Abstract

Purpose

   This study aimed to examine the differential expression profiles of
   plasma metabolites in rat models of post-traumatic osteoarthritis
   (PTOA) and elucidate the roles of metabolites and their pathways in the
   progression of PTOA using bioinformatics analysis.

Method

   Plasma samples were collected from 24 SD female rats to model PTOA, and
   metabolomic assays were conducted. The samples were divided into three
   groups: the surgically induced mild PTOA group (Group A: 3 weeks
   postoperative using the modified Hulth model; age 2 months), the
   surgically induced severe PTOA group (Group B: 5 weeks postoperative
   using the modified Hulth model; age 2 months), and the normal control
   group (Group C: healthy rats aged 2 months). Metabolites were
   structurally identified by comparing the retention times, molecular
   masses, secondary fragmentation spectra, collision energies, and other
   metabolite data with a database (provided by Shanghai Applied Protein
   Technology Co., Ltd.). Target prediction and pathway analysis were
   subsequently performed using bioinformatics analysis.

Results

   The experiment revealed that in the mild PTOA group, levels of
   Alpha-ketoglutarate, Isocitric acid, Dichloroacetate, and other
   metabolites increased significantly compared with the normal group,
   whereas Linolenic acid, Lactose, and others decreased significantly.
   These findings suggest that these metabolites can serve as biomarkers
   for the diagnosis of early PTOA. In the severe PTOA group, Diosgenin,
   Indoleacrylic acid, Alpha-ketoglutarate, Isocitric acid, and others
   were elevated and may also be used as biomarkers for PTOA diagnosis.
   Adrenosterone, (+)-chlorpheniramine, and Phenanthridine levels were
   higher in the severe PTOA group compared to the mild PTOA group, while
   Menadione, Adenosine 5’-monophosphate, and Arg-Gly-Asp levels were
   lower.

Conclusions

   Taurocholate, indoleacrylic acid, alpha-ketoglutarate, and isocitric
   acid may serve as biomarkers for PTOA joint injury in rats. Menadione,
   adenosine 5’-monophosphate, and Arg-Gly-Asp exhibited differential
   expression between severe and mild PTOA groups in rats, potentially
   reflecting the injury’s severity. Further investigation into these
   molecules in human tissues is warranted to ascertain their utility as
   biomarkers for PTOA in humans.

1. Introduction

   Post-traumatic osteoarthritis (PTOA) is a joint disease precipitated by
   year, which has brought a huge and heavy burden to the economy and
   society [[34]1]. The rising incidence of traffic accidents and
   high-energy knee injuries has significantly increased the prevalence of
   PTOA, imposing a substantial economic and societal burden (1). The
   hallmark of PTOA is the degradation and loss of articular cartilage;
   however, it is a comprehensive joint disorder that typically involves
   inflammation, damage to the soft tissues of the joint capsule and
   synovium, hypertrophic bone alterations, and subchondral bone changes
   [[35]2]. PTOA often presents with an asymptomatic phase, which hampers
   early diagnosis and intervention [[36]3]. As the disease advances,
   treatment options and their effectiveness become limited. Timely
   detection and diagnosis, however, could facilitate early intervention
   and potentially halt disease progression [[37]4]. Advancements in
   research, particularly in the analysis of metabolite expression changes
   at disease onset, foster optimism that new technologies and multi-omics
   approaches will enhance future PTOA patient care [[38]5], which is
   strongly tied to an individual’s trauma history and the accurate and
   prompt management of said trauma [[39]6].

   Metabolomics emerged in the late 20th century, as suggested by
   Professor Jeremy Nicholson from Imperial College London [[40]7].
   Metabolomics, an omics field that followed genomics, transcriptomics,
   and proteomics, represents an extension of transcriptomics and
   proteomics, enabling more precise and direct assessments of an
   organism’s physiological state [[41]8]. Currently, metabolomics is
   extensively applied across various research domains. By utilizing
   metabolome technology to analyze metabolic level differences between
   experimental and control groups, researchers can identify differential
   metabolites. This approach is instrumental for biomarker discovery and
   for investigating the biological processes involving these metabolites,
   including the identification of regulatory enzymes and metabolic
   pathways [[42]9]. Consequently, increasing attention is being directed
   toward diverse research areas, such as disease diagnosis, drug target
   identification, nutrition and health management, personalized
   medication, plant growth and development, and stress resistance
   [[43]10]. However, due to variations in study populations and sizes, as
   well as differences in detection methods and ranges, the results
   require further validation. Determining how to translate these findings
   into treatments for PTOA will be the focus of subsequent research.

   Our study aimed to differentially screen the expression profiles of
   various metabolites in plasma from rat models of PTOA. We conducted
   targeted prediction and pathway analysis using bioinformatics to
   elucidate the role of metabolites in the development and progression of
   PTOA.

