RNA Analysis

The RNA studies included in this thesis investigate both coding and non-coding RNAs. These studies relied on RNA isolated from the same mice and were used for hybridization on Affymetrix GeneChip® Mouse Gene 1.0 ST arrays and Affymetrix GeneChip® miRNA 2.0 arrays respectively. The RNA expression analysis primarily examined the CPD PAE paradigm (3 arrays with samples pooled in duplicate); however, the binge injection paradigms were also examined (2 arrays with samples pooled in triplicate).

2.3.2.1

Gene Expression Arrays

Affymetrix gene expression arrays are single channel arrays that have proved useful for the study of ethanol exposure in mice (Kerns & Miles 2008). The Affymetrix GeneChip® Mouse Gene 1.0 ST array (Affymetrix, Santa Clara, CA) consists of 770,317 unique 25-mer perfect match probes that are targeted across 28,853 transcripts, with 1,305 of those being ncRNA, and an average of 27 probes for a gene.

The Affymetrix GeneChip® Mouse Gene 1.0 ST array was used to analyze PAE induced alterations to gene expression. Single-stranded complementary DNA (sscDNA) was prepared from 200 ng of total RNA as per the Ambion WT Expression Kit for Affymetrix GeneChip Whole Transcript WT Expression Arrays (Applied Biosystems, Carlsbad, CA) and the Affymetrix GeneChip WT Terminal Labeling kit and Hybridization User Manual (Affymetrix, Santa Clara, CA). Total RNA was first converted to complementary DNA (cDNA), followed by in vitro transcription to make cRNA. By subjecting the cDNA to a cRNA conversion, a biotinylated Uracil analogue is incorporated. The biotin molecules can then be used to attach fluorescent molecules to label the cRNA. Subsequently, 5.5 µg of sscDNA was synthesized, end-labeled, and hybridized for 16 hours at 45°C to Mouse Gene 1.0 ST arrays. GeneChip Fluidics Station 450 performed all liquid handling steps, and GeneChips were scanned with the GeneChip Scanner 3000 7G using the Command Console v1.1 (Affymetrix, Santa Clara, CA). All hybridizations were performed at the London Regional Genomics Centre at the Robarts Research Institute, University of Western Ontario. Probe level (.CEL file) data were

generated using the Affymetrix Command Console v1.1. The .CEL files identify intensities of individual probes.

Probes, which have multiple replicates and target exons, were assembled into probesets that represent single genes. This summarization of the .CEL files was done in Partek Genomics Suite v6.5 (Partek Inc., St Louis, MO) by quantile normalization using the RMA algorithm adjusted for GC content and log2-transformed (Irizarry et al. 2003). Partek software was then used to determine gene level one-way ANOVA p-values and fold changes. All arrays were analyzed by using a 1.2-fold cut-off with a significance threshold of p=0.05 to generate a list of genes of interest to be assessed further using bioinformatic analysis of gene ontologies and pathways. The gene expression array results for the T1, T2, T3, and CPD paradigms were deposited within the NCBI GEO database under accessions GSE34469 (T1 & T2), GSE34549 (T3) and GSE34305 (CPD, experiment 2). The data from CPD mice were generated by Ben Laufer, T1 & T2 mice by Katarzyna Mantha, and T3 mice by Morgan Kleiber.

2.3.2.2

miRNA Expression Arrays

The Affymetrix GeneChip® miRNA 2.0 array is based on a similar design as the Affymetrix GeneChip® Mouse Gene 1.0 ST array, and has probes that are 25-mer. However, smaller probes are used if the miRNA is less than 25 bases. The miRNA array targets 722 mouse mature miRNAs, 690 mouse pre-miRNA, and 2,334 small RNAs from humans that include snoRNAs. It also has 9 probes per a probe set for a miRNA. The workflow is also similar to Affymetrix GeneChip® Mouse Gene 1.0 ST arrays.

In order to investigate the effect of PAE on miRNAs, all samples were analyzed at the London Regional Genomics Centre (Robarts Research Institute, London, Ontario) using Affymetrix miRNA 2.0 arrays. Briefly, 1 µg total RNA from each treatment paradigm was labeled using the Flash Tag Biotin HSR kit (Genisphere, Hatfield, PA) and hybridized to Affymetrix miRNA 2.0 arrays for 16 hours at 45°C. Probe level (.CEL file) data were generated using Affymetrix Command Console v1.1. Probes were summarized to gene level data in Partek Genomics Suite v6.6 (Partek Inc., St Louis, MO) by using the RMA algorithm (Irizarry et al. 2003). Partek software was used to determine differences between control and ethanol-treated samples by a one-way ANOVA and corresponding

P-values and fold changes. For each treatment model, the miRNAs present on this array were filtered using stringency criteria of 1.2-fold change (p<0.05) and subjected to a hierarchical clustering analysis by using Euclidean distance and average linkage to assess consistency in ethanol response between the arrays of different treatment paradigms. The miRNA expression array results for all treatment protocols were deposited within the NCBI Gene Expression Omnibus (GEO) database under accession GSE34413. Analyses of CPD mice were performed by Ben Laufer, T1 & T2 mice by Katarzyna Mantha and Ben Laufer, and T3 mice by Morgan Kleiber and Ben Laufer.

2.3.2.3

Quantitative PCR Confirmation

The quantitative polymerase chain reaction (qPCR) allows for real-time quantification of DNA amplification by optically measuring fluorescent dyes after laser excitation (Fraga et al. 2008). Taqman® is a methodology that utilizes sequence-specific oligonucleotide probes attached to both a fluorophore and a quencher molecule, which prevents fluorescence. During qPCR, the probe will hybridize to complementary DNA and will then be degraded by Taq polymerase, which releases the fluorophore from the quencher and allows it to be detected after excitation. qPCR technology can also be coupled with reverse transcription (RT) to enable the study of both coding and non- coding RNA expression.

In order to confirm gene expression results, cDNA was reverse transcribed from whole-brain RNA of the CPD paradigm (n=6) and matched controls (n=6). For mRNA expression, cDNA was created using Applied Biosystems High-Capacity cDNA Reverse Transcription Kit (Foster City, CA) and sequence specific TaqMan™ assays (Foster City, CA), which were used according to the manufacturer’s protocol. For miRNA expression, the Applied Biosystems TaqMan™ MicroRNA Reverse Transcription Kit (Foster City, CA), which uses stem-loop reverse transcription primers (Chen et al. 2005), and sequence specific TaqMan™ MicroRNA Assays (Foster City, CA) were used according to the manufacturer’s protocol. All TaqMan™ probes were selected using the Applied Biosystems (Carlsbad, CA) search engine to identify previously characterized TaqMan™ assays. GAPDH was chosen as an endogenous control for mRNA expression, while snoRNA 202 was chosen as an endogenous control for miRNA expression (Gao et al.

2010). The target and control reactions were run in separate tubes on the same plate for each sample as per the manufacturer’s protocol. Three technical replicates were averaged for both the endogenous control and gene of interest for each sample. qPCR reactions were performed on the Applied Biosystems StepOne™ Real-Time PCR System 2.0 according to the manufacturer’s protocol. Fold change was calculated using the ΔΔCt method (Schmittgen & Livak 2008). The ΔΔCT method allows for relative quantification of gene expression from qPCR experimentation. CT is the threshold cycle, which is the PCR cycle where fluorescence exceeds a threshold value. The ΔΔCT method compares CT’s for an experimental gene to a reference gene across both the experimental and control group. The results were analyzed using Applied Biosystems DataAssist™ Software v3.0. Statistical significance was assessed by an unpaired Student’s t-test.

2.3.2.4

Bioinformatic Analysis of Differentially Expressed Probe

Sets (mRNAs and miRNAs)

The chromosomal locations for all miRNAs were determined by using miRBase and Ensembl (Flicek et al. 2011; Kozomara & Griffiths-Jones 2011). Genomically imprinted regions were identified by the mouse imprinting catalog (www.mousebook.org).

Data from the CPD paradigm were analyzed through the use of Ingenuity® microRNA target filter [Ingenuity® Systems (www.ingenuity.com)] to generate lists of interactions between genes and miRNAs of interest from the TargetScan Human database (mouse was not available at the time). Results were filtered based on a moderate or high confidence of interaction (Lewis et al. 2005; Vergoulis et al. 2012), brain specificity, and an inverse miRNA to target mRNA expression relationship. miRNAs were filtered based on a 1.15 fold change and a p<0.3 cut-off. This analysis is referred to as the original (2012) analysis.

At a later date, updated (2016) annotations and improved software became available for the gene expression and miRNA arrays. These updates offered some improvement and thus were utilized for a re-analysis. Enrichr was used for biological, cellular, and molecular ontologies (Chen et al. 2013). Partek Pathways was used for canonical KEGG pathways. Gene annotations were obtained from UniProt (UniProt-

Consortium 2014). An updated and more stringent (1.2 fold change, p<0.05) miRNA target filter analysis was done in Partek using Mouse TargetScan 6.2. This analysis is referred to as the updated (2016) analysis.