Effect of IC50 dose of retinol on metabolomics of RAW264.7 cells

Background: Vitamin A is one of the most multifunctional Vitamin in the human body and is involved in several basic physiological processes from embryonic development to adulthood, such as embryogenesis, vision, immunity, cell differentiation, and proliferation. The aim of current study was to identify its effects on the differential metabolites and main metabolic pathways. We also analyzed the toxicity of retinol. Methods: In this study, we conducted proton nuclear magnetic resonance (NMR) to evaluate metabolomic changes in RAW264.7 cells after treatment with retinol at a half maximal inhibitory concentration (IC 50 ). Results: Our results showed that the IC 50 dose (140 μM) of retinol affected the metabolism of RAW264.7 cells, with a total of 22 differential metabolites identified via 1 H-NMR, including amino acids, sugars, organic acids, glutathione, glycerin, and creatine. Additionally, multiple metabolic pathways were affected by retinol treatment, including amino acid biosynthesis, protein synthesis, and pyruvate metabolism. We speculate that the cytotoxicity of retinol at the IC 50 dose is attributed to mitochondrial dysfunction as a result of oxidative stress or lipid peroxidation. Conclusions: RAW264.7 cells showed significant changes in protein biosynthesis, urea cycle, arginine and proline metabolism, malate-aspartate shuttle, alanine metabolism, and cellular respiration after treatment with retinol at the IC 50 dose (140 μM), indicating that retinol affects cellular physiological functions via different metabolic pathways. metabolism, and cellular respiration after treatment with retinol at the IC 50 (140 µM), indicating that retinol affects cellular physiological functions via different metabolic pathways. After treatment with the IC 50 dose (140 µM) of retinol, the cells showed significantly increased levels of oxidative stress, IL-6, and TNF-α, likely because of the mitochondrial dysfunction and lipid peroxidation caused by retinol treatment. In this study, we investigated the adverse effects of high-dose retinol from a metabolomic perspective by evaluating changes in metabolites and metabolic pathways. However, the specific molecular mechanism underlying retinol-induced mitochondrial dysfunction remains unclear and requires further study.