2. Methods

2.1. Establishment of animal models and extraction of samples

   All experiments and animal care procedures were approved by the Animal
   Experiment Ethics Committee of Changzhi Medical College (No.:
   DW2023036). Plasma samples were collected from 24 SD female rats with a
   PTOA model, and metabolomic assays were conducted. The improved Hulth
   modeling method was chosen for this modeling. The skin was cut on the
   inner side of the right posterior knee joint of the rat, and the soft
   tissue was separated layer by layer into the joint cavity. The medial
   collateral ligament, anterior cruciate ligament, and the medial
   meniscus were completely removed, and the incision was sutured layer by
   layer. The anesthetic (Pentobarbital Merck 25G) was administered
   intraperitoneally at a dose of 45mg/kg body weight. The samples were
   divided into three groups: the surgically induced mild PTOA group
   (Group A: 3 weeks postoperative using the modified Hulth model; age 2
   months), the surgically induced severe PTOA group (Group B: 5 weeks
   postoperative using the modified Hulth model; age 2 months), and the
   normal control group (Group C: normal rats, age 2 months) [[44]11]. The
   method of euthanasia for rats is to use high concentration CO2. The
   container is not pre filled with gas, and CO2 is filled at a rate of
   30% to 70% container volume per minute at an equilibrium rate.
   Post-euthanasia, knee joint cartilage tissue was harvested for tissue
   specimen preparation. Histological analysis was conducted using
   hematoxylin-eosin (HE) staining and safranin O fast green staining to
   assess the pathological grade ([45]Table 1).

Table 1. A cartilage histopathology grade assessment-dgrading methodology.

   Grade (key feature) Associated criteria (tissue reaction)
   Grade 0: surface intact, cartilage morphology intact Matrix: normal
   architecture
   Cells: intact, appropriate orientation
   Grade 1: surface intact Matrix: superficial zone intact, oedema and/or
   superficial fibrillation (abrasion), focal superficial matrix
   condensation
   Cells: death, proliferation (clusters), hypertrophy, superficial zone
   Reaction must be more than superficial fibrillation only
   Grade 2: surface discontinuity As above
   + Matrix discontinuity at superficial zone (deep fibrillation)
   ± Cationic stain matrix depletion (Safranin O or Toluidine Blue) upper
   1/3 of cartilage
   ± Focal perichondronal increased stain (mid zone) ± Disorientation of
   chondron columns
   Cells: death, proliferation (clusters), hypertrophy
   Grade 3: vertical fissures (clefts) As above
   Matrix vertical fissures into mid zone, branched fissures
   ±Cationic stain depletion (Safranin O or Toluidine Blue) into lower 2/3
   of cartilage (deep zone)
   ±New collagen formation (polarized light microscopy, Picro Sirius Red
   stain)
   Cells: death, regeneration (clusters), hypertrophy, cartilage domains
   adjacent to fissures
   Grade 4: erosion Cartilage matrix loss: delamination of superficial
   layer, mid layer cyst formation Excavation: matrix loss superficial
   layer and mid zone
   Grade 5: denudation Surface: sclerotic bone or reparative tissue
   including fibrocartilage within denuded surface. Microfracture with
   repair limited to bone surface
   Grade 6: deformation Bone remodelling (more than osteophyte formation
   only). Includes: microfracture with fibrocartilaginous and osseous
   repair extending above the previous surface
   [46]Open in a new tab

2.2. Sample preprocessing

   The blood was gradually thawed at 4°C and combined with a pre-cooled
   solution of methanol, acetonitrile, and water (2:2:1, v/v). The mixture
   was thoroughly vortexed and then sonicated at low temperatures for
   30 min, followed by incubation at −20°C for 10 min. After
   centrifugation at 14,000 g and 4°C for 20 min, the supernatant was
   vacuum-dried. Subsequently, 100 μL of an acetonitrile-water solution
   (1:1, v/v) was added for mass spectrometry analysis. The sample was
   vortexed, centrifuged at 14,000 g and 4°C for 15 min, and the resulting
   supernatant was subjected to analysis.

2.3. Chromatography—mass spectrometry analysis

   In this experiment, samples were separated using an Agilent 1290
   Infinity LC Ultra Performance Liquid Chromatography System (UHPLC)
   equipped with a HILIC column. The column temperature was set to 25°C,
   the flow rate to 0.5 mL/min, and the sample injection volume to 2 μL.
   The mobile phase consisted of component A: water with 25 mM ammonium
   acetate and 25 mM ammonia, and component B: acetonitrile. The gradient
   elution procedure was as follows: from 0 to 0.5 min, 95% B; from 0.5 to
   7 min, B linearly decreased from 95% to 65%; from 7 to 8 min, B
   linearly decreased from 65% to 40%; from 8 to 9 min, B was held at 40%;
   from 9 to 9.1 min, B linearly increased from 40% to 95%; and from 9.1
   to 12 min, B was held at 95%. The injector temperature was maintained
   at 4°C throughout the analysis. Samples were analyzed in a random
   sequence to minimize fluctuations in the instrument’s detection signal.
   System stability and data reliability were monitored by interspersing
   QC samples within the sample queue.

