CircRNA expression profile of bovine placentas in late gestation with aberrant SCNT fetus

Abstract Backgrounds One of the limitations of somatic cell nuclear transfer (SCNT) strategy to generate genetically modified offspring is the low birth rate. Placental dysfunction is one of the causes of abortion. Circular RNA (circRNA) is noncoding RNA which functions as microRNA (miRNA) sponges in biological processes. Methods Two aberrant pregnant placenta (aberrant group, AG) and three normal pregnant placenta (normal group, NG) during late gestation (180‐210 days) with bovine SCNT fetus were collected for high‐throughput sequencing and analyzed. The host genes of differentially expressed (DE) circRNAs were predicted. And the microRNAs (miRNAs) which could interact with DE circRNAs were analyzed. Then, the expressional level of partial DE circRNAs and corresponding host genes was verified through qRT‐PCR. At last, the function of host genes was analyzed through Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG). Results Altogether 123 differentially expressed circRNAs between two groups were identified, which were found related to 60 host genes and 32 miRNAs. The top 10 upregulated circRNAs were bta_circ_0012985, bta_circ_0013071, bta_circ_0013074, bta_circ_0016024, bta_circ_0013068, bta_circ_0008816, bta_circ_0012982, bta_circ_0013072, bta_circ_0019285, and bta_circ_0013067. The top 10 downregulated circRNAs were bta_circ_0024234, bta_circ_0017528, bta_circ_0008077, bta_circ_0003222, bta_circ_0007500, bta_circ_0020328, bta_circ_0011001, bta_circ_0016364, bta_circ_0008839, and bta_circ_0016049. The qRT‐PCR results showed consistent trend with sequencing analysis result, while host genes had no statistic difference. The GO and KEGG analyses of the host genes suggested that abnormal circRNA expression may play multiple roles in placental structure and dysfunction. Conclusion The abnormal circRNA expression may be one of reasons of placental dysfunction, leads to abortion of bovine SCNT fetus.


| INTRODUC TI ON
During the past two decades, tremendous progress has been achieved in animal cloning since the birth of Dolly. One major breakthrough in the field, somatic cell nuclear transfer (SCNT), has given birth to a barnyard of livestock animals, including cattle, pig, sheep, and goat. 1 Combined with gene editing technology, this technique had proven valid in developing genetically modified livestock. However, one of the bottlenecks of SCNT is low birth rate.
Only 6% of transferred cloned embryos result healthy offspring in cattle. 2 According to our previous research, the survival rate of genetically modified cloned cattle was below 5%. Incomplete reprogramming is amenable to the developmental failure of cloned embryo. 3,4 Except for the fetal aberrant development, placental dysfunction, such as reduced vascularization, placentomegaly, hypoplasia of trophoblastic epithelium, and altered basement membrane, was another cause to lead pregnancy losses. 2,4 We found that a fairly large number of cloned cattle aborted during late gestation (180-210 days). The abnormal pregnant recipient showed engorged uterus and enlarged umbilical vessels. Coincidentally, a equine clone research depicted similar symptoms. 5 It indicates that this is a relative common abnormality during SCNT fetal pregnancy, while the causes are ambiguous.
Placenta is a circular organ which temporarily exists in placental mammals during gestation. It not only supplies the space for fetus with protection and nutrition metabolism, but also secretes multiple growth factors and hormones to maintain gestation. In addition, it is the only pathway to connect the mother and the fetus.
Placental research by RNA-seq for abortion and aberrant pregnancy in livestock mainly focused on early gestation or postnatal. [6][7][8][9] little about which However, few studies have looked at the placenta during the third trimester, when large quantity of SCNT fetal abortion occur.
Circular RNAs (circRNAs) were first discovered in RNA viruses as early as the 1970s. 10 It formed as covalently closed loop structures with neither 5′-3′ polarities nor polyadenylated tails and more stable than linear RNA. 11 Serious reports showed that circRNAs could function as miRNA sponges, regulate alternative splicing, and modulate the expression of mRNAs. [12][13][14][15] The different types of RNAs serve different roles and form a network called the competing endogenous RNAs (ceRNAs). 16 Like other noncoding RNAs, circRNAs have been associated with a particular role in biological development and disease initiation and progression. 17 They have been found implicated with various cancers, including colorectal, lung, and cervical cancer. [18][19][20] Hitherto noticed features of circRNA are mainly based on evidence gathered from human, and studies on other species are insufficient. [21][22][23][24] This study aims to explore the multiple factors that potentially lead to high abortion frequency exist in SCNT fetus generation during late gestation. To this end, we collected two aberrant pregnant placenta (abnormal group, AG) and three normal pregnant placenta (normal group, NG) at late gestation (180-210 days) of bovine SCNT fetus. We acquired five bovine late gestational placental circRNA expression profiles and analyzed its differentiation. We sought to uncover the mechanism associated with this phenomenon.
The discovery may provide a new insight for SCNT fetal aberrant development and improve the SCNT efficiency.

