Abstract Retinal detachment is a serious ocular disease leading to photoreceptor degeneration and vision loss. However, the mechanism of photoreceptor degeneration remains unclear. The aim of this study was to investigate the altered metabolism pathway and physiological changes after retinal detachment. Eight-week-old male SD rats were fed, and the model of retinal detachment was established by injecting hyaluronic acid into the retinal space. The rats were euthanized 3 days after RD, and the retinal tissues were sectioned for analysis. Untargeted lipid chromatography-mass spectrometry lipidomic was performed to analyze the metabolite changes. A total of 90 significant metabolites (34 in anionic and 56 in cationic models) were detected after retinal detachment. The main pathways were (1) histidine metabolism; (2) phenylalanine, tyrosine, and tryptophan biosynthesis; and (3) glycine, serine, and threonine metabolism. The key genes corresponding to each metabolic pathway were verified from the Gene Expression Omnibus (GEO) database of human retinal samples. The results indicated that the production of histamine by histidine decarboxylase from histidine reduced after RD (p < 0.05). Xanthine, hypoxanthine, guanine, and guanosine decreased after RD (p < 0.05). Decreased xanthine and hypoxanthine may reduce the antioxidant ability. The decreased guanosine could not provide enough sources for inosine monophosphate production. Tyrosine is an important neurotransmitter and was significantly reduced after RD (p < 0.05). Citrate was significantly reduced with the increase of ATP-citrate lyase enzyme (ACLY) (p < 0.05). We inferred that lipid oxidation might increase rather than lipid biogenesis. Thus, this study highlighted the main changes of metabolite and physiological process after RD. The results may provide important information for photoreceptor degeneration. Keywords: retinal detachment, metabolomics, amino acid, histamine, retina degeneration 1. Introduction Retinal detachment results from the separation of the retinal neurosensory layer from the retinal pigment epithelium layer and is commonly detected in rhegmatogenous retinal detachment, diabetic retinopathy, age-related macular degeneration, and other retinal disorders [[36]1,[37]2]. With the rapid development of surgical intervention for rhegmatogenous retinal detachment, higher rates of successful surgery could be achieved for most patients. However, over 50% of the patients could not achieve good visual acuity even after successful reattachment surgery [[38]3]. Photoreceptor degeneration leading to cell death has been considered as the major reason for progressive visual impairment after retinal detachment [[39]4]. The retinal pigment epithelium (RPE) infiltrates into the vitreous and begins to join in the process of proliferative vitreoretinopathy (PVR), which can destroy retinal integrity and cause photoreceptor degeneration [[40]5]. However, the underlying mechanism of photoreceptor degeneration was unclear until now, and it is necessary to identify the mechanism. Metabolomics reflects the physiological and pathological process for disease and has been widely used to identify the key networks of metabolic files in diabetic retinopathy, cancer, and other diseases [[41]6]. Metabolomics helps in understanding the interactions between genes and proteins [[42]6]. Previous studies analyzed the vitreous samples from RD and PVR patients and the results pointed that, after RD, the main changes were inflammation, proliferation, and energy consumption, like the significant alterations of L-carnitine, ascorbate, and valine [[43]7,[44]8]. Increased citric acid cycle metabolism is detected in patients with choroid and rhegmatogenous retinal detachment [[45]9]. Another study pointed that the adenosine and inosine were increased from the vitreous in patients with retinal detachment than epiretinal membrane [[46]10]. However, these studies only analyzed vitreous samples from patients, not retinal samples. In addition, there were no results to reveal the early metabolomic changes of RD. Humans used rodent models of acute retinal detachment to investigate the cellular events for the difficulty in obtaining human retina tissues. The rats are the ideal model for lower ethical and financial costs [[47]11]. Our study established the rat model of RD and directly analyzed the early changes in the retina, which may reveal the key changes and help to understand the detailed mechanism for RD. In the present study, we used untargeted metabolomics to analyze the global metabolomic profiles in