Analysis of Plasma Biomarkers and Potential Pathogenesis in Children with Mycoplasma pneumoniae Pneumonia Based on Metabolomic Technology

Abstract

Background

Mycoplasma pneumoniae pneumonia (MPP) is a common cause of community-acquired pneumonia in children, yet its metabolic alterations and pathogenic mechanisms remain incompletely understood. This study aimed to characterize the plasma metabolomic profile of children with MPP and explore associated metabolic pathways.

Methods

This retrospective case-control study involved 42 children including 14 MPP patients, 14 non-MPP (MPN) patients, and 14 healthy controls (HC). Clinical and laboratory data were compared across groups. Plasma metabolomic profiling was performed using liquid chromatography-tandem mass spectrometry (LC-MS/MS). Multivariate statistical analyses, including principal component analysis (PCA), partial least squares-discriminant analysis (PLS-DA), and orthogonal partial least squares-discriminant analysis (OPLS-DA), were employed to identify metabolic differences among groups. Pathway enrichment analysis used the KEGG and Reactome databases.

Results

Clinically, the MPP group exhibited significantly higher C-reactive protein (CRP) levels and neutrophil percentages, and lower lymphocyte percentages (all P<0.05), indicating enhanced inflammatory response. Distinct plasma metabolic profiles were observed in the MPP group compared with the MPN and HC groups. A total of 127 differential metabolites were identified in MPP vs MPN, and 749 in MPP vs HC. Pathway enrichment analysis revealed that downregulated metabolites in MPP vs MPN were significantly enriched in the PPAR signaling pathway, linoleic acid metabolism, and vitamin digestion/absorption pathways. In MPP vs HC, downregulated metabolites were mainly enriched in amino acid metabolism (arginine biosynthesis, alanine-aspartate-glutamate metabolism), GABAergic synapse, and glycerophospholipid metabolism. These perturbations aligned with the enhanced inflammatory responses observed in MPP.

Conclusion

Childhood MPP is characterized by extensive metabolic disturbances involving amino acid, lipid, vitamin, and neurotransmitter metabolism. These metabolic disturbances may be associated with altered antioxidant defense, inflammatory signaling and neuroimmune crosstalk in MPP. The identified metabolites and dysregulated pathways may represent preliminary exploratory candidate signatures that require further validation to establish their relevance to childhood MPP.

Introduction

Children exhibit heightened susceptibility to respiratory infectious diseases owing to the anatomical and physiological immaturity of the respiratory tract and incomplete immune development.¹ Among these diseases, Mycoplasma pneumoniae pneumonia (MPP) has become a major global public health threat that particularly affects children. Infections caused by Mycoplasma pneumoniae (MP) typically follow a cyclical pattern every 3–5 years.² Notably, pediatric MP infections have significantly increased post the COVID-19 pandemic, resulting in a higher hospitalization rate.³ MP, a cell wall-deficient prokaryotic microorganism, is a primary cause of community-acquired pneumonia (CAP) in children. A study in China revealed that 56.9% of pathogen-positive cases in 1,070 hospitalized children with acute lower respiratory tract infections were attributed to MP, highlighting its significant pathogenicity.⁴ MPP is characterized by a slow onset, prolonged duration, and a higher likelihood of recurrence compared to pneumonia caused by other pathogens.⁵ Apart from typical respiratory symptoms like cough, fever, and wheezing, MPP can also affect various extrapulmonary organs, including the cardiovascular, nervous, and digestive systems in severe cases. A broad spectrum of extrapulmonary manifestations may occur and can even emerge without obvious pulmonary lesions. Additionally, MPP can lead to severe complications such as severe pneumonia, diffuse alveolar hemorrhage, cavitary lesions, acute respiratory distress syndrome, and bronchiolitis obliterans.^6,7 Both respiratory complications and extrapulmonary disorders constitute major adverse outcomes of MP infection. These complications pose substantial long-term risks to the health, growth, and development of children.⁸

Currently, the clinical diagnosis of MPP relies on etiological tests (such as culture methods and nucleic acid detection), serological tests (such as MP-IgM antibody detection), and imaging examinations. However, these methods have limitations. Pathogen culture is time-consuming (7–14 days) and has low sensitivity, making it unsuitable for rapid diagnosis.⁵ The reliability of nucleic acid detection depends on specimen quality and detection reagents.⁹ Serological tests have a window period, making it impossible to distinguish between acute and chronic infections.¹⁰ Imaging examinations lack specificity.¹¹ While macrolide antibiotics were previously the preferred treatment for MPP, the emergence of drug-resistant strains has led to less effective treatment and longer illness durations for some patients.¹⁰ Thus, exploring MPP pathogenesis, identifying specific candidate biomarkers and therapeutic targets, and optimizing clinical diagnosis and treatment strategies are urgently needed to improve the prognosis of pediatric MPP patients.