| Ethics statement
All experimental procedures and sample collections were conducted in accordance with the Regulations for the Administration of

| Sample information and collection
Cloned embryo, embryo transfer, and recipient cow experimental work were supplied by Inner Mongolia University. Briefly, the donor cell was fetal skin fibroblast. The recipients were 2-5 years. The procedure was followed as Wu et al. 25 A total of five late pregnant cows were used in the present study from two groups, that is, the aberrant pregnant cows (aberrant group, AG: n = 2) and normal pregnant cows (Normal group, NG: n = 3). All of the selected cows were at late pregnancy stage (180-210 days).
After the pregnant cows were slaughtered, the placenta was rapidly harvested and immediately frozen in liquid nitrogen and stored for use toward the subsequent generation of circle RNA libraries.

| RNA preparation
The total RNA was extracted using TRIzol™ reagent (Invitrogen) following the manufacturer's procedure. 26 Briefly, 50-100 mg of tissues was lysed by 1 mL of TRIzol™ reagent. 0.2 mL of chloroform per 1 mL of TRIzol™ Reagent was added after 5 minutes of incubation. Then, the samples were centrifuged for 15 minutes at 12 000 g at 4°C after 2-3 minutes incubation. The mixture separated into a lower red phenol-chloroform, and interphase, and a colorless upper aqueous phase. The RNA was contained in the aqueous phase. The aqueous phase was transferred to a new tube and added 0.5 mL of isopropanol. After incubation of 10 minutes, centrifuge for 10 minutes at 12 000 g at 4°C. The supernatant was discarded and 1 mL of 75% ethanol was added to wash RNA. Centrifuge for 5 minutes at 7500 g at 4°C. The supernatant was discarded and air-dried the RNA pellet for 5-10 minutes. At last, the RNA was resuspended in 20-50 µL of RNase-free water. The quantity and purity of total RNA were analyzed using the Bioanalyzer 2100 (Agilent) with RIN number >7.0.

| Library synthesis and highthroughput sequencing
Approximately 3 µg of total RNA was used to prepare the circRNA library. Ribo-Zero™ Gold Kits were used to degrade rRNA, and linear RNA was degraded by RNase R. Then, RNA libraries were generated according to the protocol outlined for NEBNext Ultra Directional RNA Library Prep Kit for Illumina (NEB). We then performed the single-end sequencing on an Illumina Hiseq2500 at the ANOROAD GENOME Co., Ltd. (Beijing, CN) following the vendor's recommended protocol.

| Differentially expressed circRNA analyses
The differentially expressed circRNAs between AG and NG were calculated by edge R using the likelihood ratio test (LRT) based on generalized linear model which estimates probability distributions according to mean-variance relationship of each gene. 27 Only transcripts with expression greater than 0.1 count per million (CPM) in at least one samples were selected for differential testing. Transcripts with P < 0.05 and |log2 ratio| ≥ 1 were considered differentially expressed.

| Validation of differentially expressed circRNAs through qRT-PCR
Eight differentially expressed circRNAs and relative host genes were selected for validation. Total RNA was extracted as previous.
PrimeScript™ RT reagent kit (TAKARA) was used to cDNA synthesis, and only random 6-mers were added. TB Green™ Premix Ex Taq™ II was used to qRT-PCR. The procedure was followed as the manufacturer's instruction book. The primer sequences were listed at supplemental Table S1. The GAPDH was used as reference gene. The relative expression level of each circRNA and host gene was calculated using the 2 −ΔΔCt method. The data are indicated as the means ± SE (n = 3).
The significance of the expression in two samples was calculated using a two sample t test in SPSS statistical software (Version17.0), whose difference was considered as significant when P < 0.05.

| Functional enrichment analysis of host genes of differentially expressed circRNAs
The enrichment analyses of KEGG (Kyoto Encyclopedia of Genes and Genomes) and GO (Gene Ontology) were performed using

DAVID (The Database for Annotation, Visualization, and Integrated
Discovery) with the default parameters.

