Candida-associated immune and metabolic rewiring in chronic obstructive pulmonary disease Authors: Ran Wang, Zhibin Chen, Yanwei Wang, Niangui Zhao, Kecong Li, Haiyan Gong, Yanna Zhang, Songyin Lai, Zhangqiang Ye, Hui Yang Published: 2026 ABSTRACT AbstractBackgroundChronic obstructive pulmonary disease (COPD) is characterized by persistent systemic inflammation. Candida colonization is frequently observed in patients with COPD, yet its systemic impact remains poorly understood. This study aimed to identify molecular alterations uniquely induced by Candida under COPD conditions.MethodsAn interaction-based model incorporating COPD status, Candida colonization, and their interaction term was used to define COPD-context-dependent Candida-responsive genes and metabolites. Peripheral blood mononuclear cell transcriptomes were analyzed by RNA sequencing, followed by functional enrichment, protein-protein interaction (PPI), and transcription factor (TF) motif analyses. Plasma metabolomics was performed to identify exclusive differential metabolites and enriched pathways. Multi-omic integration was performed to assess coordinated relationships between genes and metabolites, with correlation network construction and Spearman’s correlation validation.ResultsA subset of differentially expressed genes (DEGs) was exclusively altered in the CA_COPD versus COPD comparison. These genes were primarily enriched in immune regulation and inflammatory signaling pathways. PPI analysis identified limited but biologically meaningful interaction networks, while TF motif analysis implicated the zf-C2H2 family as a potential regulator of context-specific transcriptional remodeling. Metabolomic profiling revealed exclusive differential metabolites in CA_COPD. Upregulated metabolites were marginally associated with amino sugar and nucleotide sugar metabolism, whereas downregulated metabolites were significantly enriched in aminoacyl-tRNA biosynthesis, central carbon metabolism, protein digestion and absorption, and ABC transporter pathways. Integrated multi-omic analysis demonstrated tightly coordinated gene–metabolite networks linking transcriptional and metabolic reprogramming under Candida-COPD interaction conditions.ConclusionCandida colonization induces distinct COPD-context-dependent transcriptional and metabolic reprogramming, suggesting a specific host response signature that may contribute to disease progression.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12890-026-04415-8.Keywords: Chronic obstructive pulmonary disease, Candida, Peripheral blood mononuclear cells, RNA-Seq, Metabolomic analysis Introduction Chronic obstructive pulmonary disease (COPD) is a chronic inflammatory disorder characterized by structural abnormalities of the airways and alveoli that result in persistent airflow limitation and respiratory symptoms [1]. Its diagnosis is commonly based on spirometric criteria, including the Global Initiative for Chronic Obstructive Lung Disease (GOLD) fixed ratio and the lower limit of normal (LLN). Globally, COPD affects a substantial proportion of the adult population aged 30–79 years, with an estimated prevalence of 10.3% under the GOLD definition and 7.6% under the LLN definition in 2019, corresponding to hundreds of millions of cases worldwide [2]. Despite a general decline in overall mortality, COPD remains one of the leading causes of death globally, with notable sex- and region-specific trends, including decreasing mortality in men but rising rates in women in several regions such as the European Union and the United States [3]. The pathogenesis of COPD is multifactorial and involves complex interactions between environmental and host factors, and major risk contributors include smoking, low body mass index, biomass fuel exposure, occupational dust or smoke exposure, and male sex [2]. Current management strategies mainly focus on symptom relief and exacerbation prevention through inhaled bronchodilators, with inhaled corticosteroids used in selected patients, alongside non-pharmacological approaches such as smoking cessation, pulmonary rehabilitation, vaccination, and long-term oxygen therapy for severe hypoxemia [4, 5]. Nevertheless, existing therapies are largely palliative and cannot reverse airway remodeling or fully arrest disease progression, underscoring the need for novel therapeutic strategies based on a deeper understanding of the molecular mechanisms underlying COPD pathogenesis. Alterations in the airway microbiome, including shifts in both bacterial and fungal communities, are increasingly recognized as contributors to COPD pathogenesis [6]. Among airway fungi, Candida species, which are yeast-like organisms that normally colonize human epithelium, have drawn particular attention. Although typically commensal, these opportunistic fungi can overgrow under conditions such as immune dysfunction, antibiotic exposure, or microbial imbalance, leading to infections collectively termed candidiasis [7, 8]. In patients with chronic respiratory diseases, the relative abundance of Candida in the airway mycobiome is frequently elevated, and culture-based studies report isolation rates of approximately 15–25% in respiratory specimens [9]. Mechanistically, Candida species can exacerbate airway inflammation by releasing virulence factors such as candidalysin, disrupting epithelial barrier integrity, activating innate immune signaling, and engaging in complex inter-kingdom microbial interactions [10]. Together, these processes may contribute to persistent inflammation and disease progression in COPD; however, the precise pathogenic role of Candida in COPD remains to be fully elucidated. Peripheral blood mononuclear cells (PBMCs) are bone marrow-derived immune cells that mainly consist of lymphocytes and monocytes. Accumulating evidence suggests that genomic and transcriptomic alterations in PBMCs are involved in the pathogenesis of COPD [11–13]. Importantly, PBMC findings may capture bidirectional crosstalk between the systemic and pulmonary compartments: inflammatory mediators released in the lung can enter the circulation and modulate PBMC function, while activated PBMCs can, in turn, migrate into the lung and promote local inflammation [14–16]. However, the effects of Candida infection on the PBMC transcriptome in COPD patients remain unclear, and its potential influence on systemic metabolism has not been well defined. To address these gaps, we collected PBMC samples from COPD and non-COPD patients with lower respiratory tract disorders and performed RNA sequencing (RNA-seq). In parallel, plasma samples were obtained for unbiased metabolomic profiling. Through integrative analysis of transcriptomic and metabolomic data, this study aims to elucidate the overall impact of Candida infection on COPD pathophysiology. Materials and methods Human samples Between September 2025 and December 2025, patients admitted to our hospital were consecutively enrolled in this study. Fasting peripheral blood samples were obtained from each participant and processed for the isolation of PBMCs and plasma. Written informed consent was secured from all participants before inclusion. The study protocol was reviewed and approved by the Ethics Committee of The Second Affiliated Hospital of Xiamen Medical College (Approval No.2026036) and conducted in compliance with the Declaration of Helsinki. Candida colonisation was defined as the detection of Candida species from the first lower respiratory tract specimen (sputum or bronchoalveolar lavage fluid, BALF) collected after hospital admission, using either conventional microbiological culture or metagenomic next-generation sequencing (mNGS), in the absence of any evidence of invasive candidiasis. All samples for analysis were obtained upon hospital admission, before the administration of any therapeutic agents, including antibiotics, antifungals, or systemic corticosteroids. RNA-Seq Total RNA was extracted using TRIzol reagent (Invitrogen), and the purity and concentration were assessed with a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific). RNA integrity and quantity were further evaluated using the BiOptic Qsep1 system and confirmed by RNase-free agarose gel electrophoresis. For library construction, mRNA was enriched using poly-T oligo-attached magnetic beads and fragmented to approximately 300–350 bp. First-strand cDNA was synthesized with random hexamers, followed by second-strand synthesis, end repair, 3′ adenylation, and adaptor ligation. Libraries were amplified by PCR, purified, and quantified using a Qubit fluorometer and qPCR, with fragment size distribution verified by agarose gel electrophoresis. The qualified libraries were pooled and sequenced on the Illumina NovaSeq X Plus platform. Raw sequencing data were processed using Fastp to generate high-quality clean reads by removing adaptor sequences, reads containing > 10% ambiguous bases, and low-quality reads with > 50% bases having a Q-value ≤ 20. The filtered paired-end reads were aligned to the reference genome using HISAT2 with default parameters. Differential expression analysis was performed using the DESeq2 R package, and genes with a false discovery rate (Q-value) < 0.001 (corrected by Benjamini-Hochberg procedure) and an absolute fold change ≥ 2 were defined as differentially expressed genes (DEGs). For downstream functional analysis, DEGs were mapped to the STRING database to construct protein-protein interaction (PPI) networks, which were visualized in Cytoscape (v3.7.1). Open reading frames were predicted using GetORF, and potential transcription factor-encoding genes were identified by aligning ORF sequences to the Animal Transcription Factor Database (AnimalTFDB) using DIAMOND based on conserved transcription factor family features. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed using the R package ClusterProfiler (v4.10.1). Quantitative PCR (qPCR) Total RNA prepared using TRIzol-based method was reverse-transcribed into complementary DNA using the HiScript Reverse Transcription Kit (Vazyme) according to the manufacturer’s instructions. qPCR was subsequently performed on a Roche LightCycler 96 system using ChamQ Universal SYBR qPCR Master Mix (Vazyme). Both biological and technical replicates were included for each experimental group to ensure reproducibility. Relative gene expression levels were determined using the 2−ΔΔCt method with appropriate internal normalization. The primer sequences used for amplification are listed below (5’-3’): TMEM50B (Forward primer-GCGTTTGAGGAAGCTGGGATA, Reverse primer-CGACAACAGATGCCACAGCA), GZMK (Forward primer-CCCTGCGAGAAGTCACTGTT, Reverse primer-CCCCCTGAGTCACCCTTACA), HLA-DRB5 (Forward primer-TGAGCCTAAGGTGACTGTGTAT, Reverse primer-GACTTCAATGCTGCCTGGAT), CLEC2A (Forward primer-GTTCCCAAGTTGATACAAAACTGGA, Reverse primer-TGCTTGGACCATGTGGCTAT), Beta actin (Forward primer-GAGAAAATCTGGCACCACACC, Reverse primer-GGATAGCACAGCCTGGATAGCAA). Metabolomic analysis Metabolites were extracted from thawed plasma samples using a methanol-based protein precipitation method. Briefly, 100 µL of each sample was mixed with 400 µL of pre-cooled methanol/acetonitrile (1:1, v/v), followed by sonication for 10 min and incubation at -20 °C for 1 h. After centrifugation at 20,000 g for 15 min at 4 °C, the supernatant was collected and evaporated to dryness overnight at 4 °C. The dried extracts were reconstituted in 100 µL of acetonitrile/water (1:1, v/v), sonicated, and centrifuged again under the same conditions, and the final supernatant was transferred to LC-MS injection vials. A pooled quality control (QC) sample was generated by combining equal volumes of each supernatant to monitor analytical stability. LC-MS/MS analysis was conducted using an UltiMate 3000 UPLC system coupled to a Q-Exactive Plus mass spectrometer (Thermo Fisher Scientific). Metabolites were separated on an ACQUITY UPLC T3 column (100 × 2.1 mm, 1.7 μm) at 50 °C with a mobile phase consisting of 0.1% formic acid in water (A) and acetonitrile (B) at a flow rate of 0.3 mL/min under a gradient elution program. The mass spectrometer was operated in both positive and negative ion modes using data-dependent acquisition with stepped collision energies, and full-scan spectra (m/z 70–1050) were acquired at high resolution. Raw data were processed using Progenesis QI (v2.0) for peak alignment and metabolite identification against the HMDB and METLIN databases with strict mass tolerance criteria, retaining compounds supported by at least two ions. Data normalization was performed using probabilistic quotient normalization, and missing values were imputed with half of the minimum value. Differential metabolites were identified based on OPLS-DA (ropls package in R), with model validation by permutation testing (200 iterations), and significance defined as Variable Importance in Projection (VIP) > 1, P < 0.05, and fold change > 1.2 or < 0.83. Multi-omic integration and correlation analysis Supervised multi-omic integration was performed using the DIABLO framework implemented in the mixOmics package (version 6.26.0). Separate analyses were conducted for downregulated and upregulated transcriptomic and metabolomic features. Correlation networks were visualized using circos plots with a correlation threshold of r = 0.6, and variable contribution patterns were assessed using latent component analysis. Pairwise Spearman correlation analysis was performed using the Hmisc package (version 5.2.0) to validate significant gene–metabolite associations. Statistical analysis Statistical analyses were performed using GraphPad Prism (version 10.6). Continuous variables (e.g., age and PCT levels) were assessed for normality using the Shapiro-Wilk test and for homogeneity of variance before analysis. For normally distributed data with equal variances, comparisons among multiple groups were conducted using one-way analysis of variance (ANOVA) followed by the Least Significant Difference (LSD) post hoc test, and results are presented as mean ± standard deviation. For data that were not normally distributed, the Kruskal-Wallis H test was applied, followed by Dunn’s post hoc test for pairwise comparisons, with results expressed as median (interquartile range). Categorical variables (e.g., sex and smoking history) are presented as counts and percentages and were compared using Fisher’s exact test. A two-sided P-value < 0.05 was considered statistically significant. Detailed methods of statistical analysis for each variable are listed in Table S1 or the corresponding figure legends. Results Characteristics of the cohort A total of 19 participants were enrolled, including 6 controls without COPD or Candida infection (NC), 6 patients with COPD, 3 patients with Candida infection alone (CA), and 4 patients with both