Metabolic Analysis of Three-Dimensional Cultured Gastrointestinal Cancer Cells Suggests that l-Arginine Inhibits Tumor Growth by Affecting the Urea Cycle

The semi-essential amino acid l-arginine (arginine) is synthesized in the body and is involved in the production of various substances, such as nitric oxide (NO) and urea; it has diverse roles in cellular processes and has been suggested to exert anticancer effects [1, 2]. In animal models, arginine administered as a nutritional supplement suppresses the development and growth of various cancers, including colorectal cancer and squamous cell carcinoma of the skin [3‒5]. Recent evidence suggests that arginine supplementation enhances the effectiveness of chemoimmunotherapy [6]. Clinical studies have demonstrated that arginine exerts anticancer effects and improves survival rates in patients with head and neck cancer or colon cancer [7‒9]. However, in vitro analyses of the effects of arginine on cancer cells are limited. Furthermore, in vivo studies have shown that in addition to cellular-level changes, several factors, such as metabolic processes and the immune system, contribute to the observed effects, making it challenging to pinpoint specific mechanisms. To investigate the direct effects of arginine on cancer cells, herein, we utilized the conventional two-dimensional (2D) culture method as well as our recently developed three-dimensional (3D) culture method, called the “tissueoid cell culture system” [10]. This 3D culture method allows the studying of cells in vitro under microenvironments that closely resemble in vivo conditions [11, 12]. In this study, we used six human upper gastrointestinal cancer cell lines and compared the effects of arginine on proliferation using the MTT assay. We selected cell lines with different sensitivities to arginine and performed a histological evaluation, mitochondrial function analysis, cell cycle assay, comprehensive metabolic analysis, and tracer study using (¹³C₆) l-arginine to explore the mechanism underlying the anticancer effects of arginine.

Six human upper gastrointestinal cancer cell lines (OE19, OE33, MKN1, MKN45, MKN74, and AGS) were used. The characteristics, sources, and culture methods are summarized in Table 1 [13‒16]. For the 3D culture system, previously reported methods were used [17]. Cellbed^® (Japan Vilene Co., Tokyo, Japan) was used as the culture carrier in the “tissueoid cell culture system” [10].

In the 2D and 3D culture systems, OE19, OE33, MKN1, MKN45, MKN74, and AGS cells were seeded at a density of 3 × 10³ cells/well in a 96-well plate, with Cellbed used in the 3D culture system. Two groups were created: one cultured in regular medium and another cultured in medium supplemented with 5 mmol/L or 10 mmol/L l(+)-arginine hydrochloride (FUJIFILM Wako Pure Chemical Corp., Osaka, Japan) at seeding (n = 4). Because previous literature reported arginine concentrations in serum 9.25 ± 0.39 mmol/L in the study of arginine administration to adult males [18], we decided on a maximum concentration of 10 mmol/L of arginine administration. The medium was replaced with the same concentration of supplemented medium. Measurements were conducted on days 2, 4, 6, and 7 for the 2D culture system and on days 2, 4, 7, 9, 11, and 14 for the 3D culture system using the MTT Assay Kit (Nacalai Tesque, Kyoto, Japan). In the 2D culture system, assays were performed according to the kit protocol. In the 3D culture system, after incubation with the MTT reagent, each well was replaced with 200 μL of DMSO and dissolved overnight at 37°C; 100 µL was taken and used for measurements. If the volume of liquid decreased during the experiment due to evaporation and correct measurements could not be made, the sample was excluded. Proliferation was evaluated using the MTT assay and OE33 (arginine-sensitive), in which proliferation was inhibited substantially, and OE19, in which effects on proliferation were weak (arginine-insensitive) and were used for subsequent experiments. The results of the MTT assay are described in detail in the Results section.

For comparison, the effects of d-arginine (PEPTIDE INSTITUTE, INC., Osaka, Japan), a stereoisomer of l-arginine, on cell growth were evaluated. OE33 was cultured in the 2D culture system. One group was cultured in regular medium, while another group was cultured in medium supplemented with 5 or 10 mmol/L l-arginine or d-arginine (n = 6). The cultures underwent medium exchange with the same concentration of supplemented medium. Measurement was conducted on day 4 using the MTT assay kit.

