Maternal symptoms of preeclampsia (PE) are a primary consequence of excessive levels of sFlt1 protein in the maternal blood supply 1. sFlt1 is encoded by the gene Fms-related tyrosine kinase 1 (FLT1), which produces multiple protein isoforms: (1) transmembrane bound Flt1 (mFlt1), a tyrosine kinase receptor participating in vascular endothelial growth factor (VEGF) signaling; and (2) truncated and blood circulating (soluble) forms, sFlt1s, which lack the transmembrane anchoring domain and act as antagonists to mFlt1. Relative production of these isoforms results from polyadenylation sites within exon 30 (mFlt1) or within introns 13 and 14 (sFlt1s). The goals of this project are three-fold: (1) Determine whether increased sFlt1 production in PE results from an overall increase in FLT1 gene expression, from preferential use of intronic polyadenylation sites or both; (2) assess general gene expression changes in PE versus non-PE placentae; (3) determine if there is a global change in APA (alternative polyadenylation) site usage in PE. To address these goals we used a next-generation sequencing specifically targeting the 3′-ends of mRNAs: PolyAdenylation Site Sequencing (PAS-Seq). PAS-Seq allows for quantitative gene expression measurements as well as transcriptome-wide assessment of APA 2,3. Using total RNA extracted from the villous tissue of 6 normal (N) and 11 PE patient placentae, we generated and sequenced 17 PAS-Seq libraries. Among the PE patients, 8 were early onset (eoPE = delivery <34 weeks). By combining all counts from individual mRNA isoforms into a single measure (reads per million, RPM) for each gene and performing differential expression analysis (via DESeq2), we identified 288 up-regulated genes and 170 down-regulated genes in PE compared to N (padj ⩽0.05). Of these 458 differentially expressed genes, 200 (44%) overlap with the 1295 differentially expressed genes previously identified for severe PE upon meta-analysis of microarray data 4. However, only 15 (3%) of our set overlap with the 297 differentially expressed genes recently determined by RNA-Seq 5.Given the extreme heterogeneity of this disease [i.e., eoPE vs loPE; intrauterine growth restriction (IUGR) PE versus non-IUGR-PE] such a high degree of variability is perhaps unsurprising. Nonetheless, FLT1 was among the small set of differentially expressed genes common to all three datasets. In our PAS-Seq data, FLT1 was the most up-regulated gene (2.8-fold in all PE samples vs. N; 4.2-fold in eoPE vs N). Among all detectable APA isoforms, the four most abundant were: sFlt1-e15a > sFlt1-i13 long > sFlt1-i13 short > m-Flt1. All four isoforms were significantly higher (p ⩽0.05; t-test) in PE versus N; this suggests an overall increase in Flt1 gene transcription in PE. Along with this increased transcription, there also appears to be a switch toward more promoter-proximal sites. We are currently working to determine whether this switch to more promoter-proximal APA sites is a general phenomenon in PE placentae, or is solely a feature of the FLT1 gene. Our PAS-Seq data indicate that the FLT1 gene is subject to both transcriptional upregulation and a shift toward promoter-proximal polyadenylation in PE. Further data analysis regarding global APA switches and pathway analysis will be presented at the conference. R.J. Levine, et al., N. Engl. J. Med. 350, (2004) 672–683. P.J. Shepard, et al., RNA 17 (2011), 761–772. A. Derti et al., Gen. Res. 22 (2012), 1173–1183. K. Leavey, et al., PLoS One 10 (2015), e0116508. S. Sõber, et al., Sci. Rep. (2015), 1–17.
Read full abstract