a rat model of retinal detachment. It revealed 90 discriminant metabolites based on high-resolution mass spectrometry. The study revealed that histidine, phenylalanine, tyrosine, and tryptophan biosynthesis and glycine, serine, and threonine metabolism were the top changed pathways. Then, we compared the gene alternations in the metabolomics pathway after retinal detachment using human retina tissues from the GEO public database. These gene alternations and metabolomics changes indicated that decreased histamine, xanthine, and hypoxanthine and increased lipid oxidation were the main physiological processes after RD. Herein, we reported the main metabolism changes and potential alternations of key pathways verified from the human database. These results provide new information for retinal detachment and may serve as important clues to identify clinical biomarkers. 2. Materials and Methods 2.1. Animal Ethics and the Animal Model of Retinal Detachment This study was in accordance with the statement on the use of animals by American Association for Laboratory Animal Science. It was approved by the Institutional Animal Care and Use Committee of Wenzhou Medical University, China (Number: wydw2021-0068). In all, 25 large male 8- to 10-week-old SD rats weighing about 160 g were obtained from Shanghai Slake Experimental Animal Co., Ltd. (Shanghai, China). All the animals were fed in custom cages at the animal facility separately to prevent contact among them. After they arrived at the animal facility for study, all the animals were fed for 3 days to allow them to adapt to the environment. A total of 12 and 10 animals were randomly assigned to the experiment and control group. The right eye was chosen as the experiment and control eye. On day 5 after the rats were housed in the new animal facility, retinal detachment was induced in the rats according to the previous process [[48]4]. The right eyes of the 12 animals were selected as the experimental group to be studied. The rats were anesthetized with the drug sodium pentobarbital (i.p., 30 mg·kg^−1) [[49]4,[50]11]. Then, the pupils were dilated with topical tropicamide (0.5%) and phenylephrine hydrochloride (0.5%) [[51]4,[52]11]. A scleral incision was made 1.5 mm posterior to the corneal limbus and sodium hyaluronate was injected into the subretinal space to ensure that approximately two-thirds of the neurosensory retina detached from the underlying RPE and were floated in the vitreous cavity without complications such as much bleeding et al. The incision was made with a 27-gauge needle to avoid damage to the lens (10 mg·mL^−1, LG life Sciences, Seoul, Korea). Ofloxacin antibiotic ointment (Dikeluo; Sinqi Pharmaceutical Co., Ltd., Shenyang, China) was applied to the scleral and ocular surface. The rats were monitored daily after surgery. 2.2. Sample Collection and Preparation Three days after establishing the RD model, the rats were anesthetized with the drug sodium pentobarbital (i.p., 30 mg·kg^−1). Then, the pupils were dilated with topical tropicamide (0.5%) and phenylephrine hydrochloride (0.5%) [[53]4,[54]11]. The eyes were examined under the microscope. Any infectious or bleeding tissues were excluded, and the cornea was cut using scissors, the lens and vitreous were removed, and the remaining retina was separated from the RPE. The retinas were collected, placed on ice, and stored immediately at −80 °C. Retina samples were homogenized in 800 ul methanol: water (80: 20, v:v) and 0.5 μL 2-Chloro-L-phenylalanine (1 mg/mL in water) were added for internal standard by using a homogenizer by two cycles of grinding (60 s for each, 40 Hz). Then, 600 μL of supernatant was obtained and dried in a centrifugal vacuum concentrator after centrifugation at 15,000× g for 15 min. The dried extract was resuspended with 100 μL of 10% methanol/water (v/v) for analysis. Quality control samples were prepared by pooling equal aliquots of each sample together, and then they were prepared as other samples previously; every 5 samples followed by one QC sample were injected to monitor the stability [[55]12]. 