Exploring the host response in infected lung organoids using NanoString technology: A statistical analysis of gene expression data

Abstract

In this study, we used a three-dimensional airway "organ tissue equivalent" (OTE) model at an air-liquid interface (ALI) to mimic human airways. We investigated the effects of three viruses (Influenza A virus (IAV), Human metapneumovirus (MPV), and Parainfluenza virus type 3 (PIV3) on this model, incorporating various control conditions for data integrity. Our primary objective was to assess gene expression using the NanoString platform in OTE models infected with these viruses at 24- and 72-hour intervals, focusing on 773 specific genes. To enhance the comprehensiveness of our analysis, we introduced a novel algorithm, namely MAS (Magnitude-Altitude Score). This innovative approach uniquely combines biological significance, as indicated by fold changes in gene expression, with statistical rigor, as represented by adjusted p-values. By incorporating both dimensions, MAS ensures that the genes identified as differentially expressed are not mere statistical artifacts but hold genuine biological relevance, providing a more holistic understanding of the airway tissue response to viral infections. Our results unveiled distinct patterns of gene expression in response to viral infections. At 24 hours post-IAV infection, a robust interferon-stimulated gene (ISG) response was evident, marked by the upregulation of key genes including IFIT2, RSAD2, IFIT3, IFNL1, IFIT1, IFNB1, ISG15, OAS2, OASL, and MX1, collectively highlighting a formidable antiviral defense. MPV infection at the same time point displayed a dual innate and adaptive immune response, with highly expressed ISGs, immune cell recruitment signaled by CXCL10, and early adaptive immune engagement indicated by TXK and CD79A. In contrast, PIV3 infection at 24 hours triggered a transcriptional response dominated by ISGs, active immune cell recruitment through CXCL10, and inflammation modulation through OSM. The picture evolved at 72 hours post-infection. For IAV, ISGs and immune responses persisted, suggesting a sustained impact. MPV infection at this time point showed a shift towards IL17A and genes related to cellular signaling and immune responses, indicating adaptation to the viral challenge over time. In the case of PIV3, the transcriptional response remained interferon-centric, indicating a mature antiviral state. Our analysis underscored the pivotal role of ISGs across all infections and time points, emphasizing their universal significance in antiviral defense. Temporal shifts in gene expression indicative of adaptation and fine-tuning of the immune response. Additionally, the identification of shared and unique genes unveiled host-specific responses to specific pathogens. IAV exerted a sustained impact on genes from the initial 24 hours, while PIV3 displayed a delayed yet substantial genomic response, suggestive of a gradual and nuanced strategy.

Fig 1. PCA-based visualization of all 96 OTEs.

Fig 2. Volcano plots displaying BH-significant genes identified using Algorithm 1 in S1 File, with Mock-24/72 as the baseline and IAV/MPV/PIV3-24/72 as treated groups.

The top 10 MAS-ranked genes for each contrast are highlighted.

Fig 3. Comparative analysis of BH-significant genes using Algorithm 1 in S1 File with different baselines: Mock-infected and UV-infected groups.

(a) Display of shared and unique BH-significant genes between (Mock vs. active) and (UV-infected vs. active-infected) contrasts, along with the top 10 MAS-selected genes for all six comparisons. (b) Representation of the consistency among shared BH-significant genes in relation to the sign of their Log Fold Change.

Fig 4. Application of Algorithm 2 in S1 File to identify three distinctive genes pivotal for classification.

Fig 5. Heatmap representation of log2-transformed expression levels for the top 10 genes from MAS-Common-genes-24, measured at 24- and 72-hours post-infection.

The conditions are ordered vertically by descending average expression of these selected genes.

Fig 6. Comparative analysis of MAS-selected genes following IAV and PIV3 infections at 24- and 72-hours post-infection.

Panel (a) showcases the top 10 common and unique genes, while panels (b-c) illustrate gene correlation networks at both time points with significant Spearman correlations. The color bar represents connectivity degree within the selected genes.

Fig 7. Log2-transformed expression of IFIT2, IFIT1, and IL36G, highlighting significant commonalities between IAV and PIV3 at 24 and 72 hours, and showcasing IL36G’s time-specific expression (see Fig 2 for contrast with Mock samples).