Endometrial factors similarly induced by IFNT2 and IFNTc1 through transcription factor FOXS1

In ruminants, Interferon tau (IFNT) is the pregnancy recognition protein produced by the mononuclear trophectoderm of the conceptus, and is secreted into the uterine lumen during the peri-attachment period. In our previous study, the high-throughput RNA sequencing (RNA-seq) data obtained from bovine conceptuses during the peri-attachment period identified two IFNT mRNAs, IFNT2 and IFNTc1. However, how each of these IFNT variants regulates endometrial gene expression has not been characterized. Using RNA-seq analysis, we evaluated how IFNT2 and IFNTc1 affected transcript expression in primary bovine endometrial epithelial cells (EECs). IFNT treatment induced 348 differentially expressed genes (DEGs); however, there are few DEGs in IFNT2 or IFNTc1 treated EECs, indicating that IFNT2-induced DEGs were similar to those induced by IFNTc1 treatment. In in silico analysis, we identified four IFNT2- and IFNTc1-induced pathways: 1) type II interferon signaling, 2) proteasome degradation, 3) type III interferon signaling, and 4) DNA damage response. We further demonstrated that IFNT2 and IFNTc1 up-regulated several transcription factors, among which forkhead box S1 (FOXS1) was identified as the most highly expressed gene. Furthermore, the knockdown of FOXS1 in IFNT2- or IFNTc1-treated EECs similarly down-regulated 9 genes including IRF3 and IRF9, and up-regulated 9 genes including STAT1, STAT2, and IRF8. These represent the first demonstration that effects of each IFNT on EECs were studied, and suggest that endometrial response as well as signaling mechanisms were similar between two IFNT variants existed in utero.


Introduction
Interferon tau (IFNT), classified as a type I IFN along with IFN alpha, IFN beta and IFN omega, is the pregnancy recognition protein in all ruminants [1]. IFNT, produced by the mononuclear trophectoderm of the conceptus, is secreted into the uterine lumen during the peri-attachment period [1,2], although a few studies indicate that IFNT is detected in uterine vein serum [3,4]. IFNT down-regulates the expression of endometrial oxytocin receptors and a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 then maintains the corpus luteum function via inhibition of the luteolytic pulse of endometrial prostaglandin F2α [5][6][7]. Bovine IFNT increases on day 15 of pregnancy (day 0 = day of estrus), peaks on days 19-20 (conceptus attachment to endometrium begins on days 19-19.5), and is decreased soon after the initiation of conceptus attachment to the uterine epithelium [8][9][10][11].
To determine the number of IFNT genes expressed in the bovine uterus, several studies with the use of PCR have previously been executed [12,13]. We also utilized the high-throughput RNA sequencing (RNA-seq) analysis, identifying two IFNT transcripts, IFNT2 and IFNTc1, in RNAs obtained from pregnant days 17, 20, and 22 bovine conceptuses [9,14]. In addition, we showed that IFNT2 expression in utero was five-fold higher than that of IFNTc1 [9]. Over the past decade, various global analyses were performed to study the expression of transcripts in bovine endometrium [15][16][17][18][19][20]. These observations found changes in dynamic gene expression in pregnant and cyclic endometria, and identified optimal transcript and/or protein expression in utero for maintaining pregnancy at several stages in early pregnancy. In addition, those studies demonstrated the effects of IFNT and/or progesterone on differentially expressed genes in bovine endometrium. In the previous studies [12,13], IFNT variants identified were subjected to assays of antiviral or anti-luteolytic activity; however, how each of those IFNT variants affects endometrial gene expression has not been characterized.
Our previous studies showed that the upstream region of IFNTc1 gene did not possess the JUN-binding site found in the IFNT2 gene, and TEAD2 increased transcriptional activity of IFNT2 only, resulting in the differential expression between IFNT2 and IFNTc1 in in vitro and possibly in vivo [21,22]. In addition, both IFNT2 and IFNTc1 up-regulated IFN-stimulated genes (ISGs), including ISG12, ISG15, or MX dynamin-like GTPase (MX)1, while only IFNTc1 up-regulated the expression of MX2 in bovine endometrial epithelial cells [23]. We therefore hypothesized that effects of IFNT2 and IFNTc1 differ in the bovine endometrium. In this study, we evaluated how IFNT2 and IFNTc1 affected primary bovine endometrial epithelial cells using RNA-seq, followed by quantitative PCR analysis.

Cell preparation, culture condition
In this study, we did not perform any animal experiments. Bovine uterine endometrial epithelial cells (EECs) were collected from Holstein cows at local abattoir (Tsuyama Meat Center) in accordance with protocols approved by local institutional animal care [24], and the protocol for bovine cell cultures was approved by the Ethics Committee of the University of Tokyo (Permit Number: 449-2126). In brief, uteri of the early luteal phase (days 2 to 5) were used in this study. The hysterectomized uterine lumen was trypsinized (0.3% w/v) in order to detach the epithelial cells and then EECs were isolated. The isolated EECs were cultured on collagen type I-coated culture dish in DMEM/F12 (1:1) medium supplemented with 10% (v/v) FBS, 40 units/ml of penicillin, and 40 μg/ml of streptomycin at 37˚C under 5% CO 2 in humidified air [11]. Human 293T cells (CRL-3216, ATCC) were grown in DMEM supplemented with 10% (v/v) FBS and antibiotics at 37˚C in 5% CO2 [23].

Production and purification of recombinant IFNs
293T cells were transfected with the expression plasmid for IFNT2, IFNTc1 or IFNA [23] and culture media were collected at 48-72 hours after transfection [23]. Recombinant IFNs secreted to culture media from cells were purified using His-tagged protein purification reagent (Medical and Biological Laboratories, Nagoya, Japan) according to the manufacturer's instructions [23]. The titers of purified recombinant IFNT2, IFNTc1, and IFNA were determined by the assay using MDBK cells and VSV as previously reported [25].

RNA extraction and preparation for RNA-seq analysis
RNA was extracted from cultured EECs using Isogen (Nippon gene, Tokyo, Japan) according to the manufacturer's instructions. A portion of total RNA from cultured EECs treated with or without IFNT2 or IFNTc1 (n = 3 each) was pooled. High-throughput sequencing libraries were prepared using the SureSelect Strand Specific RNA Library Prep Kit (Agilent Technologies, Santa Clara, CA) according to the manufacturer's instructions, and analysis was performed by Kazusa DNA Research Institute (Chiba, Japan). Primary sequencing data were deposited to the DDBJ (DNA Data Bank of Japan) Sequence Read Archive (accession number DRA005460).

Mapping reads to the bovine genome
Nucleotide sequences identified by RNA-seq analysis were trimmed by PRINSEQ-lite v0.19.2. Trimmed sequences were generated as FASTQ outputs and analyzed on the basis of the TopHat/Cufflinks pipeline based on the bovine genome (bosTau8) and reference annotations obtained from UCSC genome browser (http://genome.ucsc.edu). Differential and significant gene expression analysis was performed with the use of gene-level FPKM (fragments per kilobase of gene locus summarized mRNA per million reads) expression levels. Genes were selected with the criteria of an absolute expression level >10 FPKM in either IFNT2-or IFNTc1-treated samples with at least 1.5-fold higher expression in IFNT2 or IFNTc1 than non-treated EECs.

RNA extraction and quantitative RT-PCR
Using ISOGEN reagent (Nippon gene), total RNAs were extracted from cultured EECs treated with IFNTs, which were performed three times independently. For real-time PCR analyses, isolated RNA (total 0.5 μg) was reverse transcribed to cDNA using the ReverTra Ace qPCR RT Kit (Toyobo, Osaka, Japan) according to the manufacturer's instructions. The cDNA reaction mixture was subjected to real-time PCR amplification using the Thunderbird SYBR qPCR Mix Kit (Toyobo) with primers listed in S1 Table, and PCR amplification was carried out on a Step One Plus real-time PCR System (Applied Biosystems, Foster City, CA). Amplification efficiencies of each target and the reference gene, bovine glyceraldehyde-3-phosphate dehydrogenase (GAPDH), were examined through their calibration curves and found to be comparable. The thermal profile for qPCR consisted of 40 cycles at 95˚C for 15 sec, and annealing and extension at 60˚C for 60 sec. Average threshold (Ct) values for each target were determined by Sequence Detection System software v2.2 (Applied Biosystems). Each run was completed with a melting curve analysis to confirm the specificity of amplification and the absence of primer dimer [11].

Statistical analysis
Data are expressed as the mean ± SEM. Significance was assessed using the Dunnet comparisons test. A P-value < 0.05 was considered statistically significant.

Differential gene expression between IFNT2-and IFNTc1-stimulated endometrial epithelial cells
To investigate effects of IFNT2 and IFNTc1 on gene expression in EECs, RNA-seq analyses were executed, detecting 348 differentially expressed genes (DEGs) among non-(Ctrl), IFNT2-and IFNTc1-treatment groups (S2 Table). The venn diagram shows the number of genes with 1.5-fold changes among these groups, and the right table lists increased or decreased genes in IFNT2 vs. IFNTc1 group, which overlap with Ctrl vs. IFNT2 or Ctrl vs. IFNTc1 group (Fig 1A). Although RNA-seq analysis found that 12 gene expressions were increased or decreased, qPCR did not detect changes in those gene expressions (Fig 1B). A pair plot comparison of IFNT2 and IFNTc1 treatments showed that there were very few DEGs, which were confirmed with high degree of correlation coefficient, 0.99 (Fig 1C). These results indicated that IFNT2-induced DEGs were similar to those of IFNTc1 treatment in EECs.

