Transcriptome meta-analysis of mouse lenses deficient for key genes to uncover regulatory pathways in lens development and pathology

Abstract

The vertebrate lens is a powerful model to study tissue-specific gene regulation, development, and disease. Recent advances in high-throughput transcriptomics have yielded a growing number of high-throughput RNA-sequencing (RNA-seq) datasets from conditional knockout (cKO) mouse models carrying deletion/mutations of specific genes essential for lens biology and perturbations of which are linked to cataract in humans or animal models. However, these studies are often analyzed in isolation, limiting the discovery of shared regulatory programs and broader biological insights. A systematic meta-analysis of these RNA-seq datasets has not been performed, representing a key knowledge-gap. Therefore, I performed a comprehensive meta-analysis of publicly available lens KO RNA-seq datasets – which spanned embryonic, early postnatal, and adult stages – to identify core transcriptional programs and misregulated pathways relevant to lens biology. Raw FASTQ files were downloaded from the NCBI Gene Expression Omnibus (GEO) using SRR accession numbers and processed through a elaborate pipeline. Batch normalization was applied to account for variation due to laboratory origin and differential expression analysis was conducted using DESeq2 and edgeR, with meta-analysis performed separately for embryonic, early postnatal, and adult datasets. Gene Ontology (GO) enrichment was assessed using enrichGO (clusterProfiler) and Database for Annotation, Visualization and Integrated Discovery (DAVID), and lens-specific expression and enrichment were evaluated using the web-based eye bioinformatics tool, iSyTE (integrated Systems Tool for Eye gene discovery). This analysis led, for the first time, to the identification of a cohort of genes that are commonly mis-expressed in the lens, regardless of the specific gene perturbation. Thus, this analysis identified a “core” set of genes whose mis-regulation correlates with lens pathology. In embryonic datasets, differentially expressed genes (DEGs) were highly enriched for processes central to lens morphogenesis, such as lens fiber cell differentiation, and visual system development. Many downregulated genes, including Crygf, Crygb, Dnase2b, and Bfsp1 – identified in the present analysis – are known for their critical role in facilitating lens transparency, serving as an independent validation of the meta-analysis having worked well for the identification of key genes in the lens. Besides the established, novel genes whose function is not yet characterized in the lens were also identified (Tmprss11e, Ermap, Uox). In early postnatal datasets, a smaller set of DEGs was identified, including ectopically upregulated retinal genes (Prph2, Pdc, Pde6b). In adult lenses, a novel gene Ces5a was identified as significantly downregulated in the knockout samples. In wild-type (WT) lenses, Ces5a shows moderate expression during development and early postnatal stages, but its expression increases sharply—by over 17-fold—by postnatal day 56 (P56), where it exhibits a 26-fold enrichment compared to whole-body expression levels. This work provides the first stage-stratified meta-analysis of lens cKO RNA-seq data, uncovering shared and unique transcriptional consequences of gene disruption in lens development. Importantly, it identifies a “core” set of genes whose misexpression correlates with lens pathology, regardless of the genetic changes that caused it, thus potentially identifying new targets that may be commonly targeted in therapeutic approaches.