Introduction
Vitamin A is a generic term for a group of fat-soluble, unsaturated organic compounds that includes retinol, retinal, and retinoic acid, among others [1].
Vitamin A and its precursors are mainly derived from livers, cod liver oil, dairy products, egg yolks, dark green leafy vegetables, dark yellow and orange vegetables and fruits, and vitamin supplements [2]. Vitamin A has various functions in the human body, such as maintaining normal vision and health of epithelial cells, promoting the synthesis of immunoglobulins, participating in the synthesis of glycoproteins [3], maintaining the normal growth and development of bones, and promoting protein biosynthesis and osteoblast differentiation [4]. The circulating form of vitamin A in the blood is retinol. After cellular uptake, retinol is converted into retinoic acid (RA), an active metabolite used by the immune system [5]. Both insufficient intake (which causes hypovitaminosis A) and excess intake (which causes hypervitaminosis A or vitamin A toxicity) of vitamin A can threaten human health [6]. Food fortification and dietary supplementation with retinyl ester are commonly employed to alleviate hypovitaminosis A. According to data from the World Health Organization (WHO), the Recommended Nutrient Intakes (RNIs) and Tolerable Upper Intake Level (UL) of vitamin A for a healthy adult males are 1998 IU and 9990 IU, respectively [7]. However, because of the rising interest in fortified foods and vitamin A supplements, most healthy individuals exceed the recommended dose of vitamin A [8].
Numerous studies have shown that excessive vitamin A intake may lead to neurological side effects [9], osteoporosis and fractures [10], and various structural abnormalities during embryogenesis [11,12]. Yeh et al. suggested that vitamin A 4 overdose (25,000-50,000 IU) can lead to liver damage and renal insufficiency in rats, and the mechanism underlying liver damage may be attributed to lipid peroxidation [13]. A study by Schuster et al. revealed that a high level of vitamin A (250 IU/g in diets) resulted in severe excessive airway inflammation and airway hyperresponsiveness by enhancing serum IgE and IgG1 responses and pulmonary eosinophilic inflammation in mice. Additionally, a high level of vitamin A shifts the Th1/Th2 balance towards Th2, leading to the aggravation of experimental asthma [14]. Bray et al. confirmed that retinol (75 mg/kg·day) significantly decreased the enzymatic activity and polypeptide levels of CYP3A in the liver and exacerbated the acetaminophen-induced liver toxicity [15]. High doses of vitamin A (1700 IU/g) directly reduced the bone diameter in rats and indirectly affected serum biomarkers of bone length, turnover, mineral density, and cortical bone [16]. A previous study demonstrated the involvement of the Wnt5a/CaMKII pathway in retinoic acid (RA)induced developmental abnormalities in mouse genioglossus [17]. Retinoic acid (10 µM) inhibits the proliferation and differentiation of C2C12 cells by upregulating the expression of miR-31-5p and miR-27b-3p and downregulating the expression of CamkIIδ and DTNA [18,19]. Exogenous RA (3 µM) promotes the apoptosis and inhibits the proliferation of epithelial cells lining the labial cervical loop by downregulating the expression of the mesenchymal factor Fgf10, thereby negatively regulating the homeostasis of the incisor stem cell niche. These effects are accompanied by inhibition of Bcl2 expression [20]. Zhang et al. suggested that alltrans retinoic acid (200 mg/kg) induces cleft palate by inhibiting degradation of the medial edge epithelium via the Notch1/RBPjk/P21 signaling pathway [21]. Liu et al. showed that excessive RA (100 mg/kg) induced cleft palate by downregulating the expression of histone methyltransferase in palate mesenchymal cells [22]. It has also been shown that RA-induced cleft palate is also attributable to downregulation of the classical Wnt/β-catenin signaling pathway [23]. Zhang et al. showed that alltrans retinoic acid (1 µM) inhibited cartilage-specific molecules (including aggrecan, Sox9, and Col2al) by regulating the expression of p53 and p21 to reduce chondrogenesis in rat embryo hindlimb bud mesenchymal cells, thereby inducing the onset of congenital club foot [24].
Metabolomics is an emerging field that can determine changes in gene expression levels by detecting changes in metabolite levels via metabolomics techniques, thereby indicating the functions of genes and their effects on metabolic flux.
Metabolomics not only allows for the comprehensive analyses of individual metabolites, but also increases the understanding of metabolic pathways and various complex interactions in organisms, as well as responses of biological systems to environmental and genetic changes [25]. Currently, most studies on vitamin A have been based on genomics, transcriptomics, and proteomics; very few studies have examined its metabolomic aspects. Additionally, most studies evaluated the metabolic pathways of endogenous vitamin A [26,27], while the toxicity of exogenous vitamin A has rarely been reported. Therefore, we evaluated the effects of vitamin A (retinol) at the IC 50 dose on the metabolomics of RAW264.7 cells via metabolomic techniques to investigate changes in metabolite levels in normal cells treated with retinol at an IC 50 dose to provide a theoretical basis for studying the toxicity of retinol.

Solution Preparation
Retinol was dissolved in ethanol solution to prepare a stock solution with a gradient concentration and stored at -80 °C. Next, the stock solution was diluted to the required concentrations (80, 100, 120, 140, 160, 180 µM) with basic medium to ensure a final volume fraction of ethanol of 0.5%. The solutions were filtered and sterilized with a 0.22-µm microfiltration membrane and then protected from light.
This cell line was originated from macrophages from ascites in the leukemia virus-7 induced tumor Abelson murine. The cells were cultured at 37 °C at 5% CO 2 and 95% humidity.

Cell Viability Assay
To test the effects of retinol on cell viability, an MTT assay was conducted. Cells in suspension were added to 96-well plates at 1 × 10 5 cells/well and incubated for 24 h at 37 °C. After removing the culture medium, fresh culture medium containing different concentrations of retinol (80, 100,120, 140, 160, and 180 µM) was added.
The cells were incubated for 24 h at 37 °C. Next, 20 µL of MTT was added to each well and the plate was incubated for 4 h at 37 °C. The culture medium in each well was removed, and 150 µL/well of dimethyl sulfoxide was added followed by incubation for 20 min with shaking at room temperature. The absorbance of the cell suspension was measured at 492 nm with a microplate reader.

Cytokine Assay
The IC 50 value of retinol in RAW264.7 cells was determined from the MTT assay results. Cells in suspension were placed in 24-well plates at 1 × 10 6 cells/well and incubated for 12 h at 37 °C. After the cells were attached, the original medium was discarded and the IC 50 dose of retinol group was added (RET group); non-treated cells were used as control (Control group). After 24 h, the culture medium was removed and centrifuged at 3000 rpm for 10 min. The supernatant was stored at -20 °C. IL-1β, IL-6, and TNF-α were quantified by using mouse ELISA kits (3 parallel groups were used for each experimental group).

1 H-NMR Spectroscopy Analysis
All 1 H-NMR spectra were obtained with a superconductor shielding Fourier transform NMR spectrometer equipped with 13 C and 1 H double resonance optimization of a 5mm CPTCI three trans detector CryoProbeTM AVANCE 600 III (Bruker) under the 9 following conditions: 600.104 MHz for proton resonance frequency, zg30 for pulse sequence, 12,019.230 Hz for spectral width, 16 repetitions, 2 dummy scans at 298 K, 1-s relaxation delay, and 12-µs pulse length. Topspin 3.2 was used to process the spectral data.

Data Analysis for 1 H-NMR Spectroscopy
The NMR spectra were processed with MestReNova 6.11 (Mestrelab Research, Santiago de Compostela, Spain). For all spectra, the exponential window function was added and processing was conducted with manual baseline correction. Using TSP as the standard, the spectra were manually cut off from the water peak and normalized. Data for the integrated peak were exported into numbered groups in Excel files. Next, the data were imported into SIMCA-P 14.1 software (Umetrics, Inc., Umea, Sweden). The data were analyzed using principal component analysis, partial To quantify the up-and downregulation of metabolites, we normalized the peak areas of identified the metabolites by summing the metabolites of the integral area multiplied by 1000 (relative value).
After identifying certain metabolites, we enriched the pathways using the Kyoto Encyclopedia of Genes and Genomes (KEGG) and MetaboAnalyst (MetPA). The metabolites were imported into KEGG to obtain a list of their KEGG ID numbers.
Next, the list was imported into MetPA to match the metabolic pathways using two methods. In pathway analysis, pathway impact characterization was used to draw graphs with the topology based on the importance of the metabolic pathways using the computed value, while the ordinated − log P revealed significant metabolic pathways in enrichment analysis. Another method for analyzing the functional enrichment analysis of humans and mammals is metabolite set enrichment analysis.
We conducted overrepresentation analysis a list of compound names was provided.

Statistical Analysis
Statistical analysis was performed using paired Student's t test or multi-way analysis of variance. Statistical analysis was conducted using Graphpad Prism software (GraphPad, Inc., La Jolla, CA, USA). Differences were considered as statistically significant when p < 0.05. All experiments were conducted independently at least three times.

Effects of Retinol (RET) on RAW264.7 Cell Viability
The results of the MTT assay revealed significant inhibition of cell viability after treatment with retinol for 24 h. According to the regression equation, the IC 50 of retinol for RAW264.7 cells was 138.69 µM (see Fig. 1), which was rounded to 140 µM for subsequent experiments.