We also evaluated cell growth using the Calcein assay, which is not affected by mitochondrial function. The seeding and culture procedures were performed in accordance with the MTT assay, and cell viability was evaluated using the Cellstain-Calcein-AM solution (DOJINDO LABORATORIES, Kumamoto, Japan) (FUJIFILM Wako Pure Chemical Corp.) (n = 5). Each well was washed once with 100 μL of PBS(+), and 100 μL of PBS(+) supplemented with Cell Counting Kit-F solution at 500-fold dilution was added. The reaction was carried out for 25 min. Measurements were performed using an Infinite 200 PRO (Tecan Trading AG, Switzerland) at an excitation wavelength of 490 nm and a fluorescence wavelength of 520 nm. Measurements were obtained on days 2, 4, and 6 for the 2D culture system and on days 2, 5, 8, 11, and 14 for the 3D culture system.

In the 2D culture system, a mitochondrial stress test was conducted using the Seahorse XF24e Extracellular Flux Analyzer (Agilent Technologies, Santa Clara, CA, USA) to measure the oxygen consumption rate (OCR). Cells were seeded in Seahorse XF24 cell culture plates (Agilent Technologies), with OE19 at 2 × 10⁴ cells/well and OE33 at 5 × 10³ cells/well. The plates were then cultured at 37°C with 5% CO₂. On day 3, the medium in the untreated group was replaced with regular medium, while the medium in the treatment group was replaced with medium supplemented with 10 mmol/L arginine before culturing for 24 h (n = 10). Measurements were taken using the Seahorse XF Cell Mito Stress Test Kit (Agilent Technologies), following the manufacturer’s instructions. Inhibitors were added by adjusting the final concentration with oligomycin (1.5 µmol/L), carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP) (1 µmol/L), and rotenone and antimycin A (0.5 µmol/L). To account for differences in cell numbers between samples, the protein level in the samples was estimated after measurement using the bicinchoninic acid (BCA) assay, before correcting the OCR values.

Cells were seeded at a density of 1 × 10⁵ cells/well for OE19 and 5 × 10⁴ cells/well for OE33 on 12-well plates containing circular Cellbed with a diameter of 19-mm. After 14 days of 3D culture, the medium was replaced with a regular medium for the untreated group and supplemented with 10 mmol/L arginine for the treatment group before culturing for an additional 24 h. As described previously [17], paraffin blocks were prepared. Sections with a thickness of 6 μm were cut and subjected to either H&E staining or immunohistochemical staining using an anti-Ki-67 (30-9) rabbit monoclonal antibody (F30644, Roche Diagnostics, Basel, Switzerland). Immunohistochemical staining was performed using the DISCOVERY ULTRA Automated IHC Stainer with a Ventana DABMap Detection Kit (Ventana Medical System, Tucson, AZ, USA) following previously described methods [10].

OE19 and OE33 were cultured with/without arginine (10 mmol/L)-containing medium for 24 h. After treatment with 5-Bromo-2’-Deoxyuridine (BrdU) (Sigma-Aldrich, St. Louis, MO, USA) (10 μmol/L) for the last hour of 24 h culture, the cells were corrected and fixed with 70% ethanol. Then, samples were treated with 2 N HCl/0.5% Triton X-100 and neutralized with 0.1 mol/L sodium tetraborate to obtain single-strand cellular DNA. Cells were permeabilized with PBS containing 1% BSA and 0.5% Tween 20. The BrdU incorporated in the DNA was detected using FITC Mouse Anti-BrdU (FITC Mouse Anti-BrdU Set 556028, BD Biosciences, Franklin Lakes, NJ, USA). The cells were also stained with propidium iodide (FUJIFILM Wako Pure Chemical Corp). Data were acquired using the BD LSRFortessa^™ X-20 Cell Analyzer (BD Biosciences) and analyzed using FlowJo ver.10.10.0 (BD Biosciences). Three similar experiments were performed, and average values were calculated.

Cells were seeded at a density of 1 × 10⁵ cells per well for OE19 and 5 × 10⁴ cells per well for OE33 in 12-well plates containing circular Cellbed with a diameter of 19-mm. After 14 days of 3D culture, the medium for the untreated group was replaced with regular medium and that for the treated group was replaced with medium supplemented with 10 mmol/L arginine. Cells were then cultured for an additional 24 h (n = 3). Samples were analyzed by Human Metabolome Technologies, Inc. (Tsuruoka, Japan) following previously described methods [11, 12]. MTT assays were performed using samples prepared in the same manner as the submitted samples. Cell counts were estimated using a calibration curve created from samples in which cell numbers were measured, and the values were corrected accordingly.