2.4. Q-TOF mass spectrometry conditions

   The primary and secondary spectra of the sample were collected using
   the AB Triple TOF 6600 mass spectrometer after separating the samples
   by UHPLC. Mass spectrometry was performed with a Triple TOF 6600 mass
   spectrometer (AB SCIEX) and detected in electrospray ionization (ESI)
   positive and negative ion modes, respectively. The ESI source setting
   parameters were as follows: atomized gas auxiliary heating gas 1
   (Gas1): 60, auxiliary heating gas 2 (Gas2): The curtain gas (CUR) was
   set to 30psi, the ion source temperature was set to 600°C, and the
   spray voltage (ISVF) was set to ± 5500 V (plus and minus modes).
   Primary mass-to-charge ratio detection range: 60–1000Da, secondary
   product ion mass-to-charge ratio detection range: 25–1000Da, primary
   mass spectrometry scanning accumulation time: 0.20 s/spectra, secondary
   mass spectrometry scanning accumulation time 0.05 s/spectra; The
   secondary mass spectrum was obtained using data-dependent acquisition
   mode (IDA), and the peak intensity value screening mode was used, the
   declustering voltage (DP): ± 60 V (plus and negative modes), the
   collision energy: 35 ± 15 eV, IDA settings are as follows: Dynamic
   exclusion of isotope ion range: 4 Da, acquisition of 10 fragment maps
   per scan.

2.5. Data analysis process

   The raw data were converted to. mzXML format from Wiff using
   ProteoWizard. Subsequently, XCMS software performed retention time
   correction, peak alignment, and peak area extraction. The data
   underwent preprocessing, metabolite structure identification, and
   quality evaluation. This research method uses metabolites from the
   three major metabolic pathways of carbohydrates, proteins, and lipids
   in the database (provided by Shanghai Applied Protein Technology Co.,
   Ltd.) as the detection catalog. The aim was to identify differentially
   expressed metabolites, including fatty acids, amino acids, nucleotides,
   and benzene compounds. Data analysis encompassed univariate and
   multidimensional statistical analyses, differential metabolite
   screening, correlation analysis of differential metabolites, and KEGG
   pathway analysis.

2.6. Evaluation of experimental data quality

   Comparing the total ion chromatogram (TIC) of QC samples with
   overlapping spectra, the higher the response intensity and retention
   time overlap of each peak, the smaller the variation caused by
   instrument errors throughout the entire experimental process. Perform
   PCA analysis on the peaks extracted from all experimental and QC
   samples. The tighter the clustering of QC samples in positive and
   negative ion modes, the better the reproducibility of the experiment.
   All metabolites identified in this project (metabolites identified by
   combining positive and negative ions) are classified and statistically
   analyzed based on their chemical taxonomy classification information.
   This study used strict Orthogonal Partial Least Squares Discriminant
   Analysis (OPLS-DA) VIP > 1 and P value < 0.05 as criteria to screen for
   differential metabolites, and used multivariate statistical methods
   such as PCA, OPLS-DA analysis, etc. to perform dimensionality reduction
   analysis on the collected multidimensional data while retaining the
   original information to the greatest extent possible. Perform KEGG
   analysis on all differential metabolites.

   In this experiment, we comprehensively evaluated the reproducibility,
   stability of the experimental instrument, and reliability of the data
   quality. Overall, the test demonstrated good reproducibility, the
   analytical system of the experimental instrument was stable, and the
   data were reliable. The differences in metabolite expression detected
   in the assay reflect the biological changes inherent to the sample.

2.7. Statistical analysis

   The statistical software used in this study is SPSS Version 22 (SPSS
   Inc.). Use Spearman rank correlation analysis to examine the
   relationship between two sets of data. The result shows the mean SD. In
   all tests, a P-value less than 0.05 is considered statistically
   significant.

3. Results

3.1. Statistics and analysis of appraisal results

3.1.1. Histological analysis and identification.

   HE and Safranin O Fast Green staining of cartilage tissue from normal,
   mild PTOA, and severe PTOA groups in rat knee joints revealed
   morphological changes. HE staining indicated a progressive decrease in
   chondrocyte number, increased clustering, and reduced columnar
   organization as PTOA severity increased. The cartilage surface layer
   thinned, tissue layers became disordered (superficial, transitional,
   radial, and calcified layers), and the tide line blurred ([47]Fig 1).
   Safranin O Fast Green staining yielded similar results ([48]Fig 2).
   Additionally, we assessed cartilage degeneration on the medial tibial
   plateau using the OARSI-recommended method, classifying the mild PTOA
   group as grade 3 and the severe PTOA group as grade 5[[49]12,[50]13].

