Maternal blood EBF1-based microRNA transcripts as biomarkers for detecting risk of spontaneous preterm birth: a nested case-control study

Abstract

Objective

Both genetic variants and maternal blood mRNA levels of EBF1 gene have been linked to sPTB. Animal and human studies suggest that specific EBF1-based miRNAs are involved in various physiological and pathophysiological processes. However, to date, we did not find any reports of EBF1-based miRNAs or miRNA transcripts in relation to sPTB. We therefore aimed to examine whether maternal blood early B cell factor 1 (EBF1) gene-based microRNA (miRNA) transcripts can be used for detecting risk of spontaneous preterm birth (sPTB).

Methods

We conducted a nested case-control study within a Canadian cohort consisting of 1878 singleton pregnancies enrolled from May 2008 to December 2010 in Calgary, Alberta, Canada. We used a public gene expression dataset (GSE59491) derived from maternal blood in trimesters 2–3 that included women with sPTB (n = 51) and term births (n = 106) matched for maternal age, race/ethnicity, pre-pregnancy body mass index, smoking during pregnancy, and parity within the Canadian cohort. Two bioinformatics tools, miRWalk and STarMirDB, with different algorithms were applied to retrieve miRNA transcripts that putatively target the EBF1 gene (i.e. EBF1-based). Limma moderated t-tests were used to examine differentially expressed (DE) miRNA transcripts (sPTB vs term) within trimesters. Logistic regression models with miRNA transcript tertiles were applied to assess threshold associations between candidate miRNA transcripts’ levels and sPTB. Receiver operating characteristic (ROC) analyses were used to identify the maximum Youden Index and its corresponding optimal sensitivity/specificity cut-point of EBF1-based miRNA transcripts for classifying sPTB, and to compare the classification performance of a linear combination (score) of miRNA transcripts with that of individual miRNA transcripts. A five-fold cross-validation was applied to examine the possible overfitting problem of the final ROC model.

Results

Four maternal blood EBF1-based miRNA transcripts (MIR4266, MIR1251, MIR601, MIR3612) in the 3rd trimester were significantly associated with sPTB. The odds ratios (95%CIs) for highest versus lowest tertile of the four miRNA transcripts were 3.01–5.25(1.21–13.14, p ≤ .018). The combined 4-miRNA transcripts’ score significantly improved the classification of sPTB compared to individual miRNA transcripts (AUC increased from 0.65–0.69 to 0.82, p ≤ .0034) and showed a sensitivity for sPTB of 0.81 and a specificity of 0.72. The final ROC model of the EBF1-based 4 miRNA transcripts’ score in cases and controls had no significant overfitting issue.

Conclusions

Maternal blood EBF1-based miRNA transcripts may, along with other biomarkers, be useful in screening for sPTB risk in 3rd trimester. Our results also provide clues for further study of potential molecular mechanisms underlying the relationship between EBF1 gene and sPTB, e.g. connecting genetic variants, mRNA expression, and miRNA regulation.

Keywords:

• EBF1 gene
• miRNA transcript
• miRNA transcripts’ score
• spontaneous preterm birth

Acknowledgements

We are thankful to Dr. Ana Vazquez for her suggestions to the manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Contribution to authorship

G.Z. developed research ideas, conducted data analyses including statistical analysis and bioinformatics analysis, interpreted the data, and drafted manuscript. C.H. provided advice on developing research ideas and conducting data analyses, interpreted the data, as well as reviewed and revised manuscript. Y.J.H. prepared some data as well as reviewed and revised manuscript. M.K. assisted with some data preparations and manuscript review/revision. S.J.L. assisted with manuscript review/revision. All authors gave final approval of the version to be published.

Details of ethics approval

Not applicable (This study was a bioinformatics analysis of a publicly available gene expression dataset and did not involve any human or animal subjects, or medical records).