Transcriptional survey of alveolar macrophages in a murine model of chronic granulomatous inflammation reveals common themes with human sarcoidosis

Abstract

Mohan A, Malur A, McPeek M, Barna BP, Schnapp LM, Thomassen MJ, Gharib SA. Transcriptional survey of alveolar macrophages in a murine model of chronic granulomatous inflammation reveals common themes with human sarcoidosis. Am J Physiol Lung Cell Mol Physiol 314: L617-L625, 2018. First published December 6, 2017; doi: 10.1152/ajplung.00289.2017 . To advance our understanding of the pathobiology of sarcoidosis, we developed a multiwall carbon nanotube (MWCNT)-based murine model that shows marked histological and inflammatory signal similarities to this disease. In this study, we compared the alveolar macrophage transcriptional signatures of our animal model with human sarcoidosis to identify overlapping molecular programs. Whole genome microarrays were used to assess gene expression of alveolar macrophages in six MWCNT-exposed and six control animals. The results were compared with the transcriptional profiles of alveolar immune cells in 15 sarcoidosis patients and 12 healthy humans. Rigorous statistical methods were used to identify differentially expressed genes. To better elucidate activated pathways, integrated network and gene set enrichment analysis (GSEA) was performed. We identified over 1,000 differentially expressed between control and MWCNT mice. Gene ontology functional analysis showed overrepresentation of processes primarily involved in immunity and inflammation in MCWNT mice. Applying GSEA to both mouse and human samples revealed upregulation of 92 gene sets in MWCNT mice and 142 gene sets in sarcoidosis patients. Commonly activated pathways in both MWCNT mice and sarcoidosis included adaptive immunity, T-cell signaling, IL-12/IL-17 signaling, and oxidative phosphorylation. Differences in gene set enrichment between MWCNT mice and sarcoidosis patients were also observed. We applied network analysis to differentially expressed genes common between the MWCNT model and sarcoidosis to identify key drivers of disease. In conclusion, an integrated network and transcriptomics approach revealed substantial functional similarities between a murine model and human sarcoidosis particularly with respect to activation of immune-specific pathways.

Keywords: alveolar macrophage; animal model; gene network; microarray; sarcoidosis.

Fig. 1.

Representative human and murine histological sections showing granulomatous and normal alveolar architecture (×40 magnification). A: human granulomatous tissue from a sarcoidosis patient. B: healthy human lung tissue. C: carbon nanotube-induced murine granulomatous lung tissue 60 days postinstillation of multiwall carbon nanotube (MWCNT); arrow indicates formation of a giant cell. D: murine control lung tissue. Note that both human and mouse granulomatous lesions show aggregates of macrophage infiltrates; however the carbon nanotubes are only present in murine lesions. Red bar = 10 μm.

Fig. 2.

Correspondence analysis demonstrates distinct segregation between control (cyan) and MWCNT-exposed (magenta) mice, implying profound alterations in global gene expression of alveolar macrophages in this model of granulomatous lung disease.

Fig. 3.

Heatmap depiction of over 1,000 differentially expressed genes [false discovery rate (FDR) <0.01] between MWCNT vs. control mice. Gene ontology (GO) analysis revealed overrepresentation of many immune-related categories among upregulated genes in MWCNT mice, including T-cell regulation and activation, immune cell differentiation, as well as cytokine binding and activity. Complete lists of differentially expressed genes and GO categories are provided in Supplemental Tables S2–S4.

Fig. 4.

Integrated gene set enrichment analysis (GSEA) and network analysis of upregulated processes in alveolar macrophages of mice exposed to chronic MWCNT. In the figure, each sphere designates an enriched gene set with its size being proportional to the number of its gene members. Since biological pathways share many common genes, connectivity lines have been drawn to show these interpathway relationships. Note that the topology of the enrichment network is characterized by groupings of highly connected gene sets based on the extent of overlap among their member genes. These modules have distinctive functional attributes including immunity, oxidative phosphorylation, extracellular matrix, cell cycle, and metabolism. A few selective pathways in each module have been labeled, with the complete list available in Supplemental Table S5.

Fig. 5.

GSEA membership depiction of common and divergent enrichment of pathways between the MWCNT murine model and human sarcoidosis. Representative upregulated gene sets are shown, highlighting overlapping functional themes between the animal model and sarcoidosis, but also demonstrating differences. Each enriched pathway (gray box) was significant at FDR < 0.05. Complete list of pathways are provided in Supplemental Tables S5 and S6.

Fig. 6.

Network analysis of differentially expressed genes common to human sarcoidosis and the MWCNT murine model. The topology of this gene product interactome is highlighted by several densely connected nodes involved in immunity including interferon-γ (IFNG), signal transducer and activator of transcription 4 (STAT4), and lymphocyte protein tyrosine kinase (LCK) but also includes matrix remodeling genes such as cathepsin K (CTSK) and macrophage elastase (MMP12). This cross species network may represent putative critical drivers of chronic granulomatous inflammation.

Fig. 7.

Quantitative PCR validation of 4 network genes (from Fig. 6) identified as commonly upregulated in mouse MWCNT model and human sarcoidosis. Alveolar macrophages from an independent set of MWCNT exposures and controls (n = 4 mice per group; A) and bronchoalveolar lavage cells from sarcoidosis patients and healthy controls (n = 6–7 subjects per group; B) were used for confirmation.