Fig 1. HE staining results of cartilage tissue in different groups of rat
PTOA (A: normal group, B: mild PTOA group, C: severe PTOA group).

   [51]Fig 1
   [52]Open in a new tab

Fig 2. Safranin O fast green staining results of cartilage tissue in
different groups of rat PTOA (A: normal group, B: mild PTOA group, C: severe
PTOA group).

   [53]Fig 2
   [54]Open in a new tab

3.1.2.Metabolite identification quantity and classification statistics.

   The metabolite was identified by comparing its molecular mass (with a
   molecular mass error of < 10 ppm), retention time, secondary
   fragmentation spectrum, collision energy, and additional data with
   metabolites in the database (provided by Shanghai Applied Protein
   Technology Co., Ltd). The identification results were subsequently
   verified manually on two separate occasions. The assessment level is
   Level 2 or higher.

   Following the merger of positive and negative ion modes in this
   project, a total of 925 metabolites were identified. Of these, 587 were
   detected in the positive ion mode and 338 in the negative ion mode. All
   identified metabolites, encompassing both positive and negative ions,
   have been categorized and quantified according to their chemical
   classification and associated information. [55]Fig 3 depicts the
   percentage distribution of the various metabolites.

Fig 3. The proportion of identified metabolites in various chemical
classifications.

   [56]Fig 3
   [57]Open in a new tab

3.2. Analysis of inter group differences

   Univariate statistical analysis assesses variability within a group and
   differences between groups for a single variable. In contrast,
   multidimensional statistical analysis evaluates these aspects across
   multiple variables. Metabolomic data, characterized by high
   dimensionality and variable correlation, challenges the effectiveness
   of traditional univariate analysis in extracting meaningful
   information. Consequently, employing statistical techniques like PCA,
   PLS-DA, and OPLS-DA for dimensionality reduction is essential to glean
   potential insights from complex multidimensional data without
   compromising the integrity of the original information.

3.2.1. Univariate statistical analysis.

   Univariate statistical analysis is a commonly used method for analyzing
   data sets. When comparing two samples, popular univariate methods
   include Fold Change Analysis (FC Analysis) and T-tests/nonparametric
   tests. These methods help identify significant differences in the data
   and provide valuable insights for further analysis. Conduct a
   differential analysis on all identified and unidentified metabolites
   detected in both positive and negative ion modes based on univariate
   analysis. Visualize differential metabolites with FC > 1.5 or FC < 0.67
   and p-value < 0.05 in the form of volcano plots. For substances with
   descriptive names that are significantly different, select the top 10
   with increased expression changes and the top 10 with decreased
   expression changes for labeling. Display the results in [58]Fig 4.

Fig 4. a. Positive ion mode volcanic map (color and up-regulation correlation
with differential metabolites) (A) Mild PTOA Group VS Severe PTOA Group (B)
Mild PTOA Group VS Normal Group (C) Severe PTOA Group VS Normal Group. b.
Negative ion mode volcanic map (correlation between color and up-regulation
of differential metabolites) (A) Mild PTOA Group VS Severe PTOA Group (B)
Mild PTOA Group VS Normal Group (C) Severe PTOA Group VS Normal Group.

   [59]Fig 4
   [60]Open in a new tab

   The x-axis in the graph shows the logarithmic value of log2 for the
   fold change, while the y-axis represents the logarithmic value of
   -log10 for the significance p value. Metabolites that show significant
   differences are shown in rose red if they meet the criteria of FC >1.5
   and p value <0.05. Metabolites that meet the criteria of FC < 0.67 and
   p value < 0.05 are shown in blue. Metabolites that are not
   significantly different are shown in black. The marked metabolites in
   the graph are the ones that show significant differences, and they are
   labeled with their qualitative names. The top 10 upregulated and top 10
   downregulated metabolites are indicated with their respective fold
   change values (FC).

3.2.2. Multidimensional statistical analysis.

   Principal Component Analysis (PCA) linearly combines all metabolites to
   form new comprehensive variables, achieving dimensionality reduction.
   The PCA model, obtained through 7-fold cross-validation, demonstrated
   high cohesion within each group and significant intergroup separation,
   indicating the model’s reliability and the presence of significant
   differences between groups. Partial Least Squares Discriminant Analysis
   (PLS-DA) uses partial least squares regression to relate metabolite
   expression levels to sample categories, enabling category prediction.
   Model evaluation parameters from lesions revealed that most models had
   a Q^2 greater than 0.5, suggesting stability and reliability. To
   prevent overfitting in the supervised model, we performed a permutation
   test. We noted a gradual decrease in both R^2 and Q^2 with reduced
   permutation retention, confirming the absence of overfitting and
   demonstrating the model’s stability. Orthogonal Partial Least Squares
   Discriminant Analysis (OPLS-DA), an enhancement of PLS-DA, filters out
   unrelated noise to improve analytical ability and effectiveness. The
   OPLS-DA model also showed most Q^2 values above 0.5, indicating
   stability and reliability. Permutation tests on each model confirmed no
   overfitting, and [61]Fig 5 illustrates the model’s robustness.

