Systems biology of interstitial lung diseases: integration of mRNA and microRNA expression changes

Abstract

Background: The molecular pathways involved in the interstitial lung diseases (ILDs) are poorly understood. Systems biology approaches, with global expression data sets, were used to identify perturbed gene networks, to gain some understanding of the underlying mechanisms, and to develop specific hypotheses relevant to these chronic lung diseases.

Methods: Lung tissue samples from patients with different types of ILD were obtained from the Lung Tissue Research Consortium and total cell RNA was isolated. Global mRNA and microRNA were profiled by hybridization and amplification-based methods. Differentially expressed genes were compiled and used to identify critical signaling pathways and potential biomarkers. Modules of genes were identified that formed a regulatory network, and studies were performed on cultured cells in vitro for comparison with the in vivo results.

Results: By profiling mRNA and microRNA (miRNA) expression levels, we found subsets of differentially expressed genes that distinguished patients with ILDs from controls and that correlated with different disease stages and subtypes of ILDs. Network analysis, based on pathway databases, revealed several disease-associated gene modules, involving genes from the TGF-β, Wnt, focal adhesion, and smooth muscle actin pathways that are implicated in advancing fibrosis, a critical pathological process in ILDs. A more comprehensive approach was also adapted to construct a putative global gene regulatory network based on the perturbation of key regulatory elements, transcription factors and microRNAs. Our data underscores the importance of TGF-β signaling and the persistence of smooth muscle actin-containing fibroblasts in these diseases. We present evidence that, downstream of TGF-β signaling, microRNAs of the miR-23a cluster and the transcription factor Zeb1 could have roles in mediating an epithelial to mesenchymal transition (EMT) and the resultant persistence of mesenchymal cells in these diseases.

Conclusions: We present a comprehensive overview of the molecular networks perturbed in ILDs, discuss several potential key molecular regulatory circuits, and identify microRNA species that may play central roles in facilitating the progression of ILDs. These findings advance our understanding of these diseases at the molecular level, provide new molecular signatures in defining the specific characteristics of the diseases, suggest new hypotheses, and reveal new potential targets for therapeutic intervention.

Figure 1

Hierarchical clustering with 1423 differentially expressed genes between ILD patient and normal control lung samples. Each row represents the expression profile of a gene across 29 samples and each column represents a sample. The sample IDs and clinical information (FVC group and diagnosis) are listed below the heatmap. Red or green colors indicate either higher or lower expression levels of the gene. Based on the dendrogram, the samples can be further separated into four subgroups indicated by red, green, blue and black bars below the heatmap which correlate with ILD other than UIP/IPF, less severe UIP/IPF (FVC 2 and 3), more severe UIP/IPF (FVC 1), and normal control, respectively.

Figure 2

Hierarchical clustering result of 125 differentially expressed miRNAs. The sample IDs and clinical information (FVC group and diagnosis) are listed below the heatmap. Red or green colors indicate either higher or lower expression levels of the miRNA. Samples are well separated into control and ILD patient groups except two.

Figure 3

Integrated view of Wnt pathway. All (present) genes associated with the Wnt pathway are denoted with by circular nodes and their fold-change values are represented by node color where red and green indicate higher and lower expression in ILD patient samples compared to controls, respectively. Bold-faced gene symbols indicate DEGs. Triangular nodes represent DEmiRNAs which have reciprocal expression fold-change values relative to their presumptive mRNA targets or that are predicted to be regulated by differentially expressed transcription factors (FDR<0.1, see Additional file 1). The transcription factors used in construction of this regulatory network are shown as squares and their fold-changes were expressed by node border colors. Gray and black lines indicate protein-protein interactions and signaling information from the KEGG database, respectively. Red, blue and green lines with arrows represent predicted transcriptional activation, transcriptional repression and miRNA-mediated repression, respectively. Open circles represent transcription factors that were not used in prediction of regulatory interactions.

Figure 4

A hypothetical network based on differentially expressed transcription factors and DEmiRNAs. Molecular interactions obtained from protein-protein interactions, KEGG pathway interactions, putative transcriptional regulatory interactions derived from 1423 DEGs, and predicted DEmiRNA and DETF interactions were combined. (A). The network containing 689 nodes composed of 22 DETFs (open squares), 618 DEGs (filled circles) and 49 DEmiRNAs (filled triangles) and 1391 non-redundant interactions was generated. The network can be grouped into 7 modules (as indicated) based on the connectivity of nodes - the more interacted nodes are grouped together. (B) This figure displays only the network consisting of the DETF (with known binding sites) and their cognate DEmiRNAs. The transcription factors, which do not have well characterized binding site information were not used in prediction of interactions, and are shown as open circles.

Figure 5

Examples of potential feed forward regulatory loops. A potential feed forward loop (FFL) was generated if a DEmiRNA and its predicted regulatory transcriptional factor targeted the same DEG (see Additional file 1). Red, blue and gray lines represent the predicted transcriptional activation (i.e. positive correlation), predicted transcriptional repression (i.e. negative correlation) and relationships between DEmiRNAs and their predicted targets (without considering expression correlations), respectively. The left inserts illustrate coherent and incoherent FFLs. JUN, miR-195 and AXUD1 form a coherent FFL in which transcriptional (from JUN to AXUD1) and miRNA-mediated regulation (from JUN to AXUD1 via miR-195) are synergistic. In contrast, NFATC4, miR-29b and COL3A1 form an incoherent FFL in which two opposite regulatory interactions occur.

Figure 6

The Zeb-1 mediated EMT is partially regulated by miRNAs in the miR-23a cluster. (A-C) Phase contrast images showing the morphology of an epithelial MDCK clone (A; 3E11), the same epithelial clone stably over-expressing Zeb-1 (B; 3E11+Zeb1), or a mesenchymal MDCK clone (C; 2F7). (D) Over-expression of Zeb-1 causes the levels of miR-23a, miR-24 and miR-27a to rise dramatically, while over-expression of the miR-23a cluster of genes had no effect on the level of Zeb-1 mRNA, relative to mock-transfected cells. (E) Nedd4L Western: MDCK clone 3E11 has an epithelial morphology and expresses 4 bands that are recognized by an anti-Nedd4l antibody (green). One of these bands (marked with asterisk) was dramatically reduced in MDCK cells with mesenchymal morphology (clone 2F7) or in an epithelial MDCK cell line stably over-expressing Zeb-1 (3E11+Zeb1). The significance of each of the four bands recognized by the anti-Nedd4L antibody remains unclear. However, a similar pattern of bands was observed using a second anti-Nedd4L antibody (data not shown). Anti-Argonaute 2 (Ago2; red) antibody was used as loading control. (F) A hypothetical model for the role of the Zeb-1, the miR-23a cluster, and Nedd4L in TGF-β mediated EMT. The genes and miRNAs involved in the process are listed and the colors indicate the relative expression changes in ILD samples compared to control, red indicates higher level in ILDs, yellow indicates no significant changes, while green represents lower levels in ILD samples compared to control.

Figure 7

Examples of sub-networks associated with different patient classifications. Examples of sub-networks that show significant changes in some key genes associated with FVC groups 1 and 2 (A), MYOCD (B), or ZEB 1 (C) expression levels. The corresponding FVC group and MYOCD or ZEB 1 genes are labeled in the sub-networks. The levels of expression are indicated as in the previous figures.