Multivariate Statistical Analysis
The integrated areas under the chemical shifts derived from the NMR spectra were imported in an Excel format into SIMCA-P14.1 software for multivariate statistical analysis. Figure 3b shows the PCA score plots of RAW264.7 cells. The retinol group and control group significantly differed based on the first and second principal components. To achieve more accurate grouping, we used supervised analytical methods, such as PLS-DA and OPLS-DA, for data processing. Figures 3c and 3d show the PLS-DA and OPLS-DA score plots of RAW264.7 cells. The results showed that the retinol group and the control group significantly differed based on the first and second principal components with clear separations, indicating altered metabolite production in retinol-treated cells. We employed the permutation test to examine these models to evaluate their reliability. Figures 3e and 3f show the results for the PLS-DA and OPLS-DA models for RAW264.7 cells. The random R 2 and Q 2 values were lower than the rightmost point in both models. Furthermore, R 2 > 0, Q 2 < 0, and the rightmost values of the regression equation of R 2 and Q 2 were close to 1, suggesting that these models were not overfitted and are reliable.
In summary, the retinol group (RAW264.7 cells treated with retinol at the IC 50 concentration) and control group differed and were significantly distinguishable from one another.

Qualitative and Metabolic Pathway Analyses of Differential Metabolites
We ranked the VIP values from multivariate statistical analyses in descending order to identify the differential metabolites. Because of the large number of substances with VIP ≥ 1, chemical shifts with VIP ≥ 2 were used to quantitatively analyze the metabolites. The confirmed differential metabolites are listed in Table 1. A total of 22 differential metabolites, including amino acids, sugars, organic acids, glutathione, glycerin, and creatine, were identified.
13 Table 1 Differential metabolites in RAW264.7 cell line. To study the metabolic pathways of biomarkers associated with retinol treatment in RAW264.7 cells, we explored their potential interactions using the KEGG database.
We further confirmed the effects of retinol on these metabolic pathways using MetPA (Fig. 4). The 22 differential metabolites were mapped to a total of 24 metabolic pathways, such as the metabolism of various amino acids (including arginine and proline), citric acid cycle, metabolism of various organic acids (such as pyruvic acid), carbohydrate metabolism, lipid metabolism, nitrogen metabolism, and biosynthesis of unsaturated fatty acids (