Fig 5. a. Positive ion multi-dimensional analysis chart (A) Mild PTOA group
VS Severe PTOA group (B) Mild PTOA group VS normal group (C) Severe PTOA
group VS Normal group (a) 3D PCA score chart (b) PLS-DA score chart (c)
PLS-DA replacement test (d) OPLS-DA score chart (e) OPLS-DA replacement test.
b. Multidimensional analysis of negative ions (A) Mild PTOA group VS Severe
PTOA group (B) Mild PTOA group VS normal group (C) Severe PTOA group VS
normal group (a) 3D PCA score chart (b) PLS-DA score chart (c) PLS-DA
replacement test (d) OPLS-DA score chart (e) OPLS-DA replacement test.

   [62]Fig 5
   [63]Open in a new tab

3.3. Screening for differential metabolites

   The OPLS-DA model assesses the impact and explanatory capacity of each
   metabolite’s expression pattern on the classification and
   discrimination of sample groups using Variable Importance for the
   Projection (VIP). It identifies biologically significant differential
   metabolite molecules. Metabolites with an OPLS-DA VIP > 1 and a P
   value < 0.05 are considered to exhibit significant differences,
   according to the metabolite screening criteria. A bar chart visually
   represents the identified significant metabolic alterations, as
   depicted in [64]Fig 6, [65]Table 2, and [66]Table 3.

Fig 6. a. Positive ion mode significant difference metabolite expression
difference factor analysis (A) mild PTOA group VS severe PTOA group (B) mild
PTOA group VS normal group (C) Severe PTOA group VS normal group. b. Negative
ion mode significant difference metabolite expression difference factor
analysis (A) mild PTOA group VS severe PTOA group (B) mild PTOA group VS
normal group (C) Severe PTOA group VS normal group.

   [67]Fig 6
   [68]Open in a new tab

   The abscissa in the figure indicates the difference expression
   multiple, red indicates that the difference expression multiple is
   greater than 1, and green indicates that the difference expression
   multiple is less than 1. Ordinates indicate significantly differential
   metabolites.

Table 2. Negative ion mode significant differential metabolites (Fold Change
greater than 2 or less than 0.5).

   Group Metabolite name
   A vs B UP Phenol 1-octadecanoyl-sn-glycero-3-phospho
   DOWN Taurocholate –
   A vs C UP 1-octadecanoyl-sn-glycero-3-phospho-(1’-myo-inositol) Phenol
   Alpha-ketoglutarate Dichloroacetate Isocitric acid
   DOWN Cis-9-palmitoleic acid Linolenic acid
   Cis-4,7,10,13,16,19-docosahexaenoic acid
   B vs C UP Indoleacrylic acid Indolelactic acid Taurocholate
   N-Acetyltryptophan 4-Hydroxymandelate
   DOWN –
   [69]Open in a new tab

   NA: not available

Table 3. Positive ion mode significant differential metabolites (Fold Change
greater than 2 or less than 0.5).

   Group Metabolite name
   A vs B UP Tuberostemonine Pro-Trp Adenosine 5’-monophosphate
   Arg-Gly-Asp Menadione
   DOWN Phenanthridine Trans-4-(aminomethyl) cyclohexanecarboxylic acid
   (+)-chlorpheniramine Adrenosterone –
   A vs C UP gamma.-L-Glu-.epsilon.-L-Lys Gamma-l-glutamyl-l-glutamic acid
   2-aminoadipic acid 7-chloro-3-methylquinoline-8-carboxylic acid
   N6-(1-iminoethyl)-l-lysine
   DOWN N.alpha.-Acetyl-L-lysine Phenanthridine Anserine 2-furanpropanoic
   acid, 3-carboxy-4-methyl-5-propyl- Lactose
   B vs C UP Diosgenin Oxyquinoline 7.alpha.-hydroxy-3-oxo-4-cholestenoic
   acid –
   DOWN N.alpha.-Acetyl-L-lysine DL-2,4-Diaminobutyric acid Arg-Gly-Asp
   L-homocitrulline 2-furanpropanoic acid, 3-carboxy-4-methyl-5-propyl-
   Adenosine 5’-monophosphate
   [70]Open in a new tab