Metabolomics, a crucial element of systems biology, reveals metabolic disorders in the body under various conditions by analyzing endogenous metabolites. This method presents a novel perspective on disease research, helping to understand pathogenesis, diagnosis, and treatment.^12,13 Unlike genomics, transcriptomics, and proteomics, metabolomics closely reflects the organism’s phenotype, capturing the dynamic changes in metabolic networks following pathogen infection and host-pathogen interactions.¹⁴ Widely used in infectious disease research for its high efficiency, sensitivity, and comprehensiveness, metabolomics has greatly advanced the elucidation of pathogenic mechanisms, identification of candidate biomarkers, and assessment of treatment efficacy.^15–17 For instance, in studies on tuberculosis and influenza virus pneumonia,^16,17 metabolomics has effectively identified metabolic biomarkers with diagnostic significance, enhancing precision in disease diagnosis and treatment.

Diagnosing and treating MPP present clinical challenges that highlight the importance of utilizing metabolomics. Currently, metabolomic research on MPP remains in its early stages, characterized by small sample sizes and a lack of systematic analysis of metabolic features in MPP children, particularly in distinguishing metabolic variations from other infections. This preliminary exploratory study utilizes untargeted metabolomics to analyze plasma metabolite profiles in children with MPP, non-MPP, and healthy controls. We aim to identify plasma metabolic signatures for differentiating MPP from other community-acquired pneumonia and screen candidate diagnostic biomarkers, while revealing MPP-related metabolic disturbances to offer preliminary clues for individualized treatment and improve clinical prognosis of pediatric MPP.

Materials and Methods

Study Design and Participants

This retrospective case-control study was performed in the pediatric ward of Suzhou Hospital, the Affiliated Hospital of Nanjing University Medical School, from November 2023 to January 2024. The study protocol was approved by the Institutional Ethics Committee (Approval No. IRB2023060). Written informed consent was obtained from the parents or legal guardians of all participants in accordance with the Declaration of Helsinki.

A total of 42 children were enrolled in this study, including 28 patients with pneumonia and 14 healthy controls (HC). The patients with pneumonia were further divided into two groups: the MPP group (n=14) and the non-MPP (MPN) group (n=14). The diagnosis of MPP was established according to the following criteria: (i) fulfillment of the clinical diagnostic criteria for CAP;¹⁸ (ii) evidence of MP infection, defined as positive targeted next-generation sequencing (tNGS) results together with a four-fold increase in MP-specific IgM antibody titer between paired acute- and convalescent- phase serum samples; and (iii) no antibiotic or glucocorticoid exposure before admission. Patients in the MPN group fulfilled the clinical and radiological criteria for CAP but showed negative results for both MP tNGS and serological IgM testing. The pneumonia etiologies of children in the MPN group included other common respiratory pathogens such as bacteria and viruses, excluding Mycoplasma pneumoniae. HC (n = 14) were recruited from children who underwent routine physical examinations at the same hospital during the same period and had no respiratory tract infection within the previous month. Participants were excluded if they had (i) underlying chronic diseases, including congenital heart disease, bronchial asthma, immunodeficiency, or chronic hepatic or renal disease; (ii) a history of immunosuppressive therapy; (iii) malignancy or recent exposure to cytotoxic agents; (iv) incomplete clinical information; and (v) documented co-infection with other respiratory pathogens based on available clinical microbiological testing.

Clinical Data and Sample Collection

Clinical data for all participants, including demographic characteristics, medical history, clinical manifestations, diagnosis information, laboratory findings, and treatment records, were extracted from the electronic medical records system.

All blood samples were collected on the second morning of admission after overnight fasting. Peripheral venous blood (3–5 mL) was collected from each participant into EDTA anticoagulant tubes prior to antibiotic or glucocorticoid administration. The samples were then centrifuged at 3,000 × g for 10 minutes at 4°C to obtain the plasma, which was aliquoted and stored at −80°C until metabolomic analysis. All samples were processed within 2 hours of collection, and repeated freeze-thaw cycles were avoided.

