Morphological changes of lung in different groups
HE staining indicated that the morphology of pulmonary alveoli in control group was regular on day 7 with even sizes (Fig. 1a). On day 14, the structure of pulmonary alveoli was normal, and there was narrowing in the alveolar septum in control (Fig. 1b). On day 21, there was increase in number of pulmonary alveoli, together with narrowing of alveolar septum. There were no aberrant changes in terminal bronchus in control group (Fig. 1c). For the HE staining in BPD model group, part of pulmonary alveoli showed fusion on day 7, combined with infiltration of inflammatory cells (Fig. 1a). On day 14, the number of pulmonary alveoli showed decrease and the structure was not regular. There was massive interstitial cell hyperplasia, together with thickening in alveolar septum (Fig. 1b). On day 21, pulmonary alveoli were no longer available, and there was obvious dilatation in terminal bronchus. In addition, structural disorder was noticed in pulmonary tissues, indicating block in pulmonary development (Fig. 1c).
Alpha diversity
Multi-sample Shannon curves indicated that the data volume was adequate for the sequencing, and the sample traits would not increase with the elevation of sequencing volume (Fig. 2a). Alpha diversity was analyzed to evaluated overall differences between the gut microbiota in model group and control group. The ACE index showed no statistical differences in richness of gut microbiota between control group and model group on day 7, 14, and 21, respectively (p>0.05, Fig. 2b). PD_whole_tree index evaluated the diversity of gut microbiota, homogeneity and evolution. There were no statistical differences between control group and model group on day 7 and 21 (p>0.05). The difference between the two groups was statistically significant on day 14 (p<0.05, Fig. 2c). This implied that there were no changes in gut microbiota richness in BPD mice compared with control, however, the genetic diversity showed statistical differences between the two groups.
Beta diversity
In this section, Beta diversity analysis was performed based on un-weighted unifrac distance. PCoA plot showed that there was no obvious separation between two groups on day 7 and day 21, respectively. In contrast, there was significant separation of PC1 between two groups on day 14 (Fig. 3a-3c). Analysis of similarities indicated that there was no significant difference in gut microbiota between two groups on day 7 (R=-0.028, p=0.628), while the difference was statistically significant between two groups on day 14 (R=0.368, p=0.021). On day 21, there was no difference in gut microbiota between two groups (R=0.188, p=0.079, Table 1).
Difference analysis
LDA analysis was performed to investigate the biomarkers, and all biomarkers had LDA scores of higher than 4. There were three biomarkers with statistical differences at all biological levels on day 7, all of which were in the control group (p<0.05). On day 14, there were statistical differences in 22 biomarkers at each biological level, among which 16 were enriched in BPD group and 6 were enriched in control group (p<0.05, Fig. 4a and 4b). On day 21, there were statistical differences in four biomarkers at all biological levels, all of which were in BPD group (Fig. 4c).
On day 14, the relative abundance of intestinal microbiota showed that the proportion of Firmicutes, Bacteroidetes and Proteobacteria in phylum level was higher than 80% (Fig. 5). Analysis of variance indicated that the relative richness of Bacteroidetes in model group was significantly lower than that of control group (11.6% vs. 54.8%, PFDR<0.01), while the relative richness of Proteobacteria in model group was significantly higher than that of control group (29.8% vs. 5.1%, PFDR <0.05). This was consistent with the LDA results. The relative richness of Cyanobacteria, Acidobacteria, Chloroflexi, Rokubacteria, Epsilonbacteraeot, Nitrospirae and Gemmatimonadetes in model group was significantly higher than that of control (all PFDR<0.05, Table 2).
Functional prediction
A total of 17 KEGG metabolic pathways associated with intestinal microbiota were significantly different between the two groups. Three of them were enriched in the model group and 14 were enriched in the control group. Signal transduction (PFDR=0.037), glycan biosynthesis and metabolism (PFDR=0.032), metabolism of terpenoids and polyketides (PFDR=0.049) are the top three metabolic pathways with statistically significant differences (Fig. 6).