miRNA signatures can predict acute liver failure in hepatitis E infected pregnant females

Abstract

Background: Acute viral hepatitis E (AVH-E) can often result in acute liver failure (ALF) during pregnancy. microRNAs serve as mediators in drug induced liver failure. We investigated their role as a biomarker in predicting ALF due to HEV (ALF-E).

Methods: We performed next generation sequencing and subsequent validation studies in PBMCs of pregnant (P) self limiting AVH-E, ALF due to HEV (ALF-E) and compared with AVH-E in non-pregnant (NP) females and healthy controls.

Findings: Eleven microRNAs were significantly expressed in response to HEV infection; importantly, miR- 431, 654, 1468 and 4435, were distinctly expressed in pregnant self-limiting AVH-E and healthy females (p = 0.0005), but not in ALF-E. Sixteen exclusive microRNAs differentiated ALF-E from self limiting AVH-E in pregnant females. miR-450b which affects cellular proliferation and metabolic processes through RNF20 and SECB was predominanlty upregulated and correlated with poor outcome (ROC 0.958, p = 0.001).

Interpretation: Our results reveal that a specific microRNA profile can predict fatality in ALF-E in pregnancy. These microRNAs could be exploited as prognostic biomarkers and help in the development of new therapeutic interventions.

Keywords: Health sciences; Virology.

Fig. 1

(A) Global analysis of small RNAs identified using deep sequencing. (B-C) Frequency of Immune cells in peripheral blood of pregnant and non-pregnant healthy controls(HC) and acute HEV patients. (D) Averaged expression signature of miRs in pooled PBMCs isolated from pregnant AVH-E, ALF-E and HC and non pregnant AVH-E. Pregnant AVH-E and non pregnant AVH-E was compared and also pregnant ALF-E and AVH-E was compared with HC. Each lane represents the average signal log intensity of five patients in each group as independent technical replicates. Expression was displayed in red and green for increased and decreased expression levels and black for no changed expression levels. Color intensity was calibrated to expression level as illustrated at the side of heat map. (E) miR expression data from different groups were applied to principal component analysis. The use of this technical analysis allowed us to visualise the observed variance between groups. (F) Averaged gene expression signature showing decreased global gene expression in AVH-E(P) compared to AVH-E(NP). (G) Heat maps of gene expression in pregnant ALF-E, AVH-E and HC. (H) Principal component analysis (PCA) visualise observed variance in groups.

Fig. 2

(A-B) Scatterplot representation of comparative global miR and gene expression profiles between AVH-E(P), ALF-E(P) and HC(P) patient groups. X and Y axis shows expression levels of miRs and genes between two groups. R2 value, a measure of goodness-of-fit of linear regression indicates the degree of correlation between the subjects.

Fig. 3

(A) Distinct miR expression in pregnant HEV infection compared to HC. (B) Difference in fold change (log₂ ratio) expression of various miRs in three groups. (C) No difference in miR profiling in AVH-E(P) and AVH-E(NP) (D) Fold change expression of miR-590 and miR-624 of various subjects. Data are expressed as box plots in which the horizontal lines indicate the 25th, 50th, and 75th percentiles of the fold change measured by qRT-PCR.

Fig. 4

(A) miRs expressed as indicators of virus self clearance. Hierarchical clustering showing high expression of specific miRs in AVH-E(P) and decrease in ALF-E (P) patients. (B) Fold change (log₂ ratio) expression of specific miRs. (C) miR profiling in AVH-E(P) and AVH-E(NP) does not show any difference. (D) Box plots showed fold change expression of miR-431 and 1468 by qRT-PCR.

Fig. 5

(A) Heat map showed comprehensive miR expression linked with ALF related to HEV infection as indicators of death signatures. (B) Fold change (log₂ ratio) expression of specific miRs. (C) Box plots showing qRT-PCR data of higher fold change expression of miR-450a, 450b and miR-365a in ALF-E(P). (D-E) ROC curve analysis of miR-450b between ALF-E (P) and AVH-E (P) showed a cut off of ΔC_t ≤ 19.5 (Sensitivity = 87.5, specificity = 87.3; p = 0.001) in ALF in pregnancy. (F) Higher expression of miR-450b in ALF non-survivors than ALF patients who survived.

Fig. 6

Differential gene expression data represented by bar graphs showes various number of genes related to both innate and adaptive immune cells as well as Type 1 and Type 2 immune signalling pathways expressed in self limiting AVH-E(P) and ALF-E(P) groups.

Fig. 7

Network analyses for pathways associated with miRs and target genes. Predicted targets of differentially expressed miRs was determined by using mirtarBase data. To determine the possible miR − mRNA interactions, the list of targets was generated using Integromatrix Software (Proprietary of Bionivid Technology Pvt Ltd, Bangalore, India). Red lines indicated miRs inhibiting the expression of target genes, and regulating different processes. (A) miRs and targeted genes related to cell death and DNA metabolic processes in HEV infection. (B-D) mRNA expression levels of targeted genes by qRT-PCR. (E) miRs and genes regulated homeostasis processes in patients who self cleared the virus. (F-H) Bar graphs showed mRNA expression of specific genes by qRT-PCR.

Fig. 8

(A) miR signatures and their targeted genes in pregnant patients with acute liver failure. (B-C) Bar Diagram showing specific gene expression by qRT-PCR.

Fig. 9

miR-450b interaction with immune cells showing involvement of various pathways and genes.