Development of Potential Biomarkers for Foetal Brain Development After Congenital CMV Infection

Rationale for the study Current tools for prediction of perinatal outcome after fetal infection with CMV are very limited. Amniocentesis is usually offered from 20 weeks gestation to diagnose fetal infection. This sampling provides an opportunity to investigate novel approaches to predicting perinatal outcome.

This study aims to develop an mRNA based approach to studying the impact of CMV on the developing fetus. Dr Hui's PhD thesis was based on the study of amniotic fluid mRNA as a gene expression "summary fluid" of the fetus that provides meaningful information about development. This work suggested that information about fetal neurodevelopment is obtainable from amniotic fluid via cell-free fetal brain specific transcripts (mRNAs).

If a woman at risk of congenital CMV chooses to have an amniocentesis for diagnosis of fetal infection, this sampling provides an opportunity to collect an aliquot of AF for RNA analysis. RNA sequencing (RNAseq) is a relatively new technology that enables detailed analysis of the genes that are actively expressed ('switched on') during a particular disease state. The investigators will apply RNAsequencing methods to amniotic fluid to search for potential gene expression differences that may assist in understanding the disease through functional analysis and identifying candidate biomarkers for future studies.

Hypothesis That fetuses infected with CMV will have an altered gene expression profile compared with noninfected fetuses, as ascertained in amniotic fluid cell-free RNA.

Aims To perform comparative whole transcriptome analysis of AF RNA from CMV-infected and uninfected fetuses using RNA sequencing technology.

Methods This will be a multicentre study involving fetal medicine units in Melbourne (Mercy Hospital for Women), Sydney (Royal Hospital for Women, Randwick) and Leuven (UZ Leuven, Leuven). The investigators aim to recruit 20 women (10 cases, 10 controls) over 2 years.

Samples will be collected during clinically-indicated amniocentesis performed for the purposes of diagnosing the presence or absence of congenital infection. This is usually performed at 20-21 weeks gestation for women with seroconversion in early pregnancy, or at the time of fetal assessment in cases of referrals with structural abnormalities. An amniotic fluid (AF) volume of 5-10ml would be required for the research study. This sample volume does not pose a significant burden on the clinician performing the procedure (only 10-20 seconds additional collection time) and does not increase the risk of the procedure for the pregnancy or woman as there is NO additional needle insertion (no increase in miscarriage or impact on fetal or maternal well-being).

The research sample will be stored in a separate vial to the clinical sample at the respective recruitment sites. It is stored at 4 degrees Celsius and then centrifuged at 300 x g for 10 minutes within 6 hours of collection. The AF supernatant will be taken off and stored in a separate tube. Both the cellular portion and the supernatant will be frozen and stored at - 80°C. The Sydney recruitment site will transport the specimens to the University of Melbourne laboratory at the Mercy Hospital for Women on dry ice, where they will be processed. All samples will be given a study sample ID number before transport to the University of Melbourne research lab at the Mercy Hospital for Women to protect the participants privacy.

Total RNA from AF samples from 10 CMV infected and 10 uninfected fetuses, matched for gestational age and fetal sex, will be analysed. Total RNA will be extracted from the AF supernatant samples will be performed using the Qiagen Circulating Nucleic Acid extraction kit. Whole transcriptome amplification and RNA-sequencing of the cell free RNA will be performed and the expression profiles of the case and control groups compared in a paired analysis as performed in previous studies of AF RNA. Pathway analysis of the differentially-regulated genes will be used for the biological interpretation. The investigators will employ both widely used commercial software (Ingenuity Pathway Analysis™) and a fetus-specific functional analysis tool developed by collaborators at Tufts University (DFLAT).

The most down- and up-regulated brain transcripts will be identified as candidate biomarkers of abnormal outcome for future validation studies.

Pregnancy outcomes to six weeks postpartum for mother and baby will be collected from the hospital medical record, including fetal abnormalities detected on ultrasound or MRI, gestation and mode of birth, birth weight, use of prenatal therapies (CMV immunoglobulin or other), newborn investigations such as urine/saliva/ cord blood test results, cranial ultrasound, physical examination and results of newborn hearing screening.

Expected outcomes: Compared to uninfected fetuses, CMV-infected fetuses will show altered neurodevelopmental pathways and gene expression on functional analysis. Fetal-brain specific transcripts that are differentially expressed in CMV-infected fetuses will be identified as candidate biomarkers of neurodevelopmental outcome for future validation studies.

Data storage and security The database will be set up within RedCap (Research Electronic Data Capture), a free, secure, web-based application designed to support patient data capture for research studies. The University of Melbourne REDCap service provides secure storage and backup policies that comply with University standards and the the United States Health Information and Privacy Act. Only the study investigators will have access to this database.

Hard copies of consent forms and other study information will be kept in locked filing cabinets. Only the study investigators and the HREC will have access to the files.

6.0 Inclusion criteria Women undergoing clinically-indicated amniocentesis for suspected congenital CMV infection through the Department of Perinatal Medicine at the Mercy Hospital for Women or the Royal Hospital for Women, Randwick will be identified from the perinatal clinic referrals by Dr Hui or Dr Shand respectively. These women will include (i) women with evidence of maternal primary CMV infection during pregnancy (ii) Fetuses with structural abnormalities suggestive of congenital CMV infection (as listed in Hui L and Wood G. Perinatal outcome after maternal primary cytomegalovirus infection in first trimester: a practical update and counseling aid. 2015 Prenatal Diagnosis; 35:1-7. ) Any English-speaking woman aged 18 or over who is capable of giving informed consent will be eligible.

Subjects with confirmed fetal CMV infection will form the case group. Fetuses without evidence of congenital infection will form the control group. Structurally abnormal fetuses that are CMV negative will be excluded from the control group analysis. The rate of infected vs noninfected fetuses following maternal infection is approximately 40%, so we anticipate a reasonable ratio of cases and controls for our target sample size.

Initial contact will be made by the clinical team, but follow up by a research midwife to discuss the research study in detail will be made, either by telephone, or face-to-face following the clinical consultation wherever possible. This is to minimize potential conflicts from having the clinicians obtain patient consent for the study. Patients will be reassured that they will be under no obligation to participate and that their decision will not affect their clinical care or relationship with their treating medical practitioner Written, informed consent will be obtained from women prior to their procedure and be stored in a locked filing cabinet in the respective Departments of Perinatal Medicine.

Exclusion criteria Women who do not give consent, who are not capable of consent for medical procedures, non-English speaking, or who are under 18 years of age.

Statistical analyses Sample size In general, the more biological replicates (larger sample size), the stronger the inferences that can be made from the gene expression data. Biological variation between samples and the expected spread of differential gene expression are important factors influencing the ideal sample size. For this study, the investigators do not know in advance how the gene expression levels would be distributed, thus restricting our ability to use sample size algorithms for RNA-seq experiments. In addition to the restrictions on sample availability discussed above, the high costs associated with laboratory processing limited the feasibility of using large numbers of AF samples.

The main factors that influence formal sample size calculations for RNA-sequencing experiments are: depth of sequencing, coefficient of variation, magnitude of differential expression detected, false positive rate and power. Assuming average read depth of 20, equal within group coefficient of variation of 0.4, an effect size of 2 fold, 0.05 false positive rate and 80% power, the number of subjects per group required is 7.1 The investigators therefore aim for a minimum target of ten samples per group.

Analysis of differential gene expression and functional analysis A consultant bioinformatics specialist within the Translational Obstetric Group will be available for statistical and analysis support for the gene expression data. Genes that are significantly dysregulated in the CMV infected cases compared with non-infected controls will be identified. Functional analyses of the differentially expressed genes will be performed using Ingenuity Pathways Analysis (IPA) Version 9.0 software (Ingenuity; Redwood City, CA). Ingenuity is a manually-curated database that identifies over-represented biological processes in a given data set and calculates a significance score for each result using the right tailed Fisher's test. The investigators will also use a publicly available fetus-specific functional annotation "Developmental FunctionaL Annotation at Tufts" (DFLAT) that has previously been shown to enhance biological interpretation of perinatal datasets (http://dflat.cs.tufts.edu).2