Candida infection and COPD (CA_COPD). There were no significant differences among groups in age, sex distribution, BMI, arterial blood gas parameters (pH, PaCO₂, PaO₂, HCO₃⁻), oxygenation index, APACHE II score, GCS score, or most laboratory indicators (all P > 0.05). Smoking prevalence differed significantly among groups (P = 0.003), with all COPD patients and 75% of CA_COPD patients being smokers, compared with only 16.7% in the NC group and none in the CA group. As expected, CAT scores were significantly higher in the COPD group compared with NC (P = 0.049), reflecting greater symptom burden. Inflammatory markers showed no statistically significant differences overall, although CRP tended to be higher in the CA group compared with NC (P = 0.055). Notably, serum albumin levels differed significantly between groups (P = 0.037), with lower levels observed in NC compared with CA_COPD. Other markers, including PCT, ESR, IL-6, TNF-α, blood cell counts, BNP, and D-dimer, did not demonstrate significant intergroup differences. We further checked the immunological status of all enrolled subjects. No patient had HIV infection, a history of solid organ or hematopoietic stem cell transplantation, or active chemotherapy or radiotherapy. None of the patients had long-term use of high-dose immunosuppressants. Only part of the COPD patients had a history of short-term systemic corticosteroid use (prednisone-equivalent dose < 10 mg/day for no more than one week), with no recent use before sample collection. Overall, the cohort was comparable across most baseline characteristics, with differences primarily observed in smoking status, symptom burden, and albumin levels. The detailed clinical features of the cohort are shown in Table 1 and S2. Table 1Clinical features of the cohortNumber of participantsNCCOPDCACA_COPDP-value6634NSGender (M/F)3/36/02/14/0NSAge74 ± 6.970.8 ± 5.577.3 ± 11.167 ± 11.6NSBMI (kg/m2)22.7 ± 3.221 ± 3.619.4 ± 1.322.3 ± 4.1NSSmoker, N (%)1 (16.7)6 (100)0 (0)3 (75)P = 0.003CAT score9.5(8.75, 19.75)33.5(30.75, 36.25)9(9,28)31(24.75, 34.25)P = 0.049(NC vs. COPD)pH7.41(7.26, 7.44)7.41(7.34, 7.44)7.45(7.37, 7.48)7.48(7.28, 7.5)NSPaCO246.8 ± 12.244.4 ± 12.243 ± 4.142.6 ± 15.6NSPaO293.3 ± 24.2100.9 ± 42.673.6 ± 7.985.8 ± 7.8NSHCO3−26.5 ± 4.625.4 ± 3.827.9 ± 3.525.4 ± 2.5NSOxygenation index287.6 ± 99.4336.5 ± 142.0245.3 ± 26.4264.3 ± 24.3NSAPACHEII12.3 ± 5.99.5 ± 2.214 ± 7.314 ± 6.0NSGCS score15(12.25, 15)15(14.75, 15)15(10, 15)14.5(14, 15)NSCRP55.9 ± 51.730.2 ± 28.7133.4 ± 40.846.1 ± 17.0P = 0.055(NC vs. CA)PCT0.23(0.04, 0.48)0.05(0.05, 0.28)1.66(0.54, 39.6)0.21(0.09, 5.07)NSESR38.8 ± 50.821.2 ± 4.272 ± 56.037.8 ± 26.1NSIL-622(10, 221.3)24(12.9, 37.75)230(32.5, 3196)22.95(15.48, 23.66)NSTNF-α2.33 ± 0.62.24 ± 0.84.16 ± 1.05.73 ± 6.3NSAbsoluteneutrophil count7.51 ± 5.88.05 ± 3.77.15 ± 3.111.6 ± 6.2NSAbsolute lymphocyte count1.08 ± 0.71.08 ± 0.50.54 ± 0.31.17 ± 0.4NSPlatelet count223.5 ± 69.7246.7 ± 50.7843.3 ± 999.7274 ± 76.9NSAlbumin35.1 ± 5.940.7 ± 3.029.5 ± 2.143.3 ± 3.1P = 0.037(NC vs. CA_COPD)BNP121.8(41.58,671)92.35(11.85,390.3)589(48.2,1846)346.5(114.3,759.5)NSD-dimer38.8 ± 50.821.2 ± 4.272 ± 56.037.8 ± 26.1NS Clinical features of the cohort 9.5 (8.75, 19.75) 33.5 (30.75, 36.25) 9 (9,28) 31 (24.75, 34.25) P = 0.049 (NC vs. COPD) 7.41 (7.26, 7.44) 7.41 (7.34, 7.44) 7.45 (7.37, 7.48) 7.48 (7.28, 7.5) 15 (12.25, 15) 15 (14.75, 15) 15 (10, 15) 14.5 (14, 15) P = 0.055 (NC vs. CA) 0.23 (0.04, 0.48) 0.05 (0.05, 0.28) 1.66 (0.54, 39.6) 0.21 (0.09, 5.07) 22 (10, 221.3) 24 (12.9, 37.75) 230 (32.5, 3196) 22.95 (15.48, 23.66) Absolute neutrophil count P = 0.037 (NC vs. CA_ COPD) 121.8 (41.58,671) 92.35 (11.85,390.3) 589 (48.2,1846) 346.5 (114.3,759.5) Identification of DEGs between the CA_COPD and COPD groups To investigate the impact of Candida infection on PBMC transcriptomic alterations in the context of COPD, we performed RNA-seq on PBMCs isolated from patients with COPD with or without Candida infection. A total of 166 DEGs were identified, including 45 upregulated and 121 downregulated genes in the CA_COPD group compared with the COPD group (Fig. 1A, B). Transcriptomic differences between CA_COPD and COPD groups. A Volcano plot showing DEGs between CA_COPD and COPD groups. Red dots represent upregulated genes and blue dots represent downregulated genes in CA_COPD compared with COPD. B Bar plot indicating the number of upregulated and downregulated DEGs. C GO enrichment analysis of DEGs, categorized based on cellular component, molecular function, and biological process. D Bubble plot showing KEGG pathway enrichment analysis results of DEGs (top 20 pathways) GO enrichment analysis indicated that these DEGs were mainly associated with cellular components such as the collagen trimer and complement component C1 complex, molecular functions including dioxygenase activity and insulin receptor binding, and biological processes related to leukocyte activation, response