2.3. LC–MS/MS Analyses for Untargeted Metabolomics Retina extracts, QC samples, and solvent blanks were analyzed by Ultimate 3000 UPLC and Orbitrap Fusion Lumos Tribrid mass spectrometer [[56]13]. Metabolites were separated on Waters^TM acquity BEH C18 column (2.1 × 100 mm, 1.7 µm) with a flow rate of 0.35 mL/min. H[2]O was used as mobile phase A and methanol was used as mobile phase B in negative ionization mode. While in positive ionization mode, 0.1% formic acid was added into both mobile phases, respectively. The LC gradient was as follows: 0–1 min, 2% B, 1–12.5 min, 2–50% B, 12.5–14.5 min, 50–98% B, 14.5–17.5 min, 98% B; then, mobile phase B returned to 2%. After chromatographic separation, metabolites were ionized b H-ESI. Spray voltage was +3.8 KV/−2.5 KV, and capillary temperature was 320 °C. Full MS scan and data-dependent acquisition was performed to acquire the ms1 m/z and MS/MS information of metabolic features. Full scan (m/z 70–1050) used resolution 120,000 with automatic gain control (AGC) target of 4 × 10^5 ions and a maximum ion injection time (IT) of 50 ms. MS/MS scan parameters were as follows: resolution 30,000; AGC 5 × 10^4 ions; maximum IT 54 ms; 1.6 m/z isolation window; Stepped HCD normalized collision energy 15%, 30%, 50%. 2.4. Data Extraction and Analysis Peak areas were extracted from the samples by converting the raw mass spectrometry files to mzXML using ProteoWizard and processed with an in-house program, which was developed using R and based on XCMS (version 3.6.2), for peak detection, extraction, alignment, and integration [[57]13]. Then, an in-house MS2 database (Biotree DB) was applied in metabolite annotation. The cutoff for annotation was set at 0.3. The resultant data were subjected to QC-based normalization (LOESS normalization) and RSD filtering of normalized peaks in the QC samples (RSD < 0.3). The matching of non-targeted metabolite substances is mainly carried out as follows: the molecular weight of the metabolite is determined according to the mass charge ratio (m/z) of the parent ion in the primary mass spectrum, the mass charge ratio of the feature ion generated after fragmentation, and the response intensity of the sub ion. The identification of metabolites and the calculation of the material matching score are based on these previous processes. Finally, to identify the substances, we will analyze the information and compare with standard products in our own and public databases. Our own database combines the information of HMDB, MONA, METLIN, and other public databases. For pathway enrichment analysis, significant metabolites under NEG or POS mode were combined on the basis of the KEGG human metabolic pathways. Metabolites containing at least two entries were used for analysis. The enrichment ratio was computed by hits/expected. A pathway with p < 0.05 was considered significant. The Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway was analyzed to reveal the enriched pathways of the altered metabolites. 2.5. RNA-Seq, Sequencing Data Extraction Analysis The sequencing data ([58]GSE28133) were download from the public database of chips and microarrays in GEO. The data included 38 human retina samples, 19 samples from RD patients, and 19 samples from patients without RD [[59]14]. Gene alternation between detached and intact retinas were compared with |log[2] Fold Change (FC)| > 1.0 and adjusted p-value < 0.05 [[60]15]. After analyzing the alternations of pathways in the main metabolomics, we looked for the key genes corresponding to metabolomics pathway in the public database. 2.6. Statistical Analysis In the untargeted metabolomics, the data were normalized to total peak intensity, and the processed data were uploaded for multivariate data analysis. The data matrix was analyzed using MetaboAnalyst v5.0 and SMICA v16.0 (Umetrics, Umea, Sweden). All the data underwent log10 transformation before analysis. Univariate analysis (ANOVA), partial least squares discriminant analysis (PLS-DA), and orthogonal partial least squares discriminant analysis (OPLS-DA) were performed using MetaboAnalyst v5.0. To evaluate the statistical parameters (accuracy, correlation coefficient (R2), and cross-validation coefficient (Q2)), PLS-DA models were analyzed using the leave-one-out cross-validation (LOOCV) method. To assess the robustness of the model, the 7-fold cross-validation and response permutation testing was carried out. The variable importance in the projection (VIP) for each variable was calculated for its contribution to the classification. VIP > 1 with significant changes was further applied to Student’s t-test at the univariate level, where p < 0.01 or p < 0.10 was considered as statistically significant. The genes related to metabolomics changes were analyzed between control and retinal detachment groups, and p < 0.05 was considered as significant. 