Metabolomic Profiling

Plasma metabolites were extracted by mixing 100 μL of thawed plasma with 400 μL of methanol/acetonitrile (2:1, v/v) containing isotopically labeled internal standards. The mixture was vortexed, ultrasonicated in an ice-water bath for 10 min, and incubated at −40°C overnight to precipitate proteins. After centrifugation at 12,000 rpm for 20 min at 4°C, the supernatant was analyzed using a Waters ACQUITY UPLC I-Class Plus system coupled with a Thermo Q Exactive HF mass spectrometer. Chromatographic separation was performed on an ACQUITY UPLC HSS T3 column (100 mm × 2.1 mm, 1.8 μm) maintained at 45°C. The mobile phases consisted of 0.1% formic acid in water (A) and acetonitrile (B), delivered at a flow rate of 0.35 mL/min with a gradient elution from 5% to 100% B over 16 min. Mass spectrometry data were acquired in data-dependent acquisition mode using a heated electrospray ionization source. The spray voltages were set at 3.8 kV in positive ion mode and −3.2 kV in negative ion mode, with a capillary temperature of 320°C. Full-scan MS data were acquired over an m/z range of 70–1050 at a resolution of 60,000. Quality control (QC) samples, prepared by pooling equal aliquots of all study samples, were injected at regular intervals throughout the analytical run to monitor instrument stability and analytical reproducibility.

Statistical Analysis

Raw liquid chromatography-mass spectrometry (LC-MS) data were preprocessed using XCMS software (version 4.6.3) for peak detection, alignment, and retention time correction. Metabolites were annotated by matching retention time, accurate mass, and MS/MS spectra against public databases, including HMDB, LIPID MAPS, and METLIN, as well as an in-house standard library.

Compound annotations were generated by integrating accurate mass, isotope-distribution similarity, MS/MS fragmentation matching and, when available, retention-time matching against reference standards. Annotation confidence was classified using a four-level system: Level 1, retention-time deviation within ±0.3 min and fragmentation score ≥45; Level 2, retention-time deviation within ±0.3 min and fragmentation score <45; Level 3, fragmentation score ≥45 without retention-time confirmation; and Level 4, fragmentation score <45 without retention-time confirmation. Unless confirmed using an authentic standard analyzed under identical LC-MS conditions, Level 2–4 annotations were regarded as putative.

To ensure data quality, features with a relative standard deviation of > 30% in QC samples or with more than 50% missing values within any group were excluded. Missing values were imputed using half of the minimum detected value, and the resulting data matrix was log2-transformed to normalize the distribution.

Multivariate statistical analyses were performed to characterize metabolic alterations. Unsupervised principal component analysis (PCA) was first conducted to assess the overall distribution and the stability of QC. Subsequently, supervised partial least squares-discriminant analysis (PLS-DA) was utilized to characterize the global metabolic patterns and maximize group separation among the three groups. To identify specific differential metabolites, orthogonal partial least squares-discriminant analysis (OPLS-DA) was conducted for pairwise comparisons. The robustness of the OPLS-DA model was evaluated using a 7-fold cross-validation and a permutation test with 200 iterations to assess potential overfitting. Differential metabolites were identified based on the following criteria: variable importance in projection (VIP) > 1, fold change (FC) ≥ 1.2 or ≤ 0.83, and P < 0.05, as determined by Student’s t-test or one-way analysis of variance (ANOVA).¹⁹ To evaluate the robustness of the exploratory findings to multiple testing, P values were adjusted separately for each pairwise comparison using the Benjamini-Hochberg false discovery rate procedure. Adjusted q values are reported in Supplementary Tables S1–S3 but were not used as an additional candidate-selection criterion. As this is an exploratory untargeted metabolomics study, strict multiple testing adjustments such as Bonferroni correction are highly conservative and may exclude biologically meaningful metabolic signals. Complementary multivariate modeling and pathway enrichment analyses were adopted to reduce false positive findings.

Biological pathway enrichment was performed using hypergeometric testing based on the KEGG and Reactome databases. For the description of clinical baseline characteristics, continuous variables were presented as the mean ± standard deviation (SD) or median with interquartile range (IQR) depending on normality, while categorical variables were presented as numbers and percentages. All statistical analyses were conducted using R software (version 4.5.0), and a two-sided P < 0.05 was considered statistically significant.

Results

Baseline Characteristics of the Study Population

A total of 42 children were enrolled in this study, including 14 patients with MPP, 14 patients with MPN, and 14 HCs. The HC group was matched for age and sex and served as the reference group for subsequent metabolomic analyses. The demographic and clinical characteristics of the two pneumonia groups are summarized in Table 1. No significant differences were observed between the MPN and MPP groups in age, sex distribution, height, or weight (all P > 0.05), indicating comparable baseline characteristics. Clinically, both groups presented with common manifestations of pneumonia, including fever and cough, with no significant differences in wheezing, crackles, or the length of hospital stay (all P > 0.05). However, significant differences were identified in several laboratory parameters. Compared with the MPN group, the MPP group had significantly higher neutrophil percentages (66.9 ± 13.4% vs 54.1 ± 15.2%, P = 0.041) and CRP levels [6.05 (4.40–15.90) mg/L vs 0.73 (0.50–9.86) mg/L, P = 0.027], but lower lymphocyte percentages (25.2 ± 11.4% vs 38.7 ± 15.9%, P = 0.015). Other parameters, including total white blood cell (WBC) count, hemoglobin, platelet (PLT) count, procalcitonin (PCT), and liver function indices (alanine aminotransferase [ALT], aspartate aminotransferase [AST]), showed no significant differences between the two groups.