to vitamins, and protein processing (Fig. 1C). Furthermore, KEGG pathway analysis revealed enrichment in the FoxO signaling pathway and inflammatory disease-related pathways, such as systemic lupus erythematosus and pertussis (Fig. 1D). Collectively, these results suggest that Candida infection may aggravate COPD pathophysiology by modulating immune-related pathways and amplifying systemic inflammatory responses. COPD-candida interaction reveals immune and metabolic transcriptomic reprogramming in PBMCs To elucidate the specific regulatory mechanisms by which Candida infection influences the PBMC transcriptome in the context of COPD, we incorporated COPD status, Candida colonization, and their interaction term into the analytical model. Genes showing significant COPD × Candida interaction effects were defined as COPD-context-dependent Candida-responsive genes, representing molecular alterations uniquely induced by Candida under COPD conditions. Notably, these genes were differentially expressed in the CA_COPD versus COPD comparison, but not in the CA versus NC or COPD versus NC comparisons. Venn diagram analysis identified 42 upregulated and 117 downregulated genes, accounting for the majority of DEGs observed between the CA_COPD and COPD groups (Fig. 2A, B, Table S2). GO enrichment analysis revealed that the upregulated genes were primarily associated with cellular components such as blood microparticles, IgM immunoglobulin complexes, and complement component C1q. These genes were enriched in molecular functions including dioxygenase, kynureninase, and hydrolase activities, and were involved in biological processes such as B cell-mediated immunity, NAD biosynthesis, and amino acid metabolism (Fig. 2C). In contrast, downregulated genes were mainly enriched in cellular components such as the collagen trimer and mitotic spindle microtubules, with molecular functions related to dioxygenase activity and insulin receptor binding, and were involved in protein processing and cell cycle transition (Fig. 2D). KEGG pathway analysis demonstrated that upregulated genes were significantly enriched in tryptophan metabolism, cytokine-cytokine receptor interaction, and pertussis pathways, whereas downregulated genes were predominantly associated with the FoxO signaling pathway (Fig. 2E, F). Identification of COPD-context–dependent Candida-responsive genes. A, B Venn diagram showing overlap of upregulated (A) and downregulated (B) DEGs among CA_COPD vs. COPD, CA vs. NC, CA_COPD vs. CA, and COPD vs. NC comparisons. C, D GO enrichment analysis of exclusively upregulated and downregulated genes with significant COPD × Candida interaction effects (outlined by red and green squares in panels A and B). E, F Bubble plots indicating KEGG enrichment analyses of exclusively upregulated and downregulated genes with significant COPD × Candida interaction effects (top 20 pathways). G qPCR data showing the expression of TMEM50B, GZMK, HLA-DRB5, and CLEC2A in different groups of PBMCs. One-way ANOVA, along with Tukey’s HSD, was used for statistical analysis To validate the RNA-seq results, we performed qPCR to assess the expression of two upregulated genes (TMEM50B and GZMK) and two downregulated genes (HLA-DRB5 and CLEC2A) that are associated with COPD or whose homologs have been implicated in COPD [17–20]. Consistent with the RNA-seq findings, qPCR confirmed that all four genes exhibited the same direction of change in response to Candida infection (Fig. 2G). Together, these findings suggest that Candida infection under COPD conditions may reshape immune-metabolic networks, enhance B cell-mediated and inflammatory responses, and simultaneously alter immune cell cycle and signaling pathways, thereby contributing to disease progression. PPI and transcription factor binding analysis To investigate the potential interactions among the identified DEGs, we performed PPI analysis. Among the upregulated DEGs, six genes, IDO1, CXCR6, CD70, CD40, KYNU, and GZMK, were predicted to interact, forming a small interaction network in which IDO1 functioned as a hub gene (Fig. 3A). In contrast, the downregulated DEGs formed several small clusters consisting of two to four genes with mainly linear interactions and limited network complexity. The only notable interaction network among downregulated genes comprised six genes, with MDM2 identified as the central hub (Fig. 3B). PPI network and TF binding analysis. A PPI network of upregulated COPD-context–dependent Candida-responsive genes, with IDO1 identified as a hub gene. B PPI network of downregulated genes. (C) Predicted TF binding motif analysis of upregulated genes. D Predicted TF binding motif analysis of downregulated genes To further elucidate the regulatory mechanisms underlying these transcriptional alterations, we conducted transcription factor (TF) binding motif analysis. Among the upregulated genes, only one TF family, zf-C2H2, was predicted