3. Results 3.1. Untargeted Metabolomics of Retina Samples of Rat Model of RD A microscope was used to evaluate RD, i.e., that the detached retinas were floating in the vitreous cavity after surgery. HE staining was also used to verify successful development of the RD model. In all, 20 retinal samples (10 RD and 10 controls) were collected for untargeted metabolomics analysis. A total of 1074 peaks were recorded by XCMS records. Principal component analysis (PCA) indicated a good repeatability both in positive and negative models. The QC samples were gathered together closely, which showed good quality control (in [61]Figure 1). The differences in metabolomics between the two groups were identified by OPLS-DA score plots. The results indicated a distinct line between the RD group and the control group, as shown in [62]Figure 1A,B. The permutation analysis of the OPLS-DA model was carried out and proved to be valid and stable in [63]Figure 1C,D. Figure 1. [64]Figure 1 [65]Open in a new tab Qualification of untargeted metabolomics analysis. (A,B) The PCA analysis of the included samples in both positive and negative models. (C,D) OPLS-DA score plots under the positive and negative models. (E,F) Permutation analysis plots of the OPLS-DA model under the positive and negative models. Significant alternations in metabolites were identified using the criteria of fold change (FC) > 1 and p < 0.01. A volcano plot analysis was performed to present the potential changes in metabolites ([66]Figure 2). All the significantly changed metabolites were compound identified within the database, where 60 altered metabolites were identified under the positive group (p < 0.01 in 56 of them; [67]Table 1; [68]Figure 3) and 36 altered metabolites were identified under the negative group (p < 0.01 in 34 of them; [69]Table 2; [70]Figure 4). The XIC and MS/MS spectra were in the [71]Supplementary Figure S1. Figure 2. [72]Figure 2 [73]Open in a new tab Volcano plots of the untargeted metabolomics under the positive model (A) and the negative model (B) according to the criteria FC > 1.5 and p < 0.05. Table 1. Significantly altered metabolites under positive mode by untargeted metabolomics. Metabolite VIP ^a Fold Change p-Value mz rt Sarcosine 1.588794734 0.753842042 0.00172 90.05489512 43.3375 Cytidine 1.809324897 0.460435551 0.00003 244.0867888 56.1297 2-Pyrrolidinone 2.031331915 0.671816243 0.00004 86.06002251 44.74485 Nb-p-Coumaroyltryptamine 1.905917252 7.582405045 0.02603 307.1416345 496.676 Guanosine 2.153682946 0.198874723 0.00062 284.0989995 154.511 Uridine 1.64987382 0.580089132 0.00488 245.0759403 92.4567 L-Carnitine 1.698905167 1.751054619 0.00012 162.1079781 42.6997 Coriandrone D 1.453741941 1.460356861 0.03866 353.1589536 910.33 Uracil 1.690394112 0.613961856 0.00439 113.0335019 92.8533 2-Hydroxypyridine 1.812738744 0.863836807 0.00079 96.04164293 69.8369 Ascorbic acid 2.04657221 0.57252687 0.00005 177.0330868 51.13885 3-Hydroxy-2-methylpyridine-4,5-dicarboxylate 1.814382779 0.475699949 0.00003 198.0372243 64.1571 L-Histidine 1.529516838 1.44624588 0.00052 156.0754384 39.2151 Niacinamide 1.341138793 0.735454619 0.00428 123.0552253 70.6486 Propionylcarnitine 1.366650349 2.162667902 0.00362 218.1380477 136.899 Pyroglutamic acid 1.825841104 0.403212341 0.00005 130.0498196 82.9935 Cytosine 1.783710152 0.47980553 0.00003 112.050413 57.1205 4-Methoxybrassinin 1.659923965 0.516514641 0.00334 267.0587392 91.8458 Guanine 2.162064502 0.208577348 0.00081 152.0564281 154.536 Hypoxanthine 1.725128074 0.611006413 0.00005 137.0466486 70.23255 m-Xylene 1.921696126 1.243960894 0.00133 107.0826798 939.389 Xanthine 1.457026673 0.743827737 0.00145 153.0433249 79.82625 Phosphorylcholine 1.311159946 1.419536743 0.00481 184.0697118 40.3446 L-Palmitoylcarnitine 1.680174185 4.978528439 0.03278 400.3412165 964.8555 Succinyladenosine 1.480016326 0.618517209 0.00201 384.1150194 330.112 (-)-Epigallocatechin 1.281914517 2.687838842 0.01162 307.0833389 105.983 gamma-Glutamylcysteine 1.16082442 0.633360213 0.02104 251.0704476 68.27515 N2-gamma-Glutamylglutamine 1.79789316 0.502733902 0.00008 276.1224994 44.7747 2-(Methylthio)propane 1.907264524 1.267761325 0.00037 91.05418475 939.711 Dimethyl dialkyl ammonium chloride 1.363211764 1.178980467 0.04779 304.3001743 940.381 S-Adenosylhomocysteine 1.152857028 0.674020148 0.00332 385.1285023 110.14 LysoPC(22:6(4Z, 7Z, 10Z, 13Z, 16Z, 19Z)) 1.289484261 0.498294474 0.01125 568.3387229 984.4805 Atroviridin 1.909820405 0.393778382 0.00000 327.0798169 100.4 Oxidized glutathione 1.188118624 4.000915552 0.02011 613.1586921 105.3015 Guggulsterone 1.38032097 1.641203242 0.04156 343.2244804 972.79 (1R, 2R, 4R, 8S)-p-Menthane-1,2,8,9-tetrol 9-glucoside 1.248626769 0.575134248 0.00270 367.1949551 1013.06 Ambonic acid 1.624811991 2.388539493 0.00620 