Table 1.

Demographic and Clinical Characteristics of Children with MPN and MPP

Characteristics	MPN Group (n=14)	MPP Group (n=14)	HC Groups (n=14)	p-Value
Demographics
Age, years	6.7 ± 2.0	8.3 ± 3.0	8.0 ± 1.3	0.168
Sex, Male/Female	10/4	9/5	8/6	0.733
Height, cm	124.7 ± 14.9	132.9 ± 20.7	130.3 ± 7.9	0.460
Weight, kg	25.4 ± 7.2	33.1 ± 15.2	27.9 ± 4.6	0.360
Clinical manifestations
Fever, n (%)	12 (85.7)	14 (100.0)	-	0.142
Cough, n (%)	13 (92.9)	14 (100.0)	-	0.309
Wheezing, n (%)	0 (0.0)	1 (7.1)	-	0.309
Crackles, n (%)	7 (50.0)	5 (35.7)	-	0.445
Hospital Length, days	5.50 (4.25–7.00)	6.00 (5.00–7.00)	-	0.314
Laboratory Findings
WBC, 10^9/L	8.9 ± 2.3	9.0 ± 3.0	-	0.909
Neutrophils, %	54.1 ± 15.2	66.9 ± 13.4	-	0.041
Lymphocytes, %	38.7 ± 15.9	25.2 ± 11.4	-	0.015
Hemoglobin, g/L	126.6 ± 12.0	127.6 ± 8.1	-	0.713
PLT, 10^9/L	332.9 ± 128.4	290.2 ± 115.5	-	0.854
CRP, mg/L	0.73 (0.50–9.86)	6.05 (4.40–15.90)	-	0.027
PCT, ng/mL	0.04 (0.03–0.10)	0.06 (0.03–0.07)	-	0.872
ALT, U/L	15.2 ± 5.8	16.1 ± 6.6	-	0.764
AST, U/L	27.2 ± 8.7	27.3 ± 11.9	-	0.782

Notes: Data are presented as mean ± SD, median (IQR) or number (%). Bold values indicate statistical significance (P < 0.05).

Quality Control and Global Metabolic Profiling

To evaluate the stability and reproducibility of the LC-MS/MS analytical system, QC samples were analyzed throughout the experiment. The PCA score plot showed that the QC samples clustered tightly, indicating good analytical stability and data reproducibility (Figure 1a). To further assess the global metabolic differences among groups, PLS-DA was performed on all study samples. The PLS-DA score plots showed a clear trend toward separation among the MPP, MPN, and HC groups, suggesting distinct metabolic profiles across the three groups (Figure 1b).

Figure 1.

Multivariate statistical analysis of plasma metabolic profiles. (a) PCA score plot of all samples, including QC samples (light green), demonstrating the high stability and reproducibility of the analytical system. (b) PLS-DA score plot illustrating clear separation among the MPP (blue, A), MPN (Orange, B), and HC (green, C) groups (n = 14 for each group) with 95% confidence intervals (shaded regions). (c) OPLS-DA score plot showing distinct metabolic differences between the MPP and MPN groups. (d) OPLS-DA score plot maximizing the separation between the MPP and HC groups.

Subsequently, OPLS-DA was conducted for pairwise comparisons. The OPLS-DA score plots showed clear separation between the MPP and MPN groups (Figure 1c), as well as between the MPP and HC groups (Figure 1d). The robustness of these models was evaluated using 7-fold cross-validation and permutation testing (n = 200). For the comparison between MPP and MPN, the model showed a good model fit (R2Y = 0.996) with moderate predictive ability (Q2 = 0.322). For the comparison between MPP and HC, similarly, the model showed excellent fit and predictive performance (R2Y = 0.999, Q2 = 0.954). In both models, the Q2 regression intercepts from permutation testing were below zero (−0.144 for MPP vs MPN and −0.349 for MPP vs HC), indicating no apparent overfitting (Figure S1).

