The sputum samples were used for pathogen testing by one-step RT-PCR using multiple respiratory pathogen detection kit (Health gene technologies, China) comprising 13 types of pathogens, including Influenza A virus (H7N9, H1N1, H3N2 and H5N2), influenza A virus H1N1 (2009), seasonal H3N2 virus, influenza B virus (Victoria strain and Yamagata strain), adenovirus (group B, group C and E), human bocavirus, rhinovirus, parainfluenza virus (type 1, 2, 3 and 4), coronavirus (type 229E, OC43, NL63 and HKU1), respiratory syncytial virus (Group A and B), metapneumovirus, Mycoplasma pneumoniae, and Chlamydia (Chlamydia trachomatis and Chlamydia pneumoniae).

Pediatric patients admitted to Suzhou Children’s Hospital were enrolled and categorized into two groups: IAV group and control group. In this study, patients met the following criteria: 1) presence of relevant epidemiological history and clinical symptoms; 2) confirmation of IAV positivity using a combination of RT-PCR and capillary electrophoresis. The control group comprised pediatric patients who tested negative for 13 types of pathogens. Exclusion criteria encompassed: 1) Pre-existing diseases affecting the respiratory microbiota and metabolome, such as asthma and cystic fibrosis; 2) use of medications prior to enrollment known to impact the respiratory microbiome and metabolome, such as immunosuppressants, probiotics, and hormones; 3) pediatric patients who declined laboratory tests after admission. A total of 85 IAV cases and 55 control were enrolled, with an average age of about 4 years. There no significant differences in age, gender or related clinical diagnosis between the two groups ( Supplementary Table S1 ). This study was approved by the Ethics Committee of Children’s Hospital of Soochow University (Approval No.2023 C143).

Respiratory specimens from the upper respiratory tract (URT) and the lower respiratory tract (LRT) were collected from IAV and control groups within 24 hours after admission. During the collection of specimens from URT, 5 mL normal saline was injected into one nostril with a pipette and the washings were collected in a dish or beaker. Due to the young age of children, it is difficult to obtain LRT specimens by alveolar lavage. The importance of oral cleanliness and deep cough were fully explained to the patients to avoid oropharyngeal bacterial contamination, and the patients were instructed how to correctly collect sputum specimens. The first sputum from pediatric patients after repeated gargle with water were collected covered to avoid the spread of microbiota. The collected specimens were immediately sent for IAV analysis or stored at -80°C for further microbiota analysis.

Fasting venous blood were collected on the morning after admission. Well-mixed anticoagulated whole blood was collected for blood routine tests using an Automatic Blood Cell analyzer BC-7500CRP. After erythrocytes were lysed with BD FACS hemolysin, the lymphocytes were stained with the lymphocyte subgroup detection Reagent (BD Multitest 6- Color TBNK Reagent), including CD3-FITC, CD4-PE-Cy7, CD8-APC-Cy7, CD56-PE, CD16-PE, CD19-APC, CD45-PerCP-Cy5.5, and analyzed by flow cytometry (BD FACSLyric). A Biochemical Analyzer ADVIA 2400 was employed to analyze the serum chemistry.

Total DNA was extracted from the collected respiratory samples. The quality, concentration and purity of DNA were detected by 1% agarose gel electrophoresis and instruments. 27F (5’-AGAGTTTGATCCTGGCTCAG-3’) and 1492R (5’-GGTTACCTTGTTACGACTT-3’) were used to amplify the full-length 16S rRNA gene. PCR products were purified by DNA gel recovery and purification kit (Majorbio, China) and quantified by Synergy HTX (Majorbio, China). Amplicon libraries were confirmed using the NEXTFLEX Rapid DNA-Seq Kit (Bioo Scientific, USA) prior to sequencing eligible libraries. Sequencing was performed using Illumina’s MiSeq (PE300).

Quality control of the original sequencing data was conducted using fastp (https://github.com/OpenGene/fastp, version 0.20.0), while FLASH (http://www.cbcb.umd.edu/software/flash, version 1.2.11) was employed for sequence merging. Noise reduction of optimized sequences after quality control was performed using the divisive amplicon denoising algorithm 2 (DADA2) in the Qiime2 pipeline, resulting in amplicon sequence variants (ASVs). To minimize the impact of sequencing depth on subsequent analyses of alpha and beta diversity, all sample sequences were rarefied to 10,000. Taxonomic classification of ASVs was performed using the Naive Bayes classifier in Qiime2, referencing the Silva 16S rRNA gene database (v138). Subsequently, the data was analyzed on the online platform of Majorbio (www.majorbio.com). Phylogenetic Investigation of Communities by Reconstruction of Unobserved States 2 (PICRUSt2) was used to predict the functional information of microbial communities, and Kyoto Encyclopedia of Genes and Genomes (KEGG) combined with STAMP software was used to analyze the species and functional composition and differences.