469.3625627 1112.09 Butyrylcarnitine 1.58194961 2.910638976 0.00087 232.154252 279.752 cis-4-Carboxymethylenebut-2-en-4-olide 1.746288625 0.692541355 0.01425 141.0175682 51.75935 Methyl phenyl disulfide 1.260869035 0.681308476 0.00151 157.0143964 51.5829 2-Methylbutyroylcarnitine 1.389115898 2.092890713 0.00103 246.1699963 402.919 Creatine 1.75441836 0.671404514 0.00078 132.076239 43.8692 N-Acetylornithine 1.538467126 1.376210914 0.00205 175.1132487 38.8724 Vitamin A 1.18216123 0.434496736 0.00634 269.2288829 1010.94 1-deoxy-1-(N6-lysino)-D-fructose 1.772390317 0.665660902 0.00050 134.0442436 45.63555 2-(2-Furanyl)-3-piperidinol 2.015285296 10.37822452 0.03335 168.1016873 97.7082 LysoPE(22:6(4Z, 7Z, 10Z, 13Z, 16Z, 19Z)/0:0) 1.68145778 0.420387817 0.00106 526.2975318 985.298 gamma-Glutamylglutamic acid 2.17322362 0.344750109 0.00000 277.1105039 53.96735 5’-Deoxy-5’-(methylsulfinyl)adenosine 1.432320977 0.49083823 0.00000 314.0810526 46.68055 Brassicanal C 1.639967821 0.402276508 0.00003 224.0375107 70.74 alpha-Methylstyrene 1.049964179 1.497412492 0.02551 119.0854447 939.6095 3-O-Acetylepisamarcandin 1.31455005 1.62786731 0.01162 460.2694428 939.137 Elaidic carnitine 1.522562393 3.678436325 0.00832 426.3562901 966.6665 L-Aspartyl-4-phosphate 2.045673216 0.473543833 0.00000 214.0126044 54.0449 Persicaxanthin 1.166835004 0.641843353 0.02979 385.2810119 985.3045 Prostaglandin G2 1.361188168 0.605822078 0.00202 351.2220203 1012.245 Histamine 1.11300757 0.385722377 0.00023 96.92917902 39.3631 Lansoprazole sulfone 1.575610946 0.447892561 0.01350 386.0851118 92.13835 Cysteic acid 1.67467486 0.658755513 0.00010 170.0084242 39.9788 [74]Open in a new tab The significant metabolites by univariate analysis under positive mode. Metabolites display a fold change greater than 1.5. ^a FC were measured using median values median values between retinal detachment group compared with control group. p value < 0.05 was considered as significant. Figure 3. [75]Figure 3 [76]Open in a new tab Heatmap of the different metabolites under the positive model in untargeted metabolomics. The blue color indicates the lower relative level of each metabolite, and the red color stands for the higher relative level of each metabolite. Table 2. Significantly altered metabolites under negative mode by untargeted metabolomics. MS2 Name VIP ^a Fold Change p-Value mz rt Glycine 1.308701658 1.418244394 0.00497 74.02476574 41.604 Pyrrolidonecarboxylic acid 1.182346811 0.73594283 0.02202 128.0353476 45.9248 L-Norleucine 1.299366092 1.359856849 0.00721 130.0870664 85.1877 Dodecanoic acid 1.013687543 1.112895178 0.02735 199.1699751 998.35 Pentadecanoic acid 1.572438882 1.238311403 0.00004 241.2170986 1032.37 Capric acid 1.22167135 1.160060897 0.00631 171.1355328 957.741 Linoleic acid 1.093421764 1.201474709 0.04900 279.2325558 1036.41 O-Phosphoethanolamine 1.016072298 1.216167453 0.01543 140.0101022 37.523 Adenine 1.446337373 1.629672015 0.01034 134.0472006 133.03 L-Tyrosine 1.272211932 1.527389323 0.00934 180.0663514 71.8062 Pyridoxal 1.319697042 1.421159579 0.00678 166.0464366 108.894 Sarcosine 1.422117232 0.761449281 0.00772 88.04041793 41.2361 L-Phenylalanine 1.429734239 1.42959732 0.00585 164.0714753 191.467 N-Acetyl-L-aspartic acid 1.738817993 0.5753958 0.00001 174.0405148 36.9303 13S-hydroxyoctadecadienoic acid 1.198241881 1.17631196 0.01217 295.2268903 984.563 D-Glutamine 1.213915884 0.729900743 0.00510 145.0688397 39.775 L-Aspartic acid 1.280263363 1.387046842 0.00218 132.0300608 37.6211 Glycolic acid 1.247166563 1.151900036 0.00991 75.00868616 1165.97 Xanthosine 1.754820236 2.161916582 0.00002 283.0677416 231.952 5,6-DHET 1.076984772 1.17425398 0.01023 337.2377561 1000.645 16-Hydroxy hexadecanoic acid 1.047890336 0.864921063 0.01094 271.2201796 982.637 Myristoleic acid 1.390731089 1.52634857 0.00146 225.1854497 1008.66 16(R)-HETE 1.010988944 2.271573423 0.04334 319.2267346 990.646 Ascorbic acid 1.18946295 0.749711824 0.03688 175.0256936 38.9044 Deoxyinosine 1.91372124 3.310838197 0.00041 251.0803053 193.8595 5Z-Dodecenoic acid 1.85788377 1.465643497 0.00000 197.1541993 981.979 Leukotriene B4 1.141316937 1.57686519 0.04224 335.2199433 974.804 Citric acid 1.334618831 0.362946196 0.00560 191.0191466 36.623 D-Myo-inositol 4-phosphate 1.329663611 2.419328351 0.02151 259.0210802 36.9303 S-Adenosylhomocysteine 1.349938422 0.779341895 0.00077 383.1136101 236.586 Thymidine 1.700378459 2.107103442 0.00024 241.0778921 256.103 Sorbitol 1.250229055 1.489605484 0.01152 181.0673795 71.9667 (10E, 12Z)-(9S)-9-Hydroperoxyoctadeca-10,12-dienoic acid 1.020361506 1.198479736 0.02153 