Differential effects of follicle-stimulating hormone glycoforms on the transcriptome profile of cultured rat granulosa cells as disclosed by RNA-seq

Abstract

It has been documented that variations in glycosylation on glycoprotein hormones, confer distinctly different biological features to the corresponding glycoforms when multiple in vitro biochemical readings are analyzed. We here applied next generation RNA sequencing to explore changes in the transcriptome of rat granulosa cells exposed for 0, 6, and 12 h to 100 ng/ml of four highly purified follicle-stimulating hormone (FSH) glycoforms, each exhibiting different glycosylation patterns: human pituitary FSH^18/21 and equine FSH (eqFSH) (hypo-glycosylated), and human FSH²⁴ and chinese-hamster ovary cell-derived human recombinant FSH (recFSH) (fully-glycosylated). Total RNA from triplicate incubations was prepared from FSH glycoform-exposed cultured granulosa cells obtained from DES-pretreated immature female rats, and RNA libraries were sequenced in a HighSeq 2500 sequencer (2 × 125 bp paired-end format, 10-15 × 10⁶ reads/sample). The computational workflow focused on investigating differences among the four FSH glycoforms at three levels: gene expression, enriched biological processes, and perturbed pathways. Among the top 200 differentially expressed genes, only 4 (0.6%) were shared by all 4 glycoforms at 6 h, whereas 118 genes (40%) were shared at 12 h. Follicle-stimulating hormone glycocoforms stimulated different patterns of exclusive and associated up regulated biological processes in a glycoform and time-dependent fashion with more shared biological processes after 12 h of exposure and fewer treatment-specific ones, except for recFSH, which exhibited stronger responses with more specifically associated processes at this time. Similar results were found for down-regulated processes, with a greater number of processes at 6 h or 12 h, depending on the particular glycoform. In general, there were fewer downregulated than upregulated processes at both 6 h and 12 h, with FSH^18/21 exhibiting the largest number of down-regulated associated processes at 6 h while eqFSH exhibited the greatest number at 12 h. Signaling cascades, largely linked to cAMP-PKA, MAPK, and PI3/AKT pathways were detected as differentially activated by the glycoforms, with each glycoform exhibiting its own molecular signature. These data extend previous observations demonstrating glycosylation-dependent differential regulation of gene expression and intracellular signaling pathways triggered by FSH in granulosa cells. The results also suggest the importance of individual FSH glycoform glycosylation for the conformation of the ligand-receptor complex and induced signalling pathways.

Keywords: Differential gene expression; Follicle-stimulating hormone glycoforms; Next Generation sequencing; RNA-seq; Transcriptome.

Figure 1.

Typical glycans attached to human pituitary FSH (hFSH), human recombinant FSH produced by Chinese hamster ovary cells (recFSH) glycoforms, and equine FSH (eFSH). The green and cyan bars and ribbons (3-D structures at right) indicate the common-α and hormone-specific FSHβ subunits, respectively. N-glycosylation sites are indicated by the numbers below the bars. The eFSHα subunit has 4 additional residues at the N-terminus, accounting for the difference in numbering. The glycan at position β7 in eqFSH (black dotted square) is absent in the bulk (90%) of the molecules contained in highly purified preparations [27]. Glycans in recFSH were taken from Mastrangeli et al [28]. Human FSH glycoform models created with the GLYCAM web tool are shown on the right with the subunits rendered as cartoons using PyMol. The same color scheme is employed for subunits and glycans, which are rendered as spheres. Partially obscured Asn²⁴ glycan is colored dark blue to distinquish it from Asn⁷ glycan (light blue).

Figure 2.

Venn diagrams of the top-200 differentially expressed genes between FSH glycoforms at 6 h (A) and 12 h (B). Numbers inside the figure represent the number of genes shared in the corresponding set. Gray scale is proportional to the number of genes in the subset compared to the total of genes. Note that the number of genes shared between the four groups (center of the figure) at 6 h is only 4 genes; meanwhile, at 12 h, the very same set is 118 genes, meaning that at 12 h, the four compounds behave much similar than at 6 h.

Figure 3.

Contrast between 6 and 12 hours for each isoform of FSH. In the four cases, the top200 more differentially expressed genes are more different between 6 hours and control than 12 hours vs control. The contrast of recombinant FSH shows that there is a strong difference at 6 and 12 hours, since the intersection between 12 and 6 hours is only 34 genes.

Figure 4.

Enriched processes of overexpressed genes for the four FSH glycoforms at 6 h, with the nodes corresponding to glycoforms in yellow. In this network representation, red squares represent significantly (p<0.01) enriched biological processes in the treatment in which there is a link. There are some shared processes i.e. biological functions that appear enriched in more than one phenotype, which is indicated by more than one link. It is important to consider the number of associated processes for each isoform: eqFSH, 8 processes (2 exclusive); recFSH, 11 processes (only 1 exclusive); FSH^18/21 13 processes, from which 6 are exclusive, and FSH²⁴ only one associated process. In this figure, the intersection between FSH^18/21 at 6 h is depicted against the other 3 isoforms at the same time.

Figure 5.

Biological processes associated with overexpressed genes at 12 hours after FSH glycoforms addition. Color code is the same as in Fig. 4. It is evident that the number of shared processes was larger at 12 h than at 6 h of exposure. See also legend of Fig. 4.

Figure 6.

Biological processes associated with underexpressed genes at 6 hours of exposure to each of the four FSH treatments with the nodes representing each glycoform in yellow. In this representation, blue color represents the enriched processes.

Figure 7.

Biological processes associated with underexpressed genes at 12 hours of exposure to each of the four treatments. In this representation, blue color represents the enriched processes.

Figure 8.

Validation of differentially expressed genes in the RNA-seq data by RT-PCR of five selected FSH glycoform-sensitive genes (Pld1, Npy1R, Amh, Vegf-B, and Bcl2l1). Genes examined that were putatively up-regulated by FSH glycoforms included: phospholipase D1 (Pld1); neuropeptide Y (NPY) receptor Y1 (Npy1R), antimüllerian hormone gene (Amh), and vascular endothelial growth factor B (Vegf-B), whereas the FSH down-regulated gene examined was BCL2 like 1 (Bcl2l1).