The SCIEX ZenoTOF 7600 system was employed to analyze serum metabolic profiling. The analytical column used was a Phenomenex F5 (2.1×100 mm, 1.8, Phenomenex, Castel Maggiore, Italy), housed in a compartment maintained at 45°C ± 1°C.

The autosampler temperature was maintained 10°C with an injection volume of 2 µL. Between each sample injection, the autosampler syringe was flushed for 5 s with a 2:1:1 (v/v) solution composed of water, methanol, and isopropyl alcohol. The gradient program was as follows: 0% B for 0–2 min, increased to 95% B from 2–14 min, held at 95% B from 14–16 min, decreased to 0% B from 16–16.1 min, and maintained at 0% B from 16.1–20 min to re-equilibrate the column. Electrospray ionization (ESI) was operated in both positive and negative ion modes with the following parameters: spray voltage +5 kV (positive) and -4.5 kV (negative), capillary temperature 550°C, gas1 50, gas2 50, and curtain gas 25. Data acquisition was performed in full scan mode (m/z 60–1200 Da) with an accumulation time of 0.2 s for MS1 and 15 ms for MS2 (m/z 3–1200 Da).

The stability of the instrument was checked by quality control samples (QC). Simca14.0 software was used for systematic analysis, such as orthogonal partial least squares discriminant analysis (OPLS-DA), to obtain variable weight importance ranking (VIP). The statistically significant differential metabolites were identified according to VIP> 1, P value< 0.05, and fold change> 1.2. The metabolic pathways were annotated using MetaboAnalyst 6.0 (www.metaboanalyst.ca) and KEGG database.

The data were analyzed by R software, and the comparison between the two groups was performed by unpaired two-tailed t test. Wilcox rank sum test was used to compare the difference of bacterium between the two groups. Correlations between metagenomics and metabolomics data were analyzed using the Mantel test and Spearman’s rank test. All data are presented as mean± SEM unless otherwise stated. P< 0.05 was considered statistically significant.

Compared with the control group, the children with IAV infection showed a significantly higher red blood cell (RBC) count and a markedly lower platelet (Plt) count. This reduction in Plt count was also observed within the H1N1 and H3N2 subtypes ( Figures 1A, B ). White blood cells (WBC), lymphocytes (LY), eosinophils (EO) and basophils (BA) counts were decreased after IAV infection, including H1N1 and H3N2 ( Figures 1C, D ). The percentage of plateletcrit (Pct) was decreased after both H1N1 and H3N2 infection, the proportion of EO and BA was decreased after IAV infection, especially H1N1 ( Figures 1E, F ). The number of B cells (CD3^-CD19⁺) was decreased in IAV group T cells (CD3⁺), CD4⁺ T cells, CD8⁺ T cells were significantly decreased in both H1N1 and H3N2 subgroups ( Figures 1G, H ). FCM analysis also revealed the significant reduction of CD4⁺ T cells in IVA-infected patients ( Figure 1I ). After IAV infection, children with bronchopneumonia (BPI) have higher counts of BA, Plt and CD8⁺ T cell ( Supplementary Figure S1A ) and percentage of Pct, T cells and CD4⁺ T cells ( Supplementary Figures S1B-D ) than those with URT infections (UI). Compared with UI group, NK cells (CD3^-CD(16 + 56)⁺) were increased in BPI group ( Supplementary Figure S1E ). Although, IAV infection failed to change the concentration of CRP, IAV-infected boys have lower CRP than girls ( Supplementary Figure S1F ). The concentrations of NE, MO, Hgb, Hct, MCV, MCH, MCHC, RDW, MPV, PDW, CD4⁺CD8⁺ and CD19⁺CD23⁺ in IAV patients were not statistically different from those of the control group ( Supplementary Tables S2-S5 ).

The concentrations of total protein (TP) and albumin (ALB) were significantly increased after IAV infection, while was decreased in both H1N1 and H3N2 subgroups ( Figure 2A ). The concentration of triglyceride (TG) was decreased in patients with IAV infection, C3 complement (C3) and calcium (Ca) was significantly decreased in both H1N1 and H3N2 subgroups ( Figures 2B-D ). The content of direct bilirubin (DBIL) was decreased after both H1N1 and H3N2 infection, while the total bilirubin (TBIL) and indirect bilirubin (IBIL) were decreased after H1N1 infection ( Figures 2E, F ). The concentration of lipase (LPS) was increased and gamma-glutamyltransferase (GGT) was decreased after IAV infection, especially those with H1N1 infection, respectively ( Figures 2G, H ). However, children with H3N2 infection displayed lower concentrations of alanine transferase (ALT) and lactate dehydrogenase (LDH). In addition, children with H3N2 subtype had lower LDH and creatine kinase (CK) contents than those with H1N1 infection ( Figures 2G, H ). After IAV infection, boys had lower C3, C4, sCRP and higher LPS than girls ( Supplementary Figures S2A-C ). Compared with BPI group, the CREA was increased in UI group, which was absent from IAV infection ( Supplementary Figure S2D ). The concentrations of PA, GB, A:G, AST, ALP, CHE, UREA, UA, HBDH, Mg, TCHOL, CG, IgA, IgG, IgM in IAV patients were not statistically different from those of the control group ( Supplementary Tables S6, S7 ).

Principal components analysis (PCA) of serum metabolomics showed good clustering of QC samples, indicating high data stability (Fig.S3A). The Orthogonal Partial Least Squares-Discriminant Analysis (OPLS-DA) plots revealed different metabolic profile between control and IAV groups ( Figure 3A ). There were 209 differential metabolites identified between control and IAV groups. Pediatric patients infected by H1N1 and H3N2 also had different serum metabolic profiles ( Figure 3B ), and 37 differential metabolites were observed. IAV infection significantly altered 6 pathways, including steroid hormone biosynthesis, primary bile acid biosynthesis, purine metabolism, alanine, aspartate and glutamate metabolism, pantothenate and COA biosynthesis and biosynthesis of unsaturated fatty acids ( Figure 3C ). Urea, adenosine, deoxyguanosine, inosine and guanosine were outlier upregulated metabolites in IAV-infected patients ( Figure 3D ). Valine, leucine and isoleucine biosynthesis was the only overlapping pathway enriched in both H1N1 and H3N2 infection ( Supplementary Figures S3B, C ). Compared with H1N1-infected patients, purine metabolism was a key pathway altered in H3N2-infected patients ( Figure 3E ). Moreover, serum concentrations of xanthine, deoxyguanosine, inosine were significantly upregulated by H3N2 infection in comparison with H1N1 infection ( Figure 3F ).

16S rRNA sequencing analysis of sputum and nasal lavage fluid was performed to determine the potential changes in the respiratory microbiome caused by IAV. Compared with the control group, the increase of Chao index in IAV group indicated an increase in microbial richness ( Figure 4A ; Supplementary Figures S4A, B ). PLS-DA analysis showed that the structure of microbiota from URT and LRT of the control group was different from that of IAV patients, including H1N1 and H3N2 ( Figures 4B, C ; Supplementary Figures S4C, D ). However, there were group differences in the proportion of bacterial communities represented by these dominant phylum. The proportion of Bacteroidetes in the IAV group was smaller than that in the control group, but the amount of Actinobacteria in the IAV group was higher than that in the control group ( Figure 4D ). The H3N2 subtype was more prevalent in Proteobacteria and fewer in Firmicutes. Compared with the control group, the amount of Actinobacteria was increased and the amount of Bacteroidetes was decreased in the H1N1 group ( Figure 4E ). In the URT, comparisons between groups showed multiple genus-level differences in bacterial taxa. Compared with the Con group, the IAV group had significantly higher abundance of Moraxella and Haemophilus. When comparing the two influenza subtypes, the abundance of Ureaplasma was higher in H3N2 group than in the H1N1 group. In the LRT, an increased abundance of Moraxella was also observed ( Supplementary Figures S4D-H ).

PICRUST2.0 was used to predict the functional characteristics of respiratory microbiota. In the URT, 180 metabolic pathways were identified between the control group and the IAV group, characterized by purine metabolism ( Supplementary Figure S5A ). There 175 and 58 metabolic pathways were identified in H1N1 and H3N2 group, respectively ( Supplementary Figures S5B, C ). 40 overlapped pathways were found between H1N1 and H3N2 group ( Supplementary Figure S5D ). Compared with H1N1 group, 14 pathways were identified in H3N2 group ( Supplementary Figure S5E ). There 34 metabolic pathways were significantly disrupted in the LRT ( Supplementary Figure S5F ).

The Mantel test was performed to assess the relationship between different respiratory bacterial genera and their differential metabolites. Moraxella in the URT was positively correlated with several metabolites between the control and IAV groups. In addition, the relative abundance of Haemophilus was proportional to inosine respectively ( Figure 5A ). Ureaplasma in the URT was found to be proportional to inosine between H1N1 and H3N2 groups ( Figure 5B ). Moraxella, Acinetobacter and Haemophilus in the URT were not associated with any differential metabolites between control and H1N1 groups ( Supplementary Figure S6A ). The relative abundance of Acinetobacter was proportional to glutathione disulfide, beta-citryl-L-glutamate, 3-carboxy-1-hydroxypropyl-thPP between the control and H3N2. Moraxella was proportional to citric acid, beta-citryl-L-glutamate ( Supplementary Figure S6B ). Only Moraxella in LRT was associated with serum differential metabolites between the control and IAV groups ( Supplementary Figure S6C ).