311.2223869 966.673 Ethylparaben 1.733042441 23.16190628 0.01128 165.0511808 789.595 Guanosine 1.809686128 0.283277591 0.00120 282.0842135 171.802 [77]Open in a new tab The significant metabolites by univariate analysis under negative mode. Metabolites display a fold change greater than 1.5. ^a FC were measured using median values median values between retinal detachment group compared with control group. p value < 0.05 was considered as significant. Figure 4. [78]Figure 4 [79]Open in a new tab Heatmap of the different metabolites under the negative model in untargeted metabolomics. The blue color indicates the lower relative level of each metabolite, and the red color stands for the higher relative level of each metabolite. 3.2. Involved Pathways Related to Changed Metabolites in RD The altered metabolites were analyzed, and KEGG was used to discover the involved signaling pathway of metabolites. Different metabolites were obtained in POS and NEG modes ([80]Figure 5). Metabolic pathways are displayed in [81]Figure 5. The top related pathways are (1) histidine metabolism; (2) phenylalanine, tyrosine, and tryptophan biosynthesis; (3) glycine, serine, and threonine metabolism; (4) vitamin B6 metabolism; (5) pyrimidine metabolism; and (6) nitrogen metabolism (p < 0.05). Figure 5. [82]Figure 5 [83]Open in a new tab KEGG pathway indicating the top 20 involved pathways of significantly changed metabolites under the positive model (A) and the negative model (B). The spot size stands for the compound number of metabolites, and the color stands for the p value. Gene Alternations Verified in the Top-Ranked PATHWAY. The pathway of metabolomics has been presented previously. The genes in the related metabolism pathways were summarized and compared between detached and intact retinas from human beings in the GEO database. 3.2.1. Histidine Metabolism Of 15 metabolites, 2 were detected in the histidine metabolism pathway. After RD, histamine and S-adenosylhomocysteine (SAH) decreased and L-histidine increased ([84]Figure 6). L-histamine could be catalyzed by histidine decarboxylase [[85]16]. SAH is the downstream product of L-histidine when methyltransferases catalyze the methyl group in SAM [[86]17]. An analysis of the gene alternations after RD from GEO indicated that histamine N-methyltransferase, histidine decarboxylase, and aldehyde dehydrogenase increased after RD ([87]Figure 6). Figure 6. [88]Figure 6 [89]Open in a new tab The alternations of histidine metabolism and pathways. (A) The alternations in histidine metabolomics by untargeted metabolomics. (B) The gene alternations of the histidine metabolomic pathway after RD. (C) The summarized pathway of histidine metabolomics, 1 stands for histidine decarboxylase, L-histamine could be catalyzed by histidine decarboxylase, 2 stands for aldehyde dehydrogenase, histamine could be oxidated to imidazoleacetic acid and changed to Imidazol acetate by Aldehyde dehydrogenase, 3 stands for histamine N-methyltransgerase, released histamine is degraded to 1,4-methyl imidazoleacetic acid and it could be changed to 1,4-methylimidazol acetate by Aldehyde dehydrogenase. 3.2.2. Purine and Pyrimidine Metabolism Of 68 metabolites, 8 were detected in purine metabolism. Xanthine, hypoxanthine, guanine, and guanosine decreased, while deoxyinosine, adenine, and xanthosine increased after retinal detachment ([90]Figure 7A,B). In pyrimidine metabolism after RD, uridine, cytidine, and uracil decreased while thymidine increased ([91]Figure 7A,B). Xanthine and hypoxanthine are catalyzed by xanthine dehydrogenase (XDH) with the byproduct of uric acid in the intracellular spaces [[92]18]. However, the gene alternations of purine and pyrimidine metabolism from GEO indicated that genes of xanthine dehydrogenase, XO, and lactoperoxidase were not changed after RD. Figure 7. [93]Figure 7 [94]Open in a new tab The alternations of main metabolomics and related genes of metabolomics in the pathway. (A,B) The changes of purine and pyrimidine metabolism. (C) Gene change of IMPDH1 after RD. (D) The changes of phenylalanine, tyrosine, and tryptophan biosynthesis. (E,F) The changes of genes in tyrosine hydroxylase, phenylalanine hydroxylase, and tyrosinase-related protein1 after RD. (G) Gene change of ATP-citrate lyase enzyme after RD. (H) The changes of glycosylate and dicarboxylate metabolites after RD. (I) Gene alternations of phosphofructokinase 1 (PFK1) and pyruvate dehydrogenase 1 (PDH1) after RD. 3.2.3. Phenylalanine, Tyrosine, and Tryptophan Biosynthesis Of four metabolites, two were detected in the