Arginine-sensitive OE33 was seeded at a density of 1 × 10⁴ cells per well in a 3D culture system and cultured for 14 days as described for the comprehensive metabolomic analysis. The untreated group was incubated in regular medium, while the treated group was incubated in medium supplemented with 10 mmol/L l-arginine-HCl (¹³C₆, 99%) CLM-2265-H-0.25 (Cambridge Isotope Laboratories, Inc., Tewksbury, MA, USA) for 24 h (n = 3). Twenty-four metabolites containing isotope-labeled substances were analyzed by Human Metabolome Technologies, Inc. Correction by cell number was performed following the methods used for the comprehensive metabolomic analysis.

The Student’s t test was used to evaluate the results of the MTT assay, calcein assay, extracellular flux assay, and cell cycle assay. Welch’s t test was utilized to evaluate the results of the metabolomic analysis (*p < 0.05, **p < 0.01, ***p < 0.001). In the tracer study, all p values were corrected for multiple hypothesis testing using the Benjamini-Hochberg method. Error bars in each figure represent the standard deviation.

In the 2D culture system, cell proliferation was significantly lower in the 5 mmol/L and 10 mmol/L arginine-treated groups than in the untreated group for OE33, MKN1, MKN45, and MKN74 cell lines (Fig. 1a; online suppl. Table S1a; for all online suppl. material, see https://doi.org/10.1159/000543006). For OE19, proliferation was significantly lower than that of the untreated group only in the 10 mmol/L arginine group on day 4, with no significant differences at other time points. No significant inhibition of proliferation was observed for the AGS cell line. In the 3D culture system, the rate of proliferation was significantly lower in the 5 and 10 mmol/L arginine-treated groups than in the untreated group for OE19, OE33, MKN45, and MKN74 cell lines as well as in the 10 mmol/L arginine-treated group for the MKN1 cell line (Fig. 1b; online suppl. Table S1b). For AGS, proliferation rates were significantly lower in the 10 mmol/L arginine-treated group than in the control group on day 9, while no significant differences were observed at other time points. In the MTT assay, the effects of arginine on cancer cell proliferation were evaluated under 2D culture conditions until day 7. Unlike OE19, MKN1, MKN74, and OE33, where the effects of arginine on cancer cell proliferation plateaued by day 6, it is likely that under 2D culture conditions, cancer cell proliferation may decrease due to various factors, such as contact inhibition when reaching a confluent state. Therefore, based on the results up to day 6, we conducted further experiments using arginine-sensitive OE33 and arginine-insensitive OE19.

The effects of d-arginine, a stereoisomer of l-arginine, on cell growth were evaluated for comparison. There were no significant differences between the l-arginine-treated and -untreated groups (shown in online suppl. Fig. S1). The results of a calcein assay were similar to those of the MTT assay. In the 2D culture system, proliferation activity was significantly lower in the treated group than in the untreated group for OE33, while no significant differences for OE19 were observed (shown in online suppl. Fig. S2a). In the 3D culture system, proliferation was significantly lower in the treated groups than in the control groups for OE33 and OE19 (online suppl. Fig. S2b).

In the arginine-treated group using OE33, the maximal respiratory capacity was significantly lower than that in the untreated group (p = 0.00001) (Fig. 2). However, there were no significant differences in basal respiration. For the OE19 cell line, no significant differences were observed between the treated and untreated groups in terms of basal respiration or maximal respiratory capacity.

Figure 3 shows representative horizontal and vertical sections after H&E staining and immunohistochemical staining for Ki-67 at 24 h post-arginine treatment for both 3D-cultured OE19 and OE33 cells. In both OE19 and OE33, there were no evident histological differences between the arginine-treated and -untreated groups at 24 h post-treatment. There were no significant differences in the Ki-67 labeling index between the groups (data not shown).

In arginine-sensitive OE33 cells, the proportion of BrdU-incorporated cells (representing S phase cells) was significantly lower (p = 0.033) and cells in G0/G1 phase were higher (p = 0.026) in arginine-treated group than in the untreated group (Fig. 4; Table 2). In arginine-insensitive OE19, significantly fewer cells were in G2/M phase in the arginine-treated group than in the untreated group (p = 0.032).