Screening and Identification of Differential Metabolites

Differential metabolites were screened based on the predefined criteria of VIP > 1, P < 0.05, and FC ≥ 1.2 or ≤ 0.83. Volcano plots were used to visualize the overall distribution of significantly differentially expressed metabolites between groups. In the MPP vs MPN comparison, 127 differential metabolites were identified, including 78 upregulated and 49 downregulated metabolites (Figure 2a). In the MPP vs HC comparison, 749 differential metabolites were detected, including 288 upregulated and 461 downregulated metabolites (Figure 2b). Hierarchical clustering analysis was then performed to visualize the expression patterns of these differential metabolites. The heatmaps showed clear group-wise clustering patterns in both the MPP vs MPN and MPP vs HC comparisons, indicating distinct metabolic profiles between the compared groups (Figures S2 and S3). To further examine the overlap of differential metabolites among the comparisons, a Venn diagram was constructed. Fifteen metabolites were shared among the three comparisons, whereas the remaining metabolites were comparison-specific to varying degrees (Figure 2c).

Figure 2.

Screening and visualization of differential plasma metabolites. (a and b) Volcano plots of differential metabolites for the (a) MPP vs MPN and (b) MPP vs HC comparisons. Red dots represent upregulated metabolites, blue dots represent downregulated metabolites (Criteria: VIP > 1, p < 0.05). (c) Venn diagram illustrating the common and unique differential metabolites among different comparisons.

To further characterize the most significantly altered metabolites, the top differential metabolites were ranked by P value and visualized using lollipop plots. In the MPP vs MPN comparison, metabolites such as adipostatin J and 2-C-Methyl-1,4-erythrono-1,4-lactone were downregulated in the MPP group, whereas ezutromid and several other metabolites were upregulated (Figure 3a). Notably, ezutromid was annotated at confidence Level 3 with MS/MS fragmentation score below 45, indicating low annotation reliability and low biological plausibility as an endogenous plasma metabolite. In the MPP vs HC comparison, metabolites including aspartylcysteine were downregulated, whereas rhein and diflubenzuron were upregulated in the MPP group (Figure 3b). Diflubenzuron was assigned to Level 4 annotation tier with insufficient MS/MS matching evidence, representing a tentative annotation with limited biological credibility. Pearson correlation analysis was subsequently performed on the top 20 differential metabolites to evaluate their interrelationships. In the MPP vs MPN comparison, the correlation matrix showed predominantly positive correlations among many of the selected metabolites (Figure 3c). Similarly, in the MPP vs HC comparison, the top differential metabolites also exhibited dense positive correlation patterns (Figure 3d).

Figure 3.

Identification of top differential metabolites and metabolite correlation analysis. (a and b) Lollipop charts showing the top differential metabolites ranked by P value in the MPP vs MPN (a) and MPP vs HC (b) comparisons. Blue indicates downregulated metabolites, and red indicates upregulated metabolites. The horizontal axis represents log2(fold change), and dot size indicates the VIP value. (c and d) Pearson correlation matrices of the top 20 differential metabolites in the MPP vs MPN (c) and MPP vs HC (d) comparisons. Red indicates positive correlation, whereas blue indicates negative correlation. ** indicates P < 0.01; *** indicates P < 0.001.

Metabolic Pathway Analysis

To further characterize the biological relevance of the differential metabolites, pathway enrichment analysis was performed using the KEGG and Reactome databases.

In the MPP vs MPN comparison (A vs B), the downregulated metabolites were significantly enriched in the PPAR signaling pathway, linoleic acid metabolism, and vitamin digestion and absorption (Figure 4a). Reactome analysis further showed enrichment in pathways related to vitamin E transport, transcriptional regulation of white adipocyte differentiation, and transcriptional regulation of brown and beige adipocyte differentiation (Figure 4b). In addition, the chord plot provided an integrated overview of the top enriched KEGG pathways and their associated differential metabolites, showing that the enriched pathways in the MPP vs MPN comparison were predominantly associated with downregulated metabolites (Figure 4c).

Figure 4.

Pathway enrichment analysis of differential metabolites. (a) Bar plot of the top KEGG pathways enriched by downregulated metabolites in the MPP vs MPN comparison. (b) Bubble plot showing the top Reactome pathways enriched by downregulated metabolites in the MPP vs MPN comparison. Bubble size represents the number of metabolites, and color indicates the P value. (c) Circular enrichment plot (chord plot) showing the top enriched KEGG pathways and their associated differential metabolites in the MPP vs MPN comparison. (d) Bubble plot showing the top KEGG pathways enriched by downregulated metabolites in the MPP vs HC comparison. Bubble size represents the number of metabolites, and color indicates the P value.

In the MPP vs HC comparison (A vs C), a greater number of differential metabolites were observed. KEGG enrichment analysis showed that the downregulated metabolites were mainly enriched in pathways related to amino acid and neurotransmitter metabolism, including GABAergic synapse, arginine biosynthesis, and alanine, aspartate and glutamate metabolism (Figure 4d). In addition, glycerophospholipid metabolism was also significantly enriched among the downregulated metabolites.