phenylalanine, tyrosine, and tryptophan biosynthesis process. After RD, L-tyrosine and L-phenylalanine increased ([95]Figure 7C). Tyrosine is synthesized from phenylalanine by the enzyme of phenylalanine hydroxylase (PHA) [[96]19]. Tyrosine hydroxylase (TH) is the enzyme catalyzing tyrosine to L-3,4-dihydroxyphenylalanine (L-DOPA) [[97]19]. However, the gene alternations after RD indicated that PHA increased ([98]Figure 1D), and TH was not altered after RD. These results indicated that the production of tyrosine from phenylalanine increased after RD. 3.2.4. Glyoxylate and Dicarboxylate Metabolism Of 16 metabolites, 2 were detected in the glyoxylate pathway. After RD, glycolic acid increased, while citric acid decreased ([99]Figure 7F). Citrate is a substrate for lipid biosynthesis by ATP-citrate lyase enzyme (ACLY) [[100]20]. The gene alternations indicated that ACLY increased after RD ([101]Figure 7E). 4. Discussion In the present study, a total of 90 significant metabolites were found in a rat model of RD. Pathway enrichment analysis by MetaboAnalyst 5.0 indicated that phenylalanine, tyrosine, and tryptophan biosynthesis; histidine metabolism; purine and phenylalanine metabolism; and glycine, serine, and threonine metabolism were profoundly altered after RD. L-histidine is an essential amino acid and was enzymatically converted to L-histamine, involved in flush sensitivity under dark-adapted conditions [[102]21]. L-histidine was transported to the retina by the L-type amino acid transporter in the retinal capillary endothelial cells [[103]22]. After RD, the blood–retinal barrier (BRB) was disrupted, and the gene of the L-type amino acid transporter increased after RD according to the RNA sequence results ([104]Figure 6). These results indicate that the increase of L-histidine might be because of the disruption of the BRB and the increased transport from capillary endothelial cells by the L-type amino acid transporter. In vasodilation, histamine is an important neurotransmitter as a regulator of ON ganglion cells and microcirculation [[105]22,[106]23,[107]24,[108]25]. The drosophila visual system recycles the histamine as a neurotransmitter between photoreceptors and other cells [[109]24,[110]25]. Decreased histamine could reduce the threshold of ON ganglion cells to scotopic, full-filed, and flash stimuli [[111]21]. Humans with decreased histamine were found to be less sensitive to light stimuli during the day, and the scotopic b-wave was reasonably reduced [[112]22,[113]23]. S-adenosylhomocysteine (SAH) was the product of S-adenosylmethionine (SAM) when DNA methyltransferases catalyzed the transfer of a methyl group [[114]17]. In our study on analyzing gene alternations in human retina tissues, we found that the metabolite of SAH decreased, and the methyltransferases protein called the isoprenylcysteine carboxyl methyltransferase (ICMT) reduced after RD ([115]Figure 6). ICMT deficiency could lead to photoreceptor dysfunction, progressively diminished rod and cone light-mediated responses, and the defect synthesis of the outer segment in the photoreceptors [[116]26]. Based on the results, we inferred that decreased histamine may influence visual function. The reduced ICMT and SAH might be the reason for the dysfunction of the outer segment of the photoreceptor. These results may explain the outer segment dysfunction and decreased vision acuity in patients with retinal detachment. Xanthine oxidase (XO) was considered an important factor for reducing oxygen free radicals after retinal injury [[117]27]. XO was mainly detected in capillary endothelium cells and cones in rabbits and played a role in reducing oxygen to toxic intermediates [[118]28]. In our study, we found that XO gene decreases after RD. The amount of xanthine and hypoxanthine decreased after RD ([119]Figure 7B). As XO could reduce oxygen to toxic intermediates, we inferred that the decreased xanthine and XO might lead to higher oxidation after RD. Guanosine was catalyzed by purine nucleoside phosphorylase (PNP) with the product of guanine [[120]29,[121]30]. In the synthesis of guanine nucleotides, inosine-5′-monophosphate dehydrogenase 1 (IMPDH1) was the rate-limiting step [[122]31]. Our study found that the level of guanosine and guanine reduced after RD and therefore could not provide enough sources for the production of inosine monophosphate (IMP). The genes of IMPDH1, guanosine and guanine, reduced after RD, and a previous study has indicated that reduced guanosine and guanine could lead to the dysfunction of photoreceptors [[123]32]. Systemic administration of