Presented at the Neonatal Society 2016 Autumn Meeting.
Montaldo P1, Hervás Marín D2, Calabria I2, Pedrola L2, Cernada M2, Lally PJ1, Oliveira V1, Chaban B1, Atreja G1, Banerjee J1, Soe A3, Pattnayak S3, Harizaj H3, Shankaran S4, Kaforou M5, Herberg J5, Vento M2, Thayyil S1 on the behalf of the MARBLE consortium
1 Centre of Perinatal Neuroscience, Imperial College London, UK
2 Health Research Institute La Fe, Valencia, Spain
3 Medway NHS Foundation Trust, Kent, UK
4 Neonatal-Perinatal Division, Wayne State University, USA
5 Infectious Diseases, Imperial College London, UK
Background: Variable responses to hypothermic neuroprotection may be related to the clinical heterogeneity of encephalopathic babies, and hence, a better disease stratification may be help in development of individualised neuroprotective therapies. Recently genome-wide analysis of gene expression has been effectively exploited for disease stratification and diagnostics in a variety of paediatric infections (1). We examined if whole transcriptome analysis can identify specific transcriptome profiling in perinatal hypoxic ischemic encephalopathy (HIE).
Methods: We extracted RNA from 0.5 ml blood (PreAnalytiX BD/QIAgen) from seven full term infants who had therapeutic hypothermia for neonatal encephalopathy and five healthy controls babies, recruited into the MARBLE (Magnetic Resonance Biomarkers in Neonatal Encephalopathy) study, after informed parental consent (NIHR; Ethics: 11/H0717/) (2). We performed whole blood RNA sequencing using Ion Proton™ Sequencer. Reads were aligned to the hg19 human genome and after pre-processing the expression abundance for each gene was acquired. We used principal component analysis (PCA) to determine the significant sources of variability in the data sets. We fitted quasi-likelihood negative binomial generalized log-linear models to the count data to identify significantly differentially expressed genes, which were then subjected to pathway analysis using the Database for Annotation, Visualization and Integrated Discovery (DAVID) method. We then used a variable selection method (logistic regression model with elastic net penalization) to identify genes predictive of HIE.
Results: Clinical characteristics of the sequenced babies are given in the Table.
PCA showed separate clusters in HIE (green) and healthy controls (blue) (Figure 1). Unsupervised hierarchical clustering based on the 710 genes identified by elastic net (rows) completely discriminated between the healthy controls (pink line) and the HIE cases (purple line). Up-regulated genes are represented in green and down-regulated genes in red (Figure 2).
The most significant pathways identified were related to endocytosis (p value 2.94 E-4) and ubiquitin mediated proteolysis (p value 3.6E-4).
Conclusion: These preliminary data suggest that babies with HIE may have different gene expression profiles compared with healthy control babies. Genome wide expression is a promising tool to identify perinatal hypoxic injury and to provide insights into the underlying molecular and cellular mechanisms, and may assist in disease stratification.
Corresponding author: firstname.lastname@example.org
Herberg J et al., Diagnostic Test Accuracy of a 2-Transcript Host RNA Signature for Discriminating Bacterial vs Viral Infection in Febrile Children. JAMA. 2016;316(8):835-45
Lally PJ et al. Magnetic Resonance Biomarkers in Neonatal Encephalopathy (MARBLE): a prospective multi-country study. BMJ open 2015;5:e008912