to bind two target genes. For the downregulated genes, three TF families were identified: CSRNP_N and RXR-like, each predicted to bind one gene, and zf-C2H2, which was predicted to bind two genes (Fig. 3C, D). These findings suggest that specific PPI networks and TF-mediated regulatory mechanisms may partially underlie the effects of Candida infection on PBMC transcriptomic remodeling in the COPD context. Identification of differential metabolites between the CA_COPD and COPD groups To evaluate the impact of Candida infection on systemic metabolism, we performed an untargeted metabolomic analysis of plasma samples from the different groups. Orthogonal partial least squares discriminant analysis (OPLS-DA) demonstrated a clear separation between the CA_COPD and COPD groups, indicating distinct metabolic profiles (Fig. 4A). Permutation test (200 times) yielded a PR2Y of 0.015 and a PQ2 of 0.03. The positive PQ2 value suggests that the model may have limited predictive robustness, likely attributable to the small sample size. A total of 57 significantly altered metabolites were identified, including 35 upregulated and 22 downregulated metabolites in the CA_COPD group compared with the COPD group (Fig. 4B). Heatmap visualization revealed that the most significantly upregulated metabolites included pregnanolone, kutilensoside H, and echinasteroside C, whereas downregulated metabolites comprised 3-(5,7-dimethoxy-4-oxo-4 H-chromen-2-yl) propanoic acid, stearoylcarnitine, and phosphatidylethanolamine (PE 18:1), among others (Fig. 4C). KEGG pathway enrichment analysis indicated that these differential metabolites were mainly involved in pathways related to ABC transporters, central carbon metabolism, and amino sugar and nucleotide sugar metabolism (Fig. 4D). Together, these findings suggest that Candida infection in the COPD context is associated with systemic metabolic reprogramming, potentially affecting energy metabolism, lipid metabolism, and transport-related pathways. Differential plasma metabolites between CA_COPD and COPD groups. A OPLS-DA score plot showing separation between CA_COPD and COPD groups. B Volcano plot of differential metabolites, with upregulated (red color) and downregulated metabolites (blue color) indicated. C Heatmap of significantly altered metabolites between CA_COPD and COPD groups. D KEGG pathway enrichment analysis of differential metabolites COPD-Candida interaction drives context-specific metabolic reprogramming in plasma. To delineate the specific metabolic pathways through which Candida infection reshapes systemic metabolism in the COPD setting, we incorporated COPD status, Candida colonization, and their interaction term into the statistical model. Metabolites exhibiting significant COPD × Candida interaction effects were defined as COPD-context-dependent Candida-responsive metabolites, representing metabolic alterations uniquely induced by Candida under COPD conditions. Venn diagram analysis identified 21 exclusively upregulated and 19 exclusively downregulated metabolites in the CA_COPD versus COPD comparison (Fig. 5A, B). KEGG pathway enrichment analysis indicated that the upregulated metabolites were marginally associated with amino sugar and nucleotide sugar metabolism. In contrast, the downregulated metabolites were significantly enriched in pathways related to aminoacyl-tRNA biosynthesis, central carbon metabolism, protein digestion and absorption, and ABC transporters (Fig. 5C, D). These findings suggest that Candida infection under COPD conditions may selectively perturb core metabolic pathways, particularly those involved in protein synthesis, energy metabolism, and substrate transport, thereby contributing to systemic metabolic dysregulation. COPD-context–dependent Candida-responsive metabolites.A, B Venn diagrams identifying metabolites with significant COPD × Candida interaction effects. C, D KEGG enrichment analysis of exclusively upregulated and downregulated metabolites outlined by red and green squares in panels A and B Integrated multi-omic analysis of transcriptomic and metabolomic profiles specifically between the CA_COPD and COPD groups To investigate coordinated relationships between the exclusive DEGs and differential metabolites in the CA_COPD versus COPD comparison, we performed supervised multi-omic integration using the DIABLO framework. The analysis revealed coordinated relationships between transcriptomic and metabolomic alterations in both downregulated and upregulated datasets. Correlation network analysis demonstrated extensive positive associations between genes and metabolites at a moderate correlation threshold (r = 0.6) in both the upregulated and downregulated datasets (Fig. 6A, B). Notably, multiple genes frequently converged on fewer metabolites, suggesting shared metabolic regulation and pathway convergence. Integrated multi-omic network analysis of DEGs and differential metabolites. A Correlation