In a principal component (PC) analysis, differences in metabolic profiles between arginine-insensitive OE19 and arginine-sensitive OE33 were evident along the first principal component (PC1), while significant differences between the arginine-treated and -untreated groups of OE33 were observed along the second principal component (PC2) (Fig. 5a). A cluster analysis revealed that the levels of many metabolites differed between OE19 and OE33, and some metabolite levels differed between the arginine-treated and -untreated groups (Fig. 5b). Table 3 lists the metabolites with significant differences between the arginine-treated and -untreated groups (with relevant p values). In both OE19 and OE33, the arginine-treated groups showed significantly elevated levels of arginine and metabolites belonging to the urea cycle, including ornithine and argininosuccinic acid. In the arginine-treated group of the arginine-sensitive OE33, citrulline/ornithine levels were significantly lower than those in the untreated group (p = 0.012, Fig. 6a). Carbamoyl aspartic acid (CAA), creatine, IMP, succinic acid, and carnitine levels were significantly lower in the arginine-treated group of OE33 than in the untreated group. There were no significant differences between the two groups for OE19. The metabolic pathways and metabolite levels centered around arginine are shown in Figure 6b. Carbamoyl phosphate (CP), a metabolite involved in the urea cycle and pyrimidine synthesis pathway, was undetectable in all groups.

The total levels of arginine, ornithine, and citrulline were significantly higher in the treated group than in the untreated group (Table 4). ¹³C₆/¹³C₅/¹³C₄-arginine, ¹³C₅/¹³C₄-ornithine, and ¹³C₆-citrulline were detected in the treated group (Fig. 7; Table 5). These findings suggest that arginine administration promotes metabolism in the urea cycle.

In this study, MTT assays revealed that arginine treatment directly inhibits the growth of several upper gastrointestinal cancer cell lines. In cell cycle analyses, arginine treatment reduced cells in the DNA synthesis phase in the arginine-sensitive OE33. A comprehensive metabolomic analysis revealed increases in metabolites belonging to the urea cycle in the arginine-treated group. In particular, in the OE33 cell line, CAA levels were lower in the arginine-treated group than in the control group. A tracer study using (¹³C₆) l-arginine demonstrated that arginine administration affects the urea cycle. These findings suggest that arginine administration affects urea cycle metabolism, consuming CP and thereby decreasing CAA, a substance necessary for pyrimidine synthesis.

In recent years, 3D culture methods have garnered attention as a technique bridging the gap between 2D culture and in vivo conditions. Conventional 2D monolayer cell cultures have shown limitations in reproducing the microenvironment in which cancer cells proliferate in vivo. 3D culture methods, such as spheroid or scaffold-based cultures, provide a more physiologically relevant environment and are beneficial for evaluating drug sensitivity [19‒22]. The 3D culture method we have developed, termed the “tissueoid cell culture system,” utilizes a culture substrate called Cellbed made of ultrafine silica fibers, resembling the sparse connective tissue structure in vivo. This method enables long-term cultivation and can closely replicate the growth patterns and metabolic functions of cancer cells under in vivo conditions [11, 12]. We employed this culture system to evaluate the anticancer effects of arginine demonstrated in vivo in previous studies under in vitro conditions that resemble the physiological microenvironment.

In the MTT assay, conducting measurements over a longer period in the 3D culture system than in the 2D culture system allowed for a clearer distinction in the effects of arginine, suggesting a concentration-dependent effect. In the mitochondrial stress test, while there was no significant difference in the basal respiratory capacity between the arginine-treated and -untreated groups, a decrease in maximal respiration was noted in the arginine-sensitive OE33 cells, hinting at a potential reduction in mitochondrial metabolic activity due to arginine treatment. The decrease in mitochondrial respiration may reflect the decreased activity of mitochondrial dehydrogenases, evaluated in the MTT assay. Furthermore, we found that arginine affects the cell cycle. In the arginine-treated group of OE33, a decrease in cells in the S phase was observed. Since BrdU is taken up during DNA replication, the decrease in BrdU-incorporated cells in the S phase may reflect the suppression of DNA replication.