guanosine could reduce cell death and inflammation after spinal cord injury [[124]33]. Based on these results, we inferred that the decrease of guanosine, guanine, and IMPDH1 might lead to higher oxidation and consequently photoreceptor degeneration. Tyrosine is a nonessential amino acid and was changed to neurotransmitters such as dopamine, epinephrine, and norepinephrine [[125]34]. It is involved in the activation of various signaling pathways by phosphorylation of the hydroxyl group of proteins called tyrosine kinases, such as VEGF, insulin, and epidermal growth factors [[126]34,[127]35]. On the basis of data from the GEO database, our study also found that tyrosinase-related protein1 (TYR1) increases after RD ([128]Figure 7). In the absence of tyrosine hydroxylase (TH), the neurotransmitters could be produced by tyrosinase [[129]36]. Tyrosinase (TYR) could be expressed by neurons under disease conditions [[130]36]. The tyrosinase-dependent dopaminergic pathway is involved in neurotransmitter and melanin metabolism [[131]37]. Melanin metabolism dysfunction contributes to retinal degeneration [[132]37]. Based on these results, we inferred that the decrease of tyrosine might cause a reduction in the neurotransmitters of retina and consequently lead to photoreceptor degeneration after RD. Citrate can be used as a substrate for lipid production by enzyme of ATP-citrate lyase (ACLY) [[133]23]. Pyruvate could be catalyzed by pyruvate dehydrogenase (PDH) with the product of AcCoA and lactate by PDH (pyruvate dehydrogenase) [[134]38]. Our gene results indicate that PDH and the rate-limiting glycolytic enzyme PFK1 decreased and ACLY increased, but PDH was not altered ([135]Figure 7). Therefore, we inferred that citrate might be used for lipid production after RD. Based on these results, we inferred that the retina might undergo lipid oxidation rather than lipid biogenesis. Citrate is a key intermediate of the tricarboxylic acid (TCA) cycle in the mitochondria and could be used to for ATP production. Reduced citrate may lead to less ATP production [[136]39]. Lipid oxidation and reduced mitochondrial energy metabolism are the main energy changes after RD. However, there were different anatomy structures between human and rats’ retinas. The cone density is higher in human fovea than rats, and more cones were detected in the periphery retina in rats [[137]40]. S-cones were densest in the periphery and retinal rim both in rats and humans [[138]41]. The latest study from Robert J. Casson pointed out that short (S)-cones were more susceptible to damage than medium- or long-wavelength cones in the rat model of retinal detachment [[139]11]. This may explain the reason why S-cones are more vulnerable to damage after intense light exposure, diabetic retinopathy, and other human retinal disease. However, the S-cones and rods were different between human and rats. These results may help us to understand why the macular was prone to be damaged in clinical practice. We adopted metabolism to seek the upstream mechanism for photoreceptor degeneration. However, the structure was different between humans and rats. All these results point to a need for more experiments to verify the pathway in retinal detachment, and further to explore the reason for photoreceptor degeneration. The limitation of the study is that metabolite changes in the animal models were not validated, and the targeted metabolomics were not used to identify the key pathway in retinal detachment. However, we comprehensively analyze the pathway alternations by metabolism and genes from the human retina. These results revealed varying pathological mechanisms involved in retinal detachment, and the neuroprotection of photoreceptors requires multiple approaches. The study helps us to understand the pathology of RD and try to find new targets for neuroprotection. 5. Conclusions In conclusion, our study has revealed the metabolomic profile changes after retinal detachment both in retina tissues of and gene alternations in human retina. Bioinformatics analysis has helped identify histidine metabolism; phenylalanine, tyrosine, and tryptophan biosynthesis; and glycine, serine, and threonine metabolism as the top pathways. All the results indicate varying pathways involved in retinal detachment, such as the decreased histamine metabolomics and xanthine, which may be related to photoreceptor dysfunction, reduced tyrosine metabolomics related to neurotransmitter, and decreased TCA after RD. More research is needed to investigate the key reason and target for neuroprotection. Acknowledgments