network (circos plot) showing relationships between the 42 upregulated genes and 21 metabolites at a correlation threshold of r = 0.6. B Circos plot showing correlation between the 117 downregulated genes and 19 metabolites at r = 0.6. C, D Variable loading plots illustrating the contribution of upregulated (C) and downregulated (D) genes and metabolites to the first two latent components, highlighting coordinated clustering patterns Correlation structure analysis further showed that specific genes and metabolites contributed similarly to the latent components in both upregulated and downregulated (Fig. 6C, D), indicating coordinated biological functions and pathway linkages. Spearman correlation analysis confirmed significant positive and negative associations between many transcriptomic and metabolomic features in both upregulated and downregulated datasets (Figure S1A, B), supporting the robustness of the identified interactions. Collectively, these findings demonstrate that both upregulated and downregulated molecular features form highly interconnected networks, indicating that transcriptional suppression or activation is consistently mirrored at the metabolic level. The identified gene-metabolite associations provide potential candidates for future functional validation studies in the context of Candida infection under COPD conditions. Discussion In this study, we investigated the impact of Candida infection on PBMC transcriptomics and systemic metabolism in patients with COPD. Using an integrative multi-omics approach, we found that Candida colonization in COPD is associated with significant immune and metabolic remodeling, providing insight into how opportunistic fungal infections may exacerbate disease severity. Our transcriptomic analysis identified DEGs between CA_COPD and COPD patients, enriched in immune-related processes including leukocyte activation, complement signaling, and B cell–mediated immunity, consistent with known host responses to Candida [21–23]. COPD-context-dependent Candida-responsive genes were primarily involved in inflammatory responses, protein processing, cell cycle regulation, and metabolic pathways such as NAD biosynthesis and amino acid metabolism. These observations align with prior evidence that PBMCs stimulated by Candida undergo rapid, coordinated responses, including activation of inflammatory signaling and cytokine secretion [24, 25]. Downregulated genes were predominantly associated with collagen trimers and cell cycle inhibitors, potentially reflecting Candida-induced collagen reorganization and enhanced immune cell proliferation [26–28]. Notably, the FoxO signaling pathway, which is mediated by FoxO transcription factors such as FoxO1 and FoxO3 [29], was suppressed. FoxO factors are critical regulators of immune homeostasis, controlling lymphocyte survival, Treg differentiation, B cell development, dendritic cell function, and oxidative stress resistance [30–32]. Reduced expression of FOXO1 and SIRT1 in PBMCs has been linked to more severe COPD [33], and polymorphisms in FOXO1 and FOXO3A are associated with disease susceptibility [34]. How Candida infection suppresses this pathway remains to be determined. PPI and TF analyses further highlighted hub genes such as IDO1 and MDM2, as well as TF families including zf-C2H2. IDO1 catalyzes the rate-limiting step of tryptophan catabolism, generating kynurenine, which modulates immune responses through AhR signaling and promotes immune tolerance [35–37]. Its activity has been correlated with COPD severity [38] and may contribute to the immune modulation observed during Candida infection. The zf-C2H2 TF family, one of the most abundant and versatile regulators [39, 40], was predicted to bind several DEGs, consistent with the reported involvement of certain members of this family in COPD pathology [41]. This suggests a potential mechanism for Candida-induced transcriptional remodeling, although further studies are required to confirm direct regulation. Our metabolomic profiling complemented the transcriptomic findings. COPD-context-dependent Candida-responsive metabolites showed upregulation of pathways related to amino sugar and nucleotide sugar metabolism, while downregulated metabolites were enriched in aminoacyl-tRNA biosynthesis, central carbon metabolism, protein digestion and absorption, and ABC transporter pathways. These data suggest that Candida infection perturbs energy metabolism, protein synthesis, and substrate transport, reinforcing its systemic effects in COPD. Previous studies have linked airway microbiome dysbiosis to enhanced inflammation and an increased risk of exacerbation [42], which is consistent with our observed upregulation of immune-related pathways in CA_COPD patients. However, our study extends current knowledge by providing mechanistic host-response data (transcriptomic and metabolomic) that are largely absent from descriptive microbiome