In a comprehensive metabolomic analysis, we compared the effects of arginine treatment in 3D culture on levels of various metabolites between arginine-sensitive OE33 and arginine-insensitive OE19. In both cell lines, the levels of ornithine and argininosuccinic acid, which are metabolites belonging to the urea cycle, along with arginine, were increased under arginine treatment. Additionally, in the arginine-treated group of OE33, the level of CAA was significantly lower than that in the untreated group, suggesting its potential contribution to the anticancer effects of arginine. Previous studies have evaluated the relationship between cancer and arginine, with a focus on NO production [9, 23]. Arginine is a substrate for both arginase and NO synthase (NOS). Although the affinity of arginine is much higher for NOS than for arginase, the maximum activity of arginase is >1,000 times that of NOS, resulting in comparable rates of substrate use [24]. A relative shift in activity between NOS and arginase has previously been noted in cell cultures following arginase inhibition or overexpression [25, 26]. Since different amino acids are produced by arginase (ornithine) and NOS (citrulline), the ratio of these amino acids will reflect the relative activity of the two enzymes. In the present study, a decrease in citrulline/ornithine levels in the treated group of OE33 suggests that the metabolic pathway of arginine toward ornithine and urea is more active than the pathway leading to citrulline and NO production [27]. In the tracer study using (¹³C₆) l-arginine, ¹³C₆/¹³C₅/¹³C₄-arginine and ¹³C₅/¹³C₄-ornithine were detected in the ¹³C₆-arginine-treated group. These findings also suggest that additional l-arginine was metabolized through the urea cycle.

In the urea cycle, CP, which is consumed during the conversion of ornithine to citrulline, is also necessary for the synthesis of pyrimidine nucleotides. The initial reaction in the de novo synthesis of pyrimidine nucleotides involves the sequential generation of CP, CAA, and dihydroorotate by carbamoyl-phosphate synthetase 2, aspartate transcarbamylase, and dihydroorotase (CAD), respectively. CP levels were below the detection limit in all samples, while CAA showed a significant decrease in the treated group of OE33. In cells with arginine-induced growth inhibition, a reduction in CAA, a substrate necessary for pyrimidine synthesis, may have been associated with the metabolism of arginine. In a previous study, the reaction catalyzed by dihydroorotate dehydrogenase (DHODH), the fourth step in pyrimidine synthesis following the CAD reaction, occurs in the inner mitochondrial membrane. DHODH transfers electrons to the electron transport chain complexes via the oxidation-reduction reaction of ubiquinone [28]. The decrease in mitochondrial respiration observed in the mitochondrial stress test may reflect not only the decreased activity of mitochondrial dehydrogenases observed in the MTT assay but also the decreased activity of DHODH in the pyrimidine synthesis pathway.

Previous studies have indicated that dysregulation of the urea cycle in cancer increases the flow of nitrogen to CAD, enhancing pyrimidine synthesis [29, 30]. Typically, carbamoyl-phosphate synthetase (CPS) 1, associated with the urea cycle, is located in the mitochondria, while CPS2 in the cytoplasm functions independently in the pyrimidine synthesis pathway. However, in certain cancer cell lines, CPS1 has been shown to maintain the pyrimidine pool [31]. Hence, further studies of the expression of enzymes and genes related to the urea cycle and pyrimidine synthesis as well as their relationship with arginine are warranted. If we can clarify the molecular differences between arginine-sensitive and arginine-insensitive strains, the anticancer effects of l-arginine can be applied in a clinical setting. This issue must be addressed in the future.

In conclusion, arginine supplementation directly inhibits the proliferation of several gastrointestinal cancer cell lines. In Figure 8, we present a speculative diagram of the mechanism underlying the effects of arginine based on the results of the present study. The consumption of CP through arginine metabolism in the urea cycle decreases CP and CAA, which are necessary for pyrimidine nucleotide synthesis, thus inhibiting the de novo synthesis of pyrimidine nucleotides.

We thank Japan Vilene Co., Ltd., which supplied Cellbed material. We also thank Mr. Takuya Iwasa and Mr Masaaki Kawabe (Japan Vilene Co., Ltd.), who provided information regarding the Cellbed characteristics. We thank Mr. Michiharu Nasu, who prepared paraffin sections and performed immunohistochemistry.

Since all cell lines used in the present study are commercially available, ethical approval is not required in accordance with local or national guidelines.

This research was done in collaboration with Japan Vilene Company, Ltd. The company provided Cellbed and information on Cellbed materials. The authors have no conflicts of interest to declare.

This study was not supported by any sponsor or funder.

Eri Tanaka: investigation, writing – original draft, and writing – review and editing. Naoko Taniura, Yusuke Kageyama, and Hirohito Ishigaki: investigation and writing – review and editing. Ken-ichi Mukaisho: conceptualization, supervision, writing – original draft, and writing – review and editing. Mai Noujima, Takahisa Nakayaka, and Ryoji Kushima: writing – review and editing.

All data generated or analyzed during this study are included in this article and its online supplementary material. Further inquiries can be directed to the corresponding author.