| Target miRNAs of differentially expressed circRNA prediction and co-expression network analysis
The target miRNAs of differentially expressed circRNAs were evaluated using miRanda (3.3a), investigating only perfect seed matching without gap of wobble pairing ("strict" parameter). 28 A hit between any expressed miRNA (including the new predicted miRNA) and a target circRNA was considered for a miRanda score of 140 or higher, corresponding to at least a perfect seed match.

| Characteristics of bovine placental circRNA expression pattern
In this study, we analyzed two aberrant pregnant placenta (AG) and three normal pregnant placenta (NG) at late gestation These circRNAs were widely scattered on almost all bovine chromosomes, and chromosome 1 was the most abundant, followed by chromosome X and 2 ( Figure 1A). The properties of circRNAs contain classic, alter exon, intron, overlap exon, antisense, and intergenic. The compositional type of each sample is shown in Figure 1B.
In total, the ratio of classic was the largest, exceeding 60% in each sample. CircRNAs transcribed from three exons  were the most abundant circRNAs in all samples, followed by 2-exon and 4-exon circRNAs ( Figure 1C).

| Identification validation of differentially expressed circRNAs between AG and NG
Hierarchical cluster analysis was used to reveal the circRNA expression levels in AG and NG (Figure 2A), which showed that these levels were distinguishable between two groups. The significantly differentially expressed (DE) circRNAs between two groups were shown in the volcano plot ( Figure 2B). In total, 123 circRNAs were identified as differentially expressed circRNAs by the filter criteria of fold change (FC) ≥2.0, P value <0.05. Among these, 49 circRNAs were upregulated, and 74 circRNAs were downregulated ( Figure 2C).
The differentially expressed circRNAs were listed in supplemental host genes with no significant difference ( Figure 2D).

| Host gene enrichment of DE circRNAs and functional analysis
Through edge R analysis, 60 host genes of DE circRNAs were enriched (Table 1). Gene ontology (GO) analysis of these genes showed that there were two genes related to protein K48-linked analysis showed that the predicted host genes were associated with virus invasion and amine acid metabolism (Table 3). Unfortunately, the result was also not significant (P > 0.05). Combined with these results, we surmised that DE circRNAs may affect placental protein metabolism.

| Prediction of differentially expressed circRNA-miRNA interaction
CircRNAs act as miRNA sponges and exert their effects via the cir-cRNA-miRNA-mRNA axis. 29 Through miRanda based on the MREs, interaction between DE circRNAs and miRNAs was theoretically predicted. We found that 32 miRNAs could be paired with eight DE circRNAs (Table 4), with the criteria of a max score ≥140 and a max energy ≤−25 (a lower max energy is indicative of a stronger correlation). The result suggested that circRNAs may play a part in causing placental dysfunction via interaction with miRNAs. Glutamine plays a vital role in carbon and nitrogen metabolism of the fetus and exhibits the highest fetal-maternal plasma ratio among all amino acids in pigs. 38 These results indicated that differentially expressed circRNAs may have multiple effects in placental both structure and function.

| D ISCUSS I ON
CircRNAs could function with miRNAs and co-regulate target genes' expression. We predicted 32 miRNAs which can pair with eight differentially expressed circRNAs. Among these miRNAs, miR-145 was reported to be related to abnormal placental development in transgenic cloned cattle. 39 Our results indicated that cir-cRNAs may play a role in abnormal bovine fetus development in late gestation through interactions with miRNAs.
For it was one type of pregnancy familiar of bovine SCNT research and occurs randomly, the sample was not sufficient. CircRNAs as noncoding RNA need to contact with other RNA to act biological function. Further study is needed to explore the mechanism. Low birth rate of SCNT is a complicated question, and relative research should pay attention to placental dysfunction. Combined with multiple strategies, such as RNA expression, protein expression, histological and hormone analysis, the mechanism of bovine SCNT fetal abortion-related placental dysfunction will be discovered. It is also helpful to improve the SCNT efficiency.

| CON CLUS ION
In this study, we acquired five circRNA expression profiles of SCNT bovine placentas (two abnormal and three normal) during late gestation. We identified 123 circRNAs were DE circRNAs between AG and NG. 60 target genes and 32 miRNAs were related to DE circRNAs. Through GO and KEGG analyses, we surmise that abnormal circRNA expression may play multiple roles in placental both structure and dysfunction. In the future, we would detect related mRNA and miRNA expression profiles to further explore its mechanism.

ACK N OWLED G M ENTS
This work was funded by the National Natural Science Foundation of China (31460599). The authors would like to thank Professor Guangpeng Li of Inner Mongolia University for providing the recipient heifers and technical assistance.