Discussion
In this study, we first demonstrated the global gene expression of primary bovine EECs treated with IFNT2 or IFNTc1, and identified a novel FOXS1 signaling pathway, resulting in IFNT response. Although several studies have shown variants of IFNT transcripts [9,12,13], their regulation and effects on EECs have not been characterized. In this study, IFNT2 and IFNTc1 induced transcripts associated with 4 enriched pathways. In addition, IFNT2 and IFNTc1 up-regulated several transcription factors, among which FOXS1 was found as the highest expressed gene. Furthermore, the knockdown of FOXS1 down-regulated 9 genes including IRF3 and IRF9, and up-regulated 9 genes including STAT1, STAT2, and IRF8. These findings suggested that upon IFNT stimulation, FOXS1 could have mediated up-or down-regulation of IFNT-stimulated transcription factors such as STAT1, STAT2 and IRFs, followed by activation of type II interferon, proteasome degradation, type III interferon, and DNA damage response signaling pathways. FOXS1 is Forkhead type transcription factor expressed in Sertoli cells and peri-endothelial cells of the developing mouse fetal testis [26]. Male and female Foxs1 knockout mice are fertile, but the mutant males accumulate blood in the fetal testis [26]. Foxs1 is also expressed in gonadal-like cells in Gata6 conditional knockout mice [27]. In addition to gonadal-like cells,  Foxs1 is expressed in neural crest derivatives [28][29][30]. However, molecular mechanisms on how FOXS1 regulates those phenotypic changes have not been characterized. In this study, FOXS1 was up-regulated by IFNT2 and IFNTc1, and its knockdown regulated general IFNTinduced transcription factors such as STAT1 and STAT2 in bovine EECs. In addition, IFNA treatment similarly increased FOXS1, STAT1 and STAT2 expression (S1 Fig); however, type I IFNs such as IFNA and IFNB do not exist in the bovine and ovine uterine lumen during periimplantation period. [12,31]. In human dermal fibroblast cells, FOXS1, up-regulated by STAT4, induces the differentiation into myofibroblast [32]. In addition, we confirmed the presence of STAT1-or STAT2-binding elements on FOXS1 promoter region (S2 Fig). These findings suggested that activation of STAT1 or STAT2 could induce FOXS1 expression and FOXS1 then down-regulate STAT1 and STAT2 expression: possible negative feedback loop between FOXS1 and STAT1/2 in the bovine endometrial epithelium (Fig 5).
In the pro-inflammatory response to control the immune system, proteasomes are replaced with immunoproteasomes, which are induced by interferon-gamma (IFNG) and increase the  production of peptides for presentation of MHC class I molecules [33]. The proteasome consists of three units, 20S proteasome, 19S regulator, and proteasome activator 28 (PA28), among which PA28 is composed of PSME1 and PSME2 [33]. 20S proteasome has 20 subunits, among which 3 subunits are replaced with PSMB8, PSMB9, and PSME10 by IFNG stimulation, resulting in immunoproteasome formation. In our results, IFNT2 and IFNTc1 up-regulated not only PSME1, PSME2, PSMB8, PSMB9, and PSMB10, but also class I antigen presentation-related genes such as BOLA-A, B2M, TAP1/2, and several ubiquitin ligases (S2 Table). These results were consistent with our and other previous studies which demonstrated changes in protein and gene expression during peri-implantation periods [11,20,34]. In addition, FOXS1 knockdown increased the expression of PSMB8, PSMB9, and PSMB10. These findings suggested that IFNT2 and IFNTc1 induced immunoproteasome formation and class I antigen presentation in endometrial epithelial cells, which could condition the EECs for interaction with semi-allogenic conceptuses.
Our previous study demonstrated that MX1, ISG12, ISG15, ISG17, IRF1, and IRF2 were similarly up-regulated by IFNT2 or IFNTc1 treatment, but MX2 was only up-regulated by IFNTc1 [23]. This study showed that IFNTc1 tended to increase the expression of MX2 compared with that of IFNT2. However, correlation coefficient of gene expression between IFNT2 and IFNTc1 was 0.99, indicating no or minimal difference in gene expression between IFNT2 and IFNTc1. The amino acid sequences of IFNT2 and IFNTc1 differ slightly, of which IFNTc1 has one casein kinase 2 phosphorylation domain whereas IFNT2 does not (S3 Fig). These results suggest that IFNT2 and IFNTc1 still differ in nucleotide structures and possibly functions, but further experiments are required to prove definitive functional differences between IFNT2 and IFNTc1.
In conclusion, this study demonstrated the global gene expression of IFNT2-or IFNTc1-treated primary bovine endometrial epithelial cell. One of these genes was transcription factor FOXS1, up-regulated by IFNT2 and IFNTc1, and its knockdown up-regulated STAT1 and STAT2. Therefore, FOXS1 could play a role as a negative feedback regulator of IFNTs signaling in bovine endometrial epithelial cells.