   NA: not available

   In this study, we found that the expression of
   gamma-L-Glu-epsilon-L-Lys and Gamma-l-glutamyl-l-glutamic acid was
   significantly upregulated in mild PTOA compared to the normal group.
   Additionally, 2-aminoadipic acid, N.alpha.- Acetyl-L-lysine,
   Phenanthridine, and Anserine were significantly down-regulated. In the
   negative ion mode, 1-octadecanoyl-sn-glycero-3-phospho-(1 ‘-myo
   -inositol), Phenol, and Alpha-ketoglutarate were also down-regulated.
   Down-regulated expressions include Cis-9-palmitoleic acid, Linolenic
   acid, and Linolenic acid. In severe PTOA, compared to the normal group,
   the expressions of Diosgenin, Oxyquinoline, and
   7.alpha.-hydroxy-3-oxo-4-cholestenoic acid were significantly
   increased, while the expressions of N.alpha.-Acetyl-L,
   DL-2,4-Diaminobutyric acid, and Arg-Gly-Asp were significantly
   decreased. In the negative ion mode, the expressions of Indoleacrylic
   acid, Indolelactic acid, and Taurocholate were significantly
   up-regulated.

3.4. Cluster analysis

   To clearly demonstrate the connections and distinctions among the
   metabolite expression patterns of the samples, we subtracted the
   group’s average value from the expression levels of all samples and
   differential metabolites. We standardized the results by dividing them
   by the root mean square of the group number. Subsequently, we
   calculated the distance matrix and employed hierarchical clustering for
   the analysis. The results of the hierarchical cluster analysis for
   metabolites with significant differences (VIP > 1, p-value < 0.05) are
   displayed in [71]Fig 7. Metabolites clustered within the same group
   exhibit similar expression patterns, suggesting potential shared
   functionalities and involvement in corresponding metabolic processes or
   cellular pathways.

Fig 7. a. Positive ion mode, significant difference, metabolite level
clustering heat map (A), mild PTOA group VS severe PTOA group, (B) mild PTOA
group VS normal group, (C) severe PTOA group VS normal group. b. Negative ion
pattern significant difference metabolite level clustering heat map (A) Mild
PTOA group VS severe PTOA group (B) Mild PTOA group VS normal group (C)
Severe PTOA group VS normal group.

   [72]Fig 7
   [73]Open in a new tab

   The figure illustrates differential metabolites, with each row
   representing a specific metabolite showing significant differential
   expression. The vertical axis depicts the metabolite, while the
   horizontal axis represents different sample groups. The color red
   indicates significant upregulation, while blue signifies significant
   downregulation, with varying shades reflecting the degree of
   upregulation or downregulation. Similar expression patterns of
   metabolites are clustered together on the left side of the figure.

3.5. Correlation analysis

   Metabolites with correlated expression may jointly participate in a
   biological process, termed functional correlation. Moreover,
   metabolites can interact synergistically or antagonistically. For
   instance, metabolites exhibiting the same trend of change are
   positively correlated, while those with opposing trends indicate a
   negative correlation. Positive correlations may imply a shared
   synthetic pathway, whereas negative correlations could indicate their
   use in synthesizing other metabolites. We employed correlation analysis
   to examine metabolites with significant differences. Using a comparison
   group as an example, we presented the results through a correlation
   heatmap in [74]Fig 8.

Fig 8. a Positive ion mode correlation heat map (A) Mild PTOA group VS Severe
PTOA group (B) Mild PTOA group VS normal group (C) Severe PTOA group VS
normal group. b. Negative ion mode correlation heat map (A) Mild PTOA group
VS Severe PTOA group (B) Mild PTOA group VS normal group (C) Severe PTOA
group VS normal group.

   [75]Fig 8
   [76]Open in a new tab

   In this visualization, the color scheme helps us understand the
   relationship between variables. Red signifies a positive correlation,
   blue indicates a negative correlation, and white indicates no
   significant correlation. The intensity of the color reflects the
   strength of the correlation coefficient - darker colors represent
   higher degrees of positive or negative correlation. Additionally, the
   size of the data points conveys the significance of the correlation. A
   smaller p-value, indicating higher significance, is represented by
   smaller data points, whereas a larger point size represents lower
   significance.

3.6. KEGG Enrichment and Analysis

   The Kyoto Encyclopedia of Genes and Genomes (KEGG) database is widely
   used for pathway research. It includes pathway information from various
   perspectives, such as metabolism, genetic and environmental information
   processing, cellular processes, biological systems, human diseases, and
   drug development. Different metabolites interact within organisms to
   perform biological functions. Analyzing the KEGG pathway offers deeper
   insights into these functions. To annotate and analyze the KEGG
   pathway, one must first combine the differential metabolites identified
   from both positive and negative ion modes. KEGG pathway enrichment
   analysis employs the KEGG pathway as a unit, considering the metabolic
   pathways present in the species or those closely related. The study
   used Fisher’s Exact Test to evaluate and quantify the significance
   level of metabolite enrichment in each pathway, aiming to identify
   metabolic and signal transduction pathways significantly affected. A
   lower P-value indicates a more substantial difference in metabolic
   pathways. [77]Fig 9 displays the results of the metabolic pathway
   enrichment analysis.