Discussion

This study employed untargeted metabolomics technology to conduct a comprehensive analysis of plasma samples from children with MPP, children with MPN, and HC. We identified the distinct metabolic characteristics and dysregulated metabolic pathways associated with MPP. Unlike previous studies that only compared metabolic differences between MPP patients and healthy individuals, the present study included an MPN control group to exclude pneumonia-associated shared metabolic alterations, thus enabling the identification of MPP-specific metabolic disturbances. This strategy allowed for a more precise identification of the specific metabolic disturbances associated with Mycoplasma pneumoniae infection, markedly improving the specificity of the research findings and strengthening their clinical relevance.

Multivariate statistical analysis identified significant differences in metabolic profiles among the MPP, MPN, and HC groups. PCA, PLS-DA, and OPLS-DA models demonstrated clear separation of samples from the three groups, with closely clustered QC samples, thus affirming the reliability of the detection system. Notably, the number of differential metabolites between the MPP and HC groups (n=749) was significantly higher than that between the MPN and HC groups (n=537). This suggests that metabolic perturbations induced by Mycoplasma pneumoniae infection are more extensive and severe compared to those caused by other respiratory pathogens. Venn diagram analysis revealed 15 metabolites differentially expressed across all three comparison groups, indicating that these metabolites might indicate the body’s non-specific metabolic reaction to pulmonary infection, thus representing a general metabolic pattern associated with pneumonia. The identification of a subset of metabolites specific to MPP validates the distinct metabolic signature of MP infection. This signature provides potential targets for differentiating MPP from other respiratory infections as preliminary candidate biomarkers.

Amino acid metabolic dysregulation was identified as a key characteristic of MPP in this study. Children with MPP showed significantly lower plasma levels of amino acids related to arginine biosynthesis and the alanine, aspartate, and glutamate metabolic pathways compared to the HC group. Additionally, there was a noticeable trend of coordinated dysregulation among amino acid derivatives. Among the various metabolites that showed differential expression, a significant decrease in aspartylcysteine was particularly noteworthy. Aspartylcysteine, a dipeptide formed from aspartate and cysteine, is crucial for the synthesis of glutathione precursors.²⁰ Glutathione is the primary antioxidant in cells, essential for combating reactive oxygen species and preserving redox balance.²¹ Thus, reduced aspartylcysteine levels may be associated with weakened cellular antioxidant defenses, increasing lung tissue susceptibility to oxidative stress-induced damage, which is consistent with previous findings linking oxidative stress to MPP-associated lung injury.^11,22,23 Furthermore, the dysregulation of the arginine biosynthesis pathway may putatively reflect an imbalance in macrophage polarization induced by MP infection. This pathway divides into two separate pathways: one regulated by nitric oxide synthase and the other by arginase, steering macrophage polarization towards the pro-inflammatory M1 and anti-inflammatory M2 phenotypes, respectively.^24,25 The perturbation of this pathway corresponds with the exaggerated inflammatory reaction characteristic of MPP. Disruption of the alanine, aspartate, and glutathione metabolic pathway may contribute to MPP pathogenesis through various interconnected mechanisms. Firstly, glutamate is a precursor for glutathione synthesis and is involved in neuroimmune regulation.²⁶ Secondly, aspartate plays a critical role in nucleotide synthesis, and its imbalance can hinder lymphocyte proliferation and function.²⁷ Lastly, alanine is essential in the glucose-alanine cycle, and its disruption may indicate a shift in energy metabolism.²⁸ The coordinated dysregulation of amino acid derivatives suggests that MP infection could globally disrupt the amino acid metabolic network by influencing key metabolic enzymes or transcription factors. This metabolic reprogramming may affect immune cell function and oxidative stress responses,²⁹ potentially contributing to the pathophysiological processes of MPP and providing a metabolic basis for disease persistence and progression.

