally derived from the colon cancer. The expression of ornithine transcarbamylase, an enzyme that uses ammonia via carbamoyl phosphate for the synthesis of citrulline, was reported to be lower in the colon than in the small intestine (Hamano et al. 1988), which might have increased ammonia production in the present study. In addition, the expression of alanine aminotransferase, which produces alanine from glutamate and pyruvic acid, is also lower in the colon than in the small intestine (Yang et al. 2009), which might have decreased alanine production in the present in vitro study. This difference might also be due to the secondary metabolism of glutamate-C and N after their metabolism in the epithelial cell layer. Although the fates of dietary and/or apical glutamate-C and N were quantitatively different between in vitro and in vivo studies, the qualitative similarity in their fates suggests the substantial impact of intestinal epithelial cells on the metabolism of dietary glutamate.

Glutamate metabolism in the intestinal epithelial cell layer model is unique compared to other amino acids. The cell layer passed hardly any glutamate to the basolateral side, as mentioned above. However, the cell layer transported significant amounts of other amino acids to the basolateral side, except in the case of aspartate, although the uptake of the other amino acids by the layer was less than that of glutamate. Similar differences in the transport of amino acids were found in the gut in vivo. Postprandial arterial-portal differences in amino acids in piglets and rats indicated that the net uptake of glutamate by the portal vein was much less than that of other amino acids (Reeds et al. 1996; Nakamura et al. 2017). Although the net uptake of glutamine was negative in these studies, a study using both [2-15N] and [5-15N]glutamine indicated the entry of significant dietary glutamine into the circulation (Nakamura et al. 2013). Thus, the present results indicate the impact of intestinal epithelial cells on the metabolism of dietary amino acids and the regulation of amino acid entry in the circulation. Indeed, the fates of the apical amino acids-C and N were similar to those in vivo. For example, aspartate-N was distributed in alanine, glutamate, glutamine, ornithine, and ammonia in both studies. Bidirectional metabolism between serine and glycine was also evident in both studies.

In all the experiment in the present study, enrichment of labelled amino acid in apical media were decreased during the culture (Tables 2, 4, 6). The decreases of the enrichments indicate isotope dilution due to entry of non-labelled amino acids from the cells into the apical media. Thus, present study suggests that bidirectional transports of amino acids take place in this intestinal epithelial cell layer model. Detailed analysis of bidirectional transport in various experimental setting, for example, at different amino acid concentrations, might provide important information to understand functions and cooperation of multiple transporters expressed in intestinal epithelial cells (Broer and Fairweather 2018). Although transwell culture of Caco-2 cells are widely used as an intestinal epithelial cell layer model because of its similarities in morphology, functional characteristics and expressions of enzymes and growth factor receptors (Hidalgo et al. 1989; Delie and Rubas 1997), the present study has limitations. Since Caco-2 cell is originally derived from colon cancer, there might be differences in amino acid metabolisms in the model from those in actual epithelial cells in small intestines, a major absorption site for amino acids. Further study using intestinal epithelial monolayer model of other cell lines and primary enterocytes will be required in future.

In conclusion, the present study demonstrated that the Caco-2 intestinal epithelial monolayer metabolized almost all the apical glutamate and limit the apical-to-basolateral transport of glutamate. Glutamate concentrations in the circulation are maintained despite the dynamic changes in dietary input of glutamate. The present study suggests that the intestinal epithelial cells contribute to the maintenance of glutamate homeostasis in the systemic circulation at least partially.