Preimplantation Endometrial Transcriptomics in Natural Conception Cycle of the Rhesus Monkey

There is no report on preimplantation phase endometrial transcriptomics in natural conception cycles of primates. In the present study, the whole-genome expression array of endometrium on Days 2, 4, and 6 post-ovulation (pov) in proven natural conception (Group 1; n = 12) and non-mated, ovulatory (Group 2; n = 12) cycles of rhesus monkeys was examined, compared, and validated. Of fifteen (15) genes showing differential expression (>2-fold; pFDR < 0.05), six genes (CHRND, FOXD3, GJD4, MAPK8IP3, MKS1, and NUP50) were upregulated, while eight genes (ADCY5, ADIPOR1, NNMT, PATL1, PIGV, TGFBR2, TOX2, and VWA5B1) were down regulated on Day 6 pov as compared to Day 2 pov in conception cycles. On Day 6 pov, four genes (ADCY5, NNMT, TOX2, and VWA5B1) were down regulated, and AVEN was upregulated in conception cycles compared with the non-conception cycle. These observations were orthogonally validated at protein expression level. Group-specifically expressed unique genes in conception cycles influence the process of induction of immune-tolerance, while the genes expressed in both groups influence processes of protein targeting and metabolism. A triad of timed-actions of progesterone, seminal plasma, and preimplantation embryo putatively regulate several input molecules to CREB, NF-kB, and STAT regulatory networks during secretory phase towards evolution of endometrial receptivity in the rhesus monkey.


Introduction
The biology of endometrium during embryo implantation in the human is a 'black box' [1,2]. Progress in this field has been limited largely due to non-availability of human endometrial samples due to ethical and practical constraints. There are also serious restrictions and scarcity in using human embryos for in vitro experimental purposes and in the study of early events in implantation. In an alternate model to supplant this obstacle, investigators examined the physiological changes occurring in endometrium in the 'window of implantation' in normal, non-conception cycles of women [3,4]. Licht et al. (1998) employed an interesting model to investigate the nature of endometrial responses following intrauterine infusion of putative embryonic factors in the secretory phase of cycle [5]. Although these studies revealed many interesting facets of implantation biology, the complete absence of a dynamic relationship between a developing preimplantation stage embryo and the primed endometrium as it attains its receptive status for nidation was a major limitation in these experimental approaches [2,6,7].
There are several excellent reviews on the nature of embryo-endometrial dialogue and establishment of receptivity for ovo-implantation using rodents, lagomorph, ruminants, and ungulates as experimental models [8][9][10][11][12]. However, the endocrine basis and the temporo-spatial aspects of embryo implantation in these species bear little resemblance to that in the human. During the past few decades, the rhesus macaque (Macaca mulatta) as a non-human primate has been one of the animals of choice for preclinical and biomedical research because monkeys and humans share significant developmental, immunological, anatomical, biochemical, and physiological similarities [13,14]. The rhesus macaque bears marked similarities to the human in their reproductive characteristics, which bear special relevance in the study of human embryo implantation and pregnancy. These include similarities in (i) the characteristics of menstrual cycle [15,16], (ii) the endocrine requirements for pregnancy establishment [17][18][19][20], (iii) the preimplantation embryonic factors [21][22][23], (iv) the temporo-spatial features of initial interaction between trophoblast and implantationstage endometrium and establishment of subsequent lacunar stage of gestation [24][25][26][27][28], and, (v) finally, the establishment of villous placenta and the interaction between trophoblast and spiral arterioles [29][30][31]. These considerations led us to examine the differential display of biochemical characteristics between the preimplantation stage endometrium in proven conception cycles and secretory phase endometrium of non-conception cycles of fertile rhesus monkeys [2]. In the present study, we have examined the whole-genome expression of preimplantation stage endometrium in natural conception cycle and compared the former with the secretory phase of normo-ovulatory non-mated cycles of proven fertile rhesus monkeys and observed that functional networks involving CREB, NF-kB, and STAT regulatory modