RNA-sequencing analysis of lung primary fibroblast response to eosinophil-degranulation products predicts downstream effects on inflammation, tissue remodeling and lipid metabolism

Abstract

Background: The association of eosinophils with inflammation and tissue remodeling is at least partially due to their release of toxic granule proteins and other mediators, including cytokines. Tissue remodeling and consequent functional defects are affected by activity of connective tissue fibroblasts. Exaggerated fibroblast activation, accumulation and change of phenotype may lead to fibrosis and loss of tissue function. So far, little information has been reported on how eosinophils affect inflammation and tissue remodeling via the activation of fibroblasts. We have recently shown that eosinophil activation with IL-3 led to a robust eosinophil degranulation on immunoglobin-G (IgG) coated plates. Thus, in the present study, we analyze the effects of IL-3-activated eosinophil degranulation products on primary human lung fibroblasts (HLF) using whole transcriptome sequencing.

Methods: Conditioned media was obtained from eosinophils that were pre-activated with IL-3 or IL-5 and subsequently cultured for 6 h on IgG to induce degranulation. This conditioned media was added on human lung fibroblasts (HLF) for 24 h and the cell lysates were then subjected to whole transcriptome sequencing to identify global changes in gene expression. Differentially expressed genes were analyzed using the Ingenuity Pathway Analysis (IPA), and validated by qPCR.

Results: In HLF, the expression level of 300 genes was changed by conditioned media from IL-3-activated eosinophils compared to control fibroblast cultures. Among these 300 genes, the expression level of 35 genes coding for known proteins was upregulated by IL-3- versus IL-5-pre-activated eosinophils. Of the 35 upregulated genes, IPA identified C3, CH25H, CXCL1, CXCL8, CYP1A1, ICAM1, IL6 and UCN2 as having downstream functions on inflammation, tissue remodeling and lipid synthesis. This analysis combined with previous RNA sequencing analyses of eosinophils suggest IL-1ß, OSM and TNFSF12 as potential upstream regulators of fibroblasts.

Conclusions: This study has identified several novel pro-inflammatory and pro-remodeling mediators produced by fibroblasts in response to activated eosinophils. These findings may have significant implications on the role of eosinophil/fibroblast interactions in eosinophilic disorders.

Keywords: Degranulation; Eosinophil; Fibroblast; IgG; Il-3; Inflammation; Lipid metabolism; Lung; RNA-sequencing; Tissue remodeling.

Fig. 1

Evaluation of eosinophil degranulation by the measurement of EDN in eosinophil conditioned media. To prepare the eosinophil conditioned media that will be tested on human lung fibroblasts, human blood eosinophils were activated with IL-3 or IL-5 for 20 h and were then added on heat-aggregated (HA) human serum IgG for 6 h. Three conditioned media were prepared, including two conditioned media from IL-3-activated eosinophils, cultured for 6 h on coated HA-IgG (IL3IgG) or uncoated wells (IL3). The third conditioned medium was from IL-5-activated eosinophils seeded for 6 h on HA-IgG (IL5IgG). EDN released from eosinophils in the three conditioned media was measured by ELISA. The graph is an average ± SEM of three different cultures from three different eosinophil donors. The three groups were statistically different from each other as determine using the One-Way Analysis of Variance followed by the Holm-Sidak method (p < 0.001, n = 3)

Fig. 2

IPA Downstream Analysis as Networks identified functions on the immune response. The association between genes of our dataset #1 (300 genes) and specific downstream functions received an overlap p value and a z-score for prediction of activation or inhibition on the functions: “immune response of cells” and “homing of cells”. Twenty-seven genes from our dataset either upregulated or downregulated in HLF by IL3IgG eosinophil conditioned media, are involved in the regulation of “immune response of cells” and “homing of cells” with a significant z-score toward increased “activation”

Fig. 3

IPA Downstream Analysis as Networks identified functions on tissue remodeling. The association between genes of our dataset #1 (300 genes) and specific functions received an overlap p value and a z-score for prediction of activation or inhibition on the functions: “angiogenesis” and “development of vasculature” and “development of connective tissue cells”. Twenty-nine genes from our first dataset either upregulated or downregulated in HLF by the IL3IgG eosinophil conditioned media, are involved in the regulation of these three functions with a significant z-score toward increased “activation”

Fig. 4

IPA Downstream Analysis as Networks identified functions on lipid metabolism. The association between genes of our dataset #1 (300 genes) and specific functions received an overlap p value and a z-score for prediction of activation or inhibition on the function: “synthesis of lipid”. Nineteen genes from our first dataset either upregulated or downregulated in HLF by the IL3IgG eosinophil conditioned media, are involved in the regulation of this function with a significant z-score toward increased “activation”

Fig. 5

PCR analysis demonstrates upregulation of CXCL1, CXCL8, IL6 and ICAM1 expression levels in HLF activated with conditioned media from eosinophils pre-activated with IL3 and degranulating on aggregated IgG. Conditioned media were prepared from eosinophils pre-activated for 20 h with IL-3 or IL-5, and seeded on heat-aggregated human IgG for 6 h (IL3IgG or IL5IgG). Conditioned media from eosinophils pre-activated with IL-3 and seeded in uncoated wells (IL3) were also prepared. HLF were cultured for 24 h with the three types of conditioned media (IL3IgG, IL5IgG and IL3), a control medium (Ø) and medium including 1 ng/ml of recombinant human IL-3 plus 0.5 μg/ml of HA-IgG (rhIL3IgG). RT-qPCR were performed from total mRNA extracted from two HLF lines (L20 and L21) cultured in the five different conditions. For each HLF lines, controls (Ø) and rhIL3IgG are n = 2. IL3, IL3IgG and IL5IgG conditioned media were prepared from three different eosinophil donors, including two donors previously used for the RNAseq analyses. IL3, IL3IgG and IL5 were compared using One Way Anova (n = 3), and * indicates that IL3IgG is upregulated compared to IL3 and IL5IgG; and # indicates that IL5IgG is upregulated compared to IL3