In the comparison between the MPP and non-MPP groups, the study observed notable alterations in lipid-related metabolites. Metabolites that were downregulated showed significant enrichment in pathways associated with lipids, including the PPAR signaling pathway, linoleic acid metabolism, and vitamin digestion and absorption. These lipid metabolic changes can be distinguished from universal metabolic shifts triggered simply by pneumonia, supporting the unique metabolic signature of MPP. The PPAR signaling pathway involves ligand-activated nuclear receptor transcription factors that regulate lipid metabolism, inflammatory responses, and immune cell function. The PPARγ isoform is particularly important in pulmonary inflammation and immune balance. PPARγ is highly expressed in lung tissue, especially in alveolar epithelial cells, macrophages, and dendritic cells.^30,31 By inhibiting pro-inflammatory transcription factors like NF-κB and AP-1, PPARγ activation decreases the release of inflammatory mediators such as TNF-α and IL-6. This leads to significant anti-inflammatory and tissue-protective effects in models of acute lung injury (ALI) and pneumonia.^32,33 Studies on animals have shown that drugs activating PPARγ, such as rosiglitazone, can effectively reduce lung inflammation caused by lipopolysaccharides in mice. Conversely, deficiency in PPARγ can worsen inflammation, highlighting the crucial role of PPARγ in maintaining lung balance and regulating excessive inflammation.^32,34 Recent multi-omics studies have also revealed lower levels of metabolites associated with the PPAR signaling pathway in children with MPP, supporting our findings. This metabolic imbalance may impair the body’s endogenous anti-inflammatory processes, resulting in prolonged and exacerbated inflammatory responses that may aggravate severe or drug-resistant pediatric MPP.³⁵ Moreover, these findings provide a novel metabolic perspective on the exaggerated inflammatory reactions observed in children with MPP. The significant decrease in the linoleic acid metabolism pathway warrants thorough investigation. Linoleic acid, an essential ω-6 polyunsaturated fatty acid, generates various bioactive metabolites, including pro-inflammatory lipid mediators such as prostaglandins and leukotrienes, as well as specialized pro-resolving lipid mediators (SPMs) including lipoxins and protectins.³⁶ These lipid derivatives play crucial roles in regulating the initiation, amplification, and resolution of inflammatory responses. Several metabolomic studies have reported profound disturbances in linoleic acid metabolism in pediatric patients with MPP.^37,38 One study identified 26 differentially expressed serum metabolites in children with MPP, highlighting linoleic acid, arachidonic acid, and tryptophan metabolism as crucial pathways related to inflammation.³⁸ Another study confirmed a notable suppression of linoleic acid metabolism in MPP patients, accompanied by the upregulation of arachidonic acid, a key component in this metabolic network, indicating a substantial reconfiguration of lipid metabolism.³⁷ In this study, the decrease in linoleic acid metabolism may hinder the production of pro-resolving mediators like lipoxins and protectins. This hindrance might delay the resolution of inflammation effectively, leading to a transition from acute, self-limited pulmonary inflammation to chronic and persistent inflammation. This mechanism aligns well with the typical clinical characteristics of MPP, such as a prolonged clinical course and a higher likelihood of recurrence. Furthermore, vitamin E functions as a crucial antioxidant by preventing lipid peroxidation, preserving cell membrane integrity, and regulating oxidative stress.³⁷ The present study revealed a significant disruption in the transport of vitamin E, suggesting that Mycoplasma pneumoniae infection could hinder the systemic delivery of antioxidant vitamins like vitamin E. This disruption may directly weaken the endogenous antioxidant defense system, exacerbating lipid peroxidation and cell membrane injury in lung tissue. Oxidative stress and lipid peroxidation are commonly associated with various pulmonary conditions such as acute respiratory distress syndrome and chronic obstructive pulmonary disease.^39,40 Combined disturbances in vitamin E transport, PPAR signaling and linoleic acid metabolism may correlate with a pro-inflammatory, pro-oxidant state linked to MPP progression.

Children with MPP showed a significant enrichment of the GABAergic synaptic pathway compared to the HC group, offering a novel perspective on the pathogenesis of this disease. While traditionally known as a primary inhibitory neurotransmitter in the central nervous system, recent research indicates that GABA signaling plays a crucial role in immunomodulation.⁴¹ Various immune cells express GABA receptors, allowing GABA signaling to inhibit pro-inflammatory cytokine release and promote anti-inflammatory responses.⁴¹ Pulmonary neuroendocrine cells produce and release GABA, influencing neighboring immune cells in a paracrine manner and establishing a neuro-immune regulatory system.⁴² The changes in the GABAergic synaptic pathway observed in this study may disrupt this regulatory network, providing a metabolic framework to explain the respiratory rhythm irregularities and wheezing often seen in children with MPP. These findings suggest that targeting neuro-immune interactions could be a promising approach for treating MPP.

The study integrated clinical baseline data and metabolomic profiles, revealing elevated CRP levels and neutrophil percentages, alongside decreased lymphocyte percentages in children with MPP. These findings suggest a robust inflammatory response, aligning with the enrichment of inflammation-related pathways like the PPAR signaling pathway and linoleic acid metabolism. This consistency implies a reinforcing cycle between metabolic dysregulation and inflammation during MPP progression. Pro-inflammatory cytokines can impact metabolic enzyme gene expression via the JAK-STAT and NF-κB signaling pathways, leading to abnormal amino acid and lipid metabolism.^43,44 Conversely, abnormal metabolites from metabolic disturbances, like lipid peroxidation products, can trigger innate immune cells through damage-associated molecular patterns, fueling the release of inflammatory mediators and creating a feedback loop.⁴⁵ Identifying this interplay provides new insights for targeted therapies. Modulating key metabolic pathways could indirectly manage inflammatory responses, disrupting the cycle and enhancing clinical outcomes in MPP.

The study identifies differential metabolites and dysregulated pathways with significant clinical translational potential for pediatric MPP. Conventional diagnostic methods for MPP have limitations,^5,9,11 whereas metabolomic profiling offers advantages such as rapid analysis, non-invasiveness, and potential for dynamic monitoring.^15–17 The metabolite signature from this study could be a valuable adjunctive diagnostic tool for early MPP detection. Additionally, the dysregulated pathways generate preliminary hypothesis-generating clues for potential therapeutic targets requiring comprehensive functional verification: downregulation of the PPAR signaling pathway suggests exploring PPARγ agonists; correcting abnormal amino acid metabolism with specific precursors may enhance antioxidant capacity; and addressing impaired vitamin E transport supports considering vitamin E supplementation as a complementary strategy. These findings highlight the significance of metabolomics in advancing innovative diagnostics and precision therapies for pediatric MPP.

This study has several limitations. Firstly, this is a preliminary exploratory metabolomic research, and the small sample size (n=14 per group) may reduce statistical power and limit the generalizability of our metabolic signatures. Future studies should validate these results with a larger sample size and multi-center prospective designs. Secondly, the cross-sectional design hinders establishing temporal relationships between metabolic perturbations and disease severity, thus limiting causal inferences. Thirdly, potential confounding effects of co-infection with other pathogens on the metabolic profiles of the study participants could not be entirely excluded. Fourthly, the absence of severity stratification among children with MPP prevented assessing associations between metabolic signatures and disease severity. Fifthly, untargeted plasma metabolomics has notable limitations for routine clinical diagnosis. The candidate metabolic signatures identified herein lack external cohort validation and standardized clinical detection schemes, so they cannot yet replace conventional clinical laboratory tests for routine MPP screening. Lastly, the biological functions and regulatory mechanisms of most differential metabolites remain unclear; their causal roles need substantiation through in vitro cellular experiments and in vivo animal models.

Conclusions

In conclusion, this study systematically characterized the plasma metabolic profiles of children with MPP using untargeted metabolomics, revealing that MP infection induces extensive metabolic disturbances involving amino acid, lipid, vitamin, and neurotransmitter metabolism. Amino acid, lipid and neurotransmitter metabolic changes may be associated with altered antioxidant and inflammatory homeostasis relevant to MPP. These metabolic alterations interact with clinical inflammatory markers to form a vicious cycle that drives MPP progression. The identified MPP-specific differential metabolites and dysregulated metabolic pathways not only deepen our understanding of MPP pathogenic mechanisms but also generate preliminary hypothesis-generating clues for potential candidate biomarkers and follow-up research on therapeutic targets for pediatric MPP. Large-scale prospective multicenter cohorts with targeted validation are needed in future research to further verify candidate biomarkers, clarify therapeutic potential and dissect causal mechanisms. This study establishes a metabolic groundwork to enhance precision medicine in managing childhood MPP clinically.

Funding Statement

This work was supported by the Suzhou Clinical Key Disease Diagnosis and Treatment Technology Special Fund (LCZX202352), the Suzhou Applied Basic Research (Medical and Health) Science and Technology Innovation Project (SYWD2024234), and the Suzhou Science and Technology Planning Project (SYWD2025251).

Abbreviations

MPP, Mycoplasma pneumoniae pneumonia; MP, Mycoplasma pneumoniae; CAP, Community-Acquired Pneumonia; HC, Healthy controls; MPN, Non-Mycoplasma pneumoniae pneumonia (non-MPP); tNGS, Targeted next-generation sequencing; QC, Quality control; LC-MS, Liquid chromatography-mass spectrometry; PCA, Principal Component Analysis; PLS-DA, Partial Least Squares-Discriminant Analysis; OPLS-DA, Orthogonal Partial Least Squares-Discriminant Analysis; VIP, Variable Importance in Projection; FC, Fold change; SD, Standard deviation; IQR, Interquartile range; CRP, C-reactive protein; WBC, White blood cell; PLT, Platelets; PCT, Procalcitonin; ALT, Alanine aminotransferase; AST, Aspartate aminotransferase.

Data Sharing Statement

The datasets used and/or analyzed during the current study are available from the corresponding author, Shuangqin Ran, upon reasonable request.

Ethics Approval and Informed Consent

This study was performed in accordance with the Declaration of Helsinki. The study protocol was approved by the Institutional Ethics Committee (Approval No. IRB2023060). Written informed consent was obtained from the parents or legal guardians of all participating children prior to enrollment.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work. All authors have read the manuscript and provided their consent for the submission.

Disclosure

The authors declare that they have no competing interests.