Fig 9. KEGG enrichment pathway chart (bubble chart) (A) mild PTOA group VS
severe PTOA group (B) mild PTOA group VS normal group (C) Severe PTOA group
VS normal group.

   [78]Fig 9
   [79]Open in a new tab

   In a bubble chart, each bubble represents a metabolic pathway. The
   selection criteria for these pathways is based on their p-values, with
   only the top 20 pathways with the highest significance being included.
   The horizontal placement and size of each bubble in the chart represent
   the influence of the pathway in topology analysis. The larger the
   bubble, the greater the influence of the pathway. The vertical
   placement and color of each bubble represent the p-value of enrichment
   analysis. The p-values are converted to their negative logarithm
   (-log10 p-value) and the darker the color, the smaller the p-value. A
   smaller p-value indicates a higher degree of enrichment and therefore
   greater significance in the analysis. Furthermore, the rich factor of
   each pathway represents the proportion of differential metabolites
   within the pathway, compared to the total number of metabolites
   annotated in the pathway.

   Differential Abundance Score is a method for analyzing metabolic
   changes based on pathways, capturing the average and overall shifts in
   all metabolites within a specific pathway. Significantly enriched
   metabolic pathways are classified and attributed according to their
   higher-level Pathway Hierarchy, as depicted in [80]Fig 10.

Fig 10. Differential abundance scores of differential metabolic pathways
(classified according to Pathway_Hierarchy) (A) mild PTOA group VS severe
PTOA group (B) mild PTOA group VS normal group (C) severe PTOA group VS
normal group.

   [81]Fig 10
   [82]Open in a new tab

   The figure's Y-axis represents the name of the differential pathway,
   while the X-axis represents the DA score. The DA score reflects the
   overall change in all metabolites within the metabolic pathway. A score
   of 1 indicates an increase in the expression of all identified
   metabolites in the pathway, while a score of −1 indicates a decrease in
   their expression. The length of the line segment represents the
   absolute value of the DA score, and the size of the dot at the end of
   the line segment indicates the number of metabolites in the pathway. A
   larger dot signifies a higher number of metabolites. The color
   intensity of the line segments and dots corresponds to the value of the
   DA score. The darker red indicates a higher likelihood of overall
   upregulation in the pathway's expression, whereas the darker blue
   denotes a higher likelihood of overall downregulation.

4. Discussion

   Osteoarthritis is a primary contributor to joint disorders, affecting
   approximately half of the population according to statistical
   data[[83]14]. The incidence of traffic accidents and sports injuries
   has recently increased, leading to a rise in cases of traumatic
   arthritis. This condition results in secondary thickening, hardening,
   and degeneration of joint cartilage, with the trend showing an annual
   increase. Traumatic osteoarthritis is a type of osteoarthritis that
   makes up about 10% of all cases of osteoarthritis[[84]15]. Traumatic
   osteoarthritis, which accounts for about 10% of all osteoarthritis
   cases (15), shares some pathophysiological changes with osteoarthritis;
   however, a complete understanding of their specific pathogenesis
   remains elusive. Traumatic arthritis typically stems from joint
   injuries, but its early identification is challenging. Consequently,
   clinical symptoms fail to improve due to the absence of early
   intervention.

   Metabolomics encompasses the dynamic alterations in metabolites
   resulting from pathophysiological disturbances, biological stimuli,
   environmental shifts, and genetic variations. Metabolites not only
   engage in enzyme-catalyzed chemical reactions but also play crucial
   roles in cellular functions, including energy production and storage,
   signal transduction, and apoptosis [[85]16]. Additionally, they serve
   as both substrates and products of cellular activities. While the host
   can produce metabolites directly, external factors such as diet and
   stress can significantly influence their levels [[86]17]. Due to its
   high sensitivity to initial biological damage, which triggers a cascade
   of metabolic responses in organisms, metabolomics contains a wealth of
   information [[87]18]. As a relatively downstream discipline within the
   “omics” hierarchy, its ultimate aim is to elucidate the collective
   impact of biological and environmental changes on the profile of low
   molecular weight molecules in organisms.