Discussion
Macrophages are monocyte-derived leukocytes present in tissues and play important roles in host defense, tissue development, and homeostasis [28]. The mouse macrophage cell line RAW264.7 is derived from peritoneal effusion of male Balb/c mice. These immune cells are commonly used for laboratory experiments because of their high adhesive and phagocytic capacities [29]. Numerous studies have evaluated the mechanism of macrophage activation using metabolomic techniques. For example, Wu et al. studied the anti-inflammatory effect of nonsteroidal anti-inflammatory drugs on RAW264.7 cells through the ultra-highperformance liquid chromatography/quadrupole time-of-flight mass spectrometry method and found that the effects were related to changes in glycerophospholipid metabolism [30]. Aarash et al. conducted gas chromatography-mass spectrometry and suggested that lipopolysaccharide activates macrophages by downregulating the de novo synthesis of nucleotides, which is regulated by the glutamate-glutamine complex and the malate-aspartate shuttle [31].
Our results showed that retinol significantly increased the levels of IL-6 and TNF-α in RAW264.7 cells but did not affect IL-1β levels. Several studies have demonstrated the effects of vitamin A on cytokines. Retinoic acid increased the transcription levels of IL-6 (by upregulating the transcription and stability of mature transcripts), monocyte chemoattractant protein-1, granulocyte-macrophage colony-stimulating factor, and IL-10 in macrophages [32]. Treatment of Wistar rats with vitamin A (2500 IU/kg·day) for 3 months enhanced monoamine oxidase activity, mitochondrial redox dysfunction, and TNF-α levels [33]. Gasparotto et al. found that vitamin A (2000 IU/kg, oral) reduced the pulmonary levels of superoxide dismutase, increased the levels of lipid peroxidation and protein carbonylation, and increased the level of TNF-α while decreasing the level of IL-10, but it did not affect the level of IL-1β in rats [34], which is consistent with our findings for cytokines.
Excessive retinol can enhance the production of free radicals and cause oxidative stress, thereby resulting in early apoptosis due to necrosis [35]. Different doses of vitamin A (1000, 2500, or 4500 IU/kg) induced the onset of a pro-oxidation state and large-scale production of superoxide anions in the hippocampus of adult rats, leading to oxidative stress and hippocampal toxicity [36]. Vitamin A supplementation induced oxidative damage to biomolecules [37], upregulation of antioxidant enzymes [38,39], and activation of phosphorylation signaling pathways in Sertoli cells [40][41][42]. These results strongly suggest that vitamin A promotes oxidative stress and has potential pro-oxidative capacity. In this study, we observed reduced glutathione levels and elevated betaine levels, both of which are differential metabolites in cells. Glutathione is a major redox "buffer" in the cytoplasm [43], while betaine is widely considered as an antioxidant that can alleviate oxidative stress [44] [45]. Therefore, we predicted that after treatment with retinol at the IC 50 , the large amount of RET entering the cells could cause mitochondrial dysfunction and lipid peroxidation to result in massive superoxide anion production because of the leakage of the electron transport chain, thereby leading to oxidative stress responses. The significantly elevated levels of IL-6 and TNF-α indicated that the onset of intracellular inflammatory responses may be attributed to the maintenance of steady-state mitochondrial activity in cells because of mitochondrial dysfunction [46].
We also found that RET significantly downregulated the levels of argininosuccinic acid and L-malic acid and significantly increased the levels of fructose-6-phosphate.
Metabolic pathway mapping of differential metabolites revealed the participation of argininosuccinic acid in the metabolism of arginine and proline; metabolism of alanine, aspartic acid and glutamate and the participation of L-malic acid in the metabolism of glyoxylic acid and dicarboxylic and tricarboxylic acid (TCA) cycles as well as the metabolism of pyruvic acid (see Table 2). The results of pathway enrichment revealed varying degrees of abnormalities in the malate-aspartate shuttle, TCA cycle, and pyruvic acid metabolism, suggesting abnormalities in cellular energy metabolism that eventually led to cell death.
Argininosuccinic acid is a basic amino acid synthesized from citrulline and aspartic acid in some cells and is used as a precursor for arginine in the urea cycle and citrulline-NO cycle. Additionally, argininosuccinic acid is the precursor for fumaric acid in TCA cycle [47].  [50]. Our experimental results partially agree with the above findings. Therefore, we predicted that retinol caused abnormalities in the metabolism of pyruvic acid and the TCA cycle by significantly downregulating the levels of argininosuccinic acid and L-malic acid, leading to cellular energy abnormalities that reduced aerobic respiration in cells. Elevation of fructose-6phosphate may be caused by the accumulation of anaerobic respiration-related molecules, suggesting a conversion from aerobic to anaerobic respiration that reduces the cellular energy supply. Previous analyses of retinol-induced mitochondrial dysfunction and electron transport chain leakage appear to support our hypothesis. Furthermore, the reduction of argininosuccinic acid can lead to urea cycle disorders, which are biochemically characterized by aggravated argininosuccinic acid and arginine deficiencies [51]. In humans, the urea cycle mainly occurs in the liver, and thus excessive retinol can lead to urea cycle abnormalities that cause hyperammonemia, leading to hepatotoxicity and neurotoxicity of the elevated creatine levels provide an basis for our prediction.

Conclusion
In conclusion, RAW264.7 cells showed significant changes in protein biosynthesis,

Acknowledgement
We thank Editage Company for its linguistic assistance during the preparation of this manuscript.

Funding
Not applicable.

Author contributions
YW conceived, designed and performed research, analyzed data and wrote the first draft of the paper. XS performed research and reviewed the manuscript. YG conceived and designed the research, and reviewed the manuscript. All authors have read the manuscript before submission.

Availability of data and materials
The datasets used and analyzed during the current study are available from the corresponding author (YG) upon reasonable request.

Ethics approval and consent to participate
This study did not involve the use of animal or human samples.

Consent for publication
All the authors gave their consent to the publication. The work described was original research that has not been published previously, and is not under consideration for publication elsewhere, in whole or in part.  Multivariate statistical analysis between the control group and the retinol treatment group. a. PCA 44 Figure 5 a. Metabolic pathways attributed to differential metabolites in RAW264.7 cells; b. Overview of me