studies. Specifically, we identify dysregulated pathways, such as tryptophan metabolism, FoxO signaling, and energy metabolism, that may mediate host responses to Candida colonisation in COPD. These findings are novel and have not been well characterized in previous COPD microbiome studies. Several limitations of this study should be noted. This is a pilot study, and the cohort size was relatively small. Clinical heterogeneity may have obscured subtle differences. The limited sample size also precludes meaningful statistical analysis of clinical outcome associations and restricts our ability to link molecular findings to adverse events. Patients were receiving various medications, which could introduce confounding effects. Finally, more genes from RNA-seq findings should be validated using qPCR or other orthogonal methods. Future studies with larger cohorts, longer follow-up, and functional validation are needed to confirm the causal role of these pathways, to assess the association between Candida status and clinically relevant endpoints (e.g., acute exacerbation, mortality), and to explore interventions targeting Candida-mediated immune-metabolic dysregulation and Future studies with dedicated microbiome sequencing are needed to fully characterize how bacterial and fungal communities interact with host immune-metabolic responses in Candida-colonised COPD patients. In conclusion, Candida infection in COPD patients may drive context-dependent transcriptional and metabolic reprogramming, likely contributing to disease severity and systemic inflammatory burden. These findings underscore the importance of monitoring fungal colonization in COPD and suggest potential therapeutic strategies targeting immune-metabolic pathways. Supplementary Information Supplementary Material 1: Table S1 Quantification method details for Table 1: additional files for details. Supplementary Material 2: Table S2 Comprehensive clinical information of the patients in the cohort. Supplementary Material 3: Table S3 Exclusive DEGs observed between the CA_COPD and COPD groups: additional files for details. Supplementary Material 4: Figure S1. Spearman correlation validation of gene-metabolite interactions. Acknowledgements Not applicable. Clinical trial number Not applicable. Abbreviations Animal Transcription Factor Database Chronic obstructive pulmonary disease Differentially expressed genes Global Initiative for Chronic Obstructive Lung Disease Gene Ontology Kyoto Encyclopedia of Genes and Genomes Lower limit of normal Least Significant Difference Protein-protein interaction Peripheral blood mononuclear cells Quality control Authors’ contributions Hui Yang and Ran Wang designed the research. Ran Wang, Zhibin Chen, Yanwei Wang, Kecong Li, and Haiyan Gong performed the experiments. Niangui Zhao, Yanna Zhang, Songyin Lai, and Zhangqiang Ye contributed to sample collection. Ran Wang and Zhibin Chen analyzed the data and wrote the paper. Hui Yang supervised the study and revised the manuscript. All authors read and approved the final manuscript. Funding This study Supported by Fujian Provincial Natural Science Foundation of China, Grant No. 2022J011390. Data availability The datasets generated and analyzed during the current study are available in the NCBI BioProject repository under the accession number PRJNA1445529. The metabolomics raw data have been deposited in the MetaboLights with the study identifier MTBLS14197. All other data supporting the findings of this study are available from the corresponding author upon reasonable request. Declarations Ethics approval and consent to participate Written informed consent was secured from all participants before inclusion. The study protocol was reviewed and approved by the Ethics Committee of The Second Affiliated Hospital of Xiamen Medical College (Approval No.2026036) and conducted in compliance with the Declaration of Helsinki. Consent for publication Not Applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Ran Wang and Zhibin Chen contributed equally to this work. References Associated Data Supplementary Materials Supplementary Material 1: Table S1 Quantification method details for Table 1: additional files for details. Supplementary Material 2: Table S2 Comprehensive clinical information of the patients in the cohort. Supplementary Material 3: Table S3 Exclusive DEGs observed between the CA_COPD and COPD groups: additional files for details. Supplementary Material 4: Figure S1. Spearman correlation validation of gene-metabolite interactions. Data Availability Statement The datasets generated and analyzed during the current study are available in the NCBI BioProject repository under the accession number PRJNA1445529. The metabolomics raw data have been deposited in the MetaboLights with the study identifier MTBLS14197. All other data supporting the findings of this study are available from the corresponding author upon reasonable request.