   In plasma samples, differential metabolites identified by positive ion
   screening–2-aminoadipic acid, Gamma-l-glutamyl-l-glutamic acid,
   phenanthridine, and lactose–were differentially expressed between mild
   PTOA and normal groups. Both 2-aminoadipic acid and
   Gamma-l-glutamyl-l-glutamic acid showed upregulated expression, while
   phenanthridine and lactose were downregulated. Differential metabolites
   alpha-ketoglutarate, isocitric acid, and dichloroacetate, identified by
   negative ion screening, were upregulated in mild PTOA compared to
   normal groups. Indoleacrylic acid, indolelactic acid, and taurocholate
   were differentially expressed in severe PTOA versus normal groups, with
   all expressions being upregulated.

   Studies indicate that elevated cholesterol levels in cartilage cells
   impede the association with chondrogenesis genes, leading to
   osteoarthritis[[88]19]. Taurocholate may lower cholesterol in the
   chondrocyte membrane, thus enhancing the activation of cartilage marker
   genes and playing a role in OA treatment by promoting chondrocyte
   proliferation [[89]20]. Our research revealed that taurocholate levels
   were higher in the severe OA group than in the normal group, suggesting
   that taurocholate is not only a biomarker for OA diagnosis but also has
   therapeutic potential. Similarly, indoleacrylic acid and indolelactic
   acid showed an increasing trend in the severe OA group compared to the
   normal group. These compounds, typically derived from tryptophan
   metabolites, have been demonstrated to significantly affect human
   health and disease[[90]21]. Diosgenin, a steroidal saponin with
   anti-inflammatory properties, has been shown to reduce mitochondrial
   oxidative stress damage and apoptosis in OA chondrocytes by activating
   the JAK2/STAT3 pathway, thereby offering protection to OA
   chondrocytes[[91]22]. An upward trend of diosgenin was also observed in
   the severe OA group relative to the normal group.

   Preventing excessive oxidation of mitochondrial peroxiredoxin(PRX) is
   crucial for preserving cellular redox homeostasis and signaling, as
   observed in chondrocytes in vivo[[92]23]. High oxidation of PRX can
   impair peroxidase function and increase intracellular ROS levels,
   contributing to oxidative stress. Menadione has been shown to induce
   chondrocyte changes through redox cycling, leading to excessive
   cellular ROS production and subsequent PRX oxidation[[93]24]. This
   effect is particularly pronounced in older individuals due to their
   reduced antioxidant capacity[[94]25]. In this study, lower menadione
   levels were observed in the severe OA group compared to the mild OA
   group. The tricarboxylic acid cycle, or Krebs cycle, is essential for
   regulating mitochondrial function and controlling oxidative stress in
   eukaryotes. Alpha-ketoglutarate and isocitric acid are two key
   intermediate metabolites in this cycle [[95]26]. Liang Liu et al. found
   that alpha-ketoglutarate can mitigate oxidative stress in chondrocyte
   mitochondria and promote chondrocyte autophagy. It can also decrease
   inflammatory responses in vivo and in vitro, enhance the extracellular
   matrix, and promote chondrocyte proliferation while inhibiting
   apoptosis[[96]27]. They suggest that alpha-ketoglutarate is a potential
   therapeutic target for OA. In our experiment, we also observed that the
   expression levels of both metabolites in the mild OA group trended
   upward compared to the normal group. This may represent a
   self-protective response, enhancing the energy supply by boosting the
   tricarboxylic acid cycle in OA chondrocytes [[97]28].

5. Conclusions

   A wide array of metabolites differ in the blood among mild PTOA, severe
   PTOA, and normal rats. Taurocholate, indoleacrylic acid,
   alpha-ketoglutarate, and isocitric acid could serve as biomarkers for
   PTOA diagnosis in rats. Menadione, adenosine 5’-monophosphate, and
   Arg-Gly-Asp exhibited differential expression between severe and mild
   PTOA groups in rats, potentially correlating with PTOA severity. These
   findings may also apply to humans.

6. Limitations

   However, this study has the following limitations: 1) The study did not
   directly screen differential miRNAs in human blood. Given the
   significant individual differences among PTOA patients in clinical
   practice, this omission could impact the results. In contrast, the rat
   model allows for better control of variables such as age and weight,
   yielding more stable outcomes; 2) Instead of identifying a specific
   metabolite or a few, the study found more than a dozen. This could be
   due to the small sample size or extraneous factors. Future research
   with larger sample sizes may refine the findings and yield more precise
   results; 3) Some outcomes did not align with expectations, which may
   also be attributable to the limited sample size; 4) The final results
   require further validation in human blood.

Data Availability

   The datasets generated during the current study are available in the
   Open Science Framework repository,
   [[98]https://doi.org/10.17605/OSF.IO/7CYDA].

Funding Statement

   This study was supported by the Scientific Research Project of Changzhi
   Health Commission (2023-16). The funding body had no role in the design
   of the study and collection, analysis, and interpretation of data and
   in writing the manuscript.

References