Swine influenza-modified pulmonary microbiota

Abstract

Influenza A virus (IAV) remains a major health concern in both humans and animals, with pigs serving as key reservoirs for generating novel reassortant viruses with pandemic potential. Respiratory microbiome alterations during infection may facilitate secondary bacterial complications. This study investigates the lung microbiota of pigs naturally infected with IAV across different regions in Spain, using Oxford Nanopore Technologies (ONT) long-read 16S rRNA sequencing to characterize associated bacterial communities. Our results show a higher bacterial genus diversity in IAV-infected animals compared to healthy controls, with significant differences in both presence and relative abundance of bacterial taxa. Infected lungs exhibited increased proportions of potential pathogens, particularly Glaesserella spp., detected in approximately 60% of infected samples, often as the dominant genus. Other pathogenic genera, including Pasteurella, Staphylococcus, Mycoplasma, and Fusobacterium, were also strongly associated with infection. Clustering analyses revealed distinct microbial profiles that clearly separated infected from non-infected animals, identifying specific bacterial signatures predictive of infection status. These findings suggest that IAV infection significantly alters the pulmonary microbiota, potentially creating a permissive environment for secondary bacterial infections. This study underscores the relevance of microbiota shifts during IAV infection in swine and highlights the importance of understanding microbial dynamics in respiratory disease progression. Additionally, we present a novel, rapid, and practical experimental pipeline based on ONT long-read sequencing to investigate the respiratory microbiota in swine infection models. This approach offers a valuable tool for future research and potential diagnostic applications in both veterinary and human medicine.

Keywords: Oxford Nanopore; coinfection; influenza virus; lung; pigs; respiratory microbiome; sequencing; swine.

Figure 1

Graphical abstract. (a) Schematic and Graphical Representation of Experimental Methodology. A piece of lung from all infected pigs was extracted, embedded in PBS (Phosphate Buffered Saline) (1) and crushed (2) using a TissueLyser. Then, DNA was extracted (3) using Qiagen extraction kit and amplified (4) with barcoded 16S primers from Oxford Nanopore Technologies. Purification of the DNA (5) was performed using magnetic beads for the library preparation and sequencing (6). Finally, data were analyzed (7). (b) Number and origin of the pigs analyzed. 92 samples were processed and analyzed: 53 were Influenza-infected and 39 were (non-infected) - healthy animals. Figure created with BioRender.com.

Figure 2

Relative bacterial abundance in the lungs of Influenza virus-infected and non-infected swine. (a) Individual 16S/18S gene ratio obtained by qPCR for 39 healthy (green) and 53 infected (red) DNA purified from lung necropsy samples. LOD: Limit of Detection (b) Stacked bar average relative abundance (%) of bacteria present in healthy (Influenza confirmed negative) pigs and Infected with influenza virus for phylum, class, order, family, and genus. (c) Violin plots of the main class groups classification of bacteria (Bacilli, Betaproteobacteria, Gammaproteobacteria, and Mollicutes) in Healthy vs Infected animals diagnosed with Influenza virus. (d) Venn diagram of bacteria in healthy or infected pigs at the genus level. (e) Individual bacterial stacked representation of genus relative abundance (%) ordered in terms of abundance. (f) Violin plots of relevant genus groups classification of bacteria (Streptococcus, Escherichia, Glaesserella, and Pasteurella) in healthy vs infected samples from Influenza virus-diagnosed swine. Statistical comparisons were performed using the Mann–Whitney U test and ANCOM-BC microbiome test. Significance levels are indicated as follows: · p ≈ 0.05, *p < 0.05; **p < 0.01; ***p < 0.001. ns, Non-signitican diferences.

Figure 3

Differential bacterial proportions and correlation networks. (a) Comparisons of the relative proportion (%) in Influenza-virus infected and healthy groups at the level of bacterial genus. (b) Partitioning Around Medoids (PAM) clustering analysis. (c) Neural network of significant bacteria based on Pearson correlation and Louvain clusterization. Each node for (b, c) represents a bacterial genus, separated by distances and grouped following clusterization methods.