4.6. Adipose Tissue Microarray Analyses

RNA was isolated from frozen eWAT of Adcy5^–/– (n = 10, 5 females, 5 males) and control (n = 10, 5 females, 5 males) CD mice. Before microarray analysis, RNA integrity and concentration were examined on an Agilent Fragment Analyzer (Agilent Technologies, Palo Alto, CA, USA) using the HS RNA Kit (Agilent Technologies) according to the manufacturer’s instructions. cRNA was prepared from 100 ng of total RNA hybridized to GeneChip Clariom S arrays (ThermoFisher Scientific). Arrays were scanned with a third generation Affymetrix GeneChip Scanner 3000 (ThermoFisher Scientific). Affymetrix GeneChip data were extracted from fluorescence intensities and were scaled to normalize data for inter-array comparison using Transcriptome Analysis Console (TAC) 4.0.2 software.

Raw data were preprocessed applying the oligo Bioconductor R package (v1.50.0, [51]), which performs a deconvolution method for background correction, quantile normalization and uses the Robust Multichip Average (RMA) algorithm for summarization [52]. Quality control of the raw and normalized data was performed using the Biobase (v2.46, [53]) and oligo R Bioconductor packages [54]. Based on the outlier detection test of the arrayQualityMetrics (v3.42, [54]) Bioconductor R package, two male Ctrl and one female Adcy5^–/– samples were excluded. The differentially expressed genes (DEGs) for the comparisons (i) female Adcy5^–/– vs. female Ctrl and (ii) male Adcy5^–/– vs. male Ctrl were screened using the Linear Models for Microarray data (LIMMA) method (v3.42, [55]). To increase the signal-to-noise ratios array weights were considered. Since only a few genes (36 female; 1 male, see Table S2) survived multiple testing (FDR < 0.05) applying the Benjamini–Hochberg procedure [56], we set the threshold for identification of DEGs as p-value < 0.01 and |FC| ≥ 2. This allowed us to be exploratory about the results, especially in the pathway analysis. A broad gene list functional enrichment analysis for KEGG mouse pathways was performed using Enrichr (www.amppharm.mssm.edu/Enrichr accessed on 19th April 2021) [57,58]. All differentially expressed genes (p-value < 0.01) irrespective of FC were used. Microarray data were deposited in the ArrayExpress database at EMBL-EBI (www.ebi.ac.uk/arrayexpress accessed on 19th April 2021) under accession number E-MTAB-9417.