Expression profiling of differentiating eosinophils in bone marrow cultures predicts functional links between microRNAs and their target mRNAs

Abstract

Background: MicroRNAs (miRNAs) are small non-coding RNAs that regulate complex transcriptional networks underpin immune responses. However, little is known about the specific miRNA networks that control differentiation of specific leukocyte subsets. In this study, we profiled miRNA expression during differentiation of eosinophils from bone marrow (BM) progenitors (bmEos), and correlated expression with potential mRNA targets involved in crucial regulatory functions. Profiling was performed on whole BM cultures to document the dynamic changes in miRNA expression in the BM microenvironment over the differentiation period. miRNA for network analysis were identified in BM cultures enriched in differentiating eosinophils, and chosen for their potential ability to target mRNA of factors that are known to play critical roles in eosinophil differentiation pathways or cell identify.

Methodology/principal findings: We identified 68 miRNAs with expression patterns that were up- or down- regulated 5-fold or more during bmEos differentiation. By employing TargetScan and MeSH databases, we identified 348 transcripts involved in 30 canonical pathways as potentially regulated by these miRNAs. Furthermore, by applying miRanda and Ingenuity Pathways Analysis (IPA), we identified 13 specific miRNAs that are temporally associated with the expression of IL-5Rα and CCR3 and 14 miRNAs associated with the transcription factors GATA-1/2, PU.1 and C/EBPε. We have also identified 17 miRNAs that may regulate the expression of TLRs 4 and 13 during eosinophil differentiation, although we could identify no miRNAs targeting the prominent secretory effector, eosinophil major basic protein.

Conclusions/significance: This is the first study to map changes in miRNA expression in whole BM cultures during the differentiation of eosinophils, and to predict functional links between miRNAs and their target mRNAs for the regulation of eosinophilopoiesis. Our findings provide an important resource that will promote the platform for further understanding of the role of these non-coding RNAs in the regulation of eosinophil differentiation and function.

Figure 1. Expansion of bmEos ex vivo.

Bone marrow cells from BALB/c mice were cultured for 14 days, see Methods and samples taken from day 4 to day 14 in cultures grown in the presence of IL-5. Eosinophils were identified by (A) flow cytometry (SiglecF⁺Gr-1⁻CD11b⁺CD11c⁻), (B) light microscopy with Giemsa staining (100×) and (C) the percentages and numbers of bmEos were determined using a haemocytometer. Values are presented as mean ±SEM (n = 4∼6), **P<0.01 (d10, d12 or d14 vs. other groups). *P<0.05 (vs. d4, d6 or d8). # P<0.05 (vs. d4 or d6).

Figure 2. Expression levels of miRNA during bmEos differentiation.

Bone marrow cells from BALB/c mice were cultured for 14 days. Heat map representation of expression levels of miRNA that were up-regulated or down-regulated by more than 5-fold from day 4 to day14 in the presence of IL-5. The fluorescence index of each miRNA at different timepoints was further normalized to that of the respective miRNAs in the control group (isolated bone marrow cells). The normalized microarray data were managed and analyzed by GeneSpring (Agilent). Data represent three independent bmEos cultures. Scale ranges from a signal value of −6.1(blue) to 6.1(red). Some of miRNAs were selected to further confirm the efficacy of miRNA array.

Figure 3. Confirmation of miRNA array expression by Taqman quantitative PCR.

8 miRNAs (miRNA -144, -223, -146a, -365, -155, -451, -194 and -494) were selected to verify the expression profile of the miRNA array. Samples were RNA isolated from cultured bone marrow cells from day 4 to day14 in the presence of IL-5, and represent three independent cultures. Values are presented as mean ±SEM (n = 4∼6), **P<0.05 (vs. other groups). *P<0.05 (vs. d4).

Figure 4. Expression of GATA -1/2, PU.1 and c/EBPε correlated with the expression levels of miRNAs that potentially target these transcripts.

Bone marrow cells were cultured as described in Methods and RNA samples were extracted from cells taken from day 4 to day14 which had been grown in the presence of IL-5. A. Expression levels of GATA -1/2, PU.1 and c/EBPε were determined by qPCR. B. Potential miRNAs targeting the 3′-UTR of the four transcription factors were identified by TargetScan, the Miranda database and IPA ingenuity system. Blue represents decreased expression of miRNAs, whereas yellow is for increased expressed miRNAs. C. The fold changes of potential regulating miRNAs were calculated based on the fluorescence index of each miRNA at different time-points, after normalization to that of the respective miRNAs in the control group (isolated bone marrow cells). Data represent three independent eosinophil cell cultures. Values are presented as mean ±SEM (n = 4∼6), **P<0.05 (vs. other groups). *P<0.05 (vs. d4).

Figure 5. Potential molecules and top canonical pathways that were predicted and targeted by the miRNAs that exhibited 5-fold changes in expression.

A. Target prediction by TargetScan database (http://www.targetscan.org/) was established on sequence data complementarity to target 3′UTR sites. Target molecules, associated with eosinophil biology, were identified by exact syntax matching in the MeSH database. (http://www.nlm.nih.gov/MeSH/MeSHhome.html). B. Top 30 canonical pathways that consist of the putatively selected 348 molecules as identified by IPA. The significance of association between selected genes and canonical pathway was evaluated by a right-tailed Fisher’s exact test to calculate a p value determining the probability that the association is not explained by chance alone (grey bars, upper y-axis). Ratios referring to the proportion of selected genes from a pathway related to the total number of molecules that make up that particular pathway were also displayed (line graph, bottom y-axis).

Figure 6. Expression levels of IL-5Rα, CCR3 and MBP correlated with the expression of miRNAs that potentially target these transcripts.

Bone marrow cells were cultured as described in Methods and RNA samples were extracted from day 4 to day14 from cells grown in the presence of IL-5. A. Expression levels of IL-5Rα, CCR3 and MBP were determined by qPCR. B. Potential miRNAs targeting the 3′-UTR of IL-5Rα and CCR3 were identified by TargetScan and MiRanda database and IPA ingenuity system. Blue represents decreased expression of miRNAs, whereas yellow is for miRNAs with increased expression. C. The fold changes of potential regulating miRNAs were calculated based on the fluorescence index of each miRNA at different time-points, after normalization to that of the respective miRNAs in the control group (isolated bone marrow cells). Data represent three independent eosinophil cell cultures. Values are presented as mean ±SEM (n = 4∼6), **P<0.001 (vs. other groups). *P<0.01 (vs. BM).

Figure 7. Expression of TLR4, TLR6 and TLR13 correlated with the expression of miRNAs that potentially target these transcripts.

Bone marrow cells were cultured as described in Methods and RNA samples were extracted from day 4 to day14 from cells grown in the presence of IL-5. A. Expression levels of TLR4, TLR6 and TLR13 were determined by qPCR. B. Potential miRNAs targeting the 3′-UTR of TLR4, TLR6 and TLR13 were identified by TargetScan, the MiRanda database and IPA ingenuity system. Blue represents decreased expression of miRNAs, whereas yellow is for miRNAs with increased expression. C. The fold changes of potential regulating miRNAs were calculated based on the fluorescence index of each miRNA at different time-points, after normalization to that of the respective miRNAs in the control group (isolated bone marrow cells). Data represent three independent eosinophil cell cultures. Values are presented as mean ±SEM (n = 4∼6), **P<0.001 (vs. d4, d6, d8 or d10).