81 JAK-STAT SIGNALLING IS CRITICAL FOR INNER CELL MASS DEVELOPMENT IN BOVINE BLASTOCYSTS

Abstract

The inner cell mass (ICM) of mammalian blastocysts comprises 2 transient lineages, namely hypoblast and epiblast, which develop into extra-embryonic and embryonic tissues, respectively. In the mouse, epiblast cells autocrinally secrete fibroblast growth factor (FGF) to induce hypoblast differentiation, and pharmacological FGF/mitogen-activated protein kinase (MAPK) signal inhibition converts all ICM cells into epiblast. We conducted a chemical screen for additional signal enhancers of epiblast identity in bovine Day 8 blastocysts. From the morula stage onwards, in vitro-fertilised (IVF) embryos were cultured in the presence of 9 small molecule inhibitors, targeting 9 principal signal pathway components. Inhibitors included SB431542, LDN193189, BIBF1120, Forskolin, BI-D1870, A66/TGX 221/ZSTK474, and AZD1480, targeting TGFβ-RI, BMP-RI, VEGFR/PDGFR/FGFR, adenylate cyclase, ribosomal S6 kinase (RSK), PI3K, and JAK2 signalling, respectively. Using (1) blastocyst quality (by morphological grading), (2) cell numbers (by differential stain), and (3) lineage-specific candidate gene expression (by quantitative PCR) as readouts, we sought to identify positive and negative regulators of ICM development and lineage determination. Based on our previous digital mRNA profiling data (McLean et al. 2014 Biol. Reprod., in press), we selected discriminatory epiblast-specific (FGF4, NANOG) and hypoblast-specific (PDGFRα, SOX17) markers for qPCR analysis. Each inhibitor was compared, alone or in combination, to an appropriately diluted dimethylsulfoxide (DMSO) vehicle control in at least 3 biological replicates. Statistical significance was determined using a generalised linear mixed model with binomial distribution and logit link for developmental data and REML for log cell counts and log gene expression data, applying fixed treatment effects and random run and sample within run effects. Blocking TGFβ1-, BMP- or VEGF-/PDGF-/FGF-signalling did not affect blastocyst development, ICM v. trophectoderm (TE) cell numbers, or gene expression. Repression of PI3K signals via AG66 and TGX, but not ZSTK alone, modestly decreased grade 1–2 blastocyst development (P < 0.05) but had no effect on cell numbers or gene expression. Stimulating adenylate cyclase activity increased NANOG levels (2.5-fold; P < 0.05), while RSK inhibition reduced FGF4 and PDGFRα expression (4-fold and 2-fold, respectively; P < 0.05). Suppressing JAK-STAT signalling, on the other hand, consistently compromised grade 1–2 blastocyst development and ICM numbers relative to DMSO controls (18/235 = 7% v. 59/159 = 29%, n = 5 IVF runs; 12 v. 47 ICM cells, N = 25 and N = 7 embryos counted, respectively; P < 0.0001). Epiblast and hypoblast markers were up to 40-fold reduced (FGF4, NANOG, SOX17; P < 0.0001) or completely abolished (PDGFRα; P < 0.0001). This effect was specific to the ICM because TE numbers and TE-specific gene expression (CDX2, KTR8) were not significantly altered. In summary, we have established Day 8 blastocysts as a useful chemical screening platform and demonstrated that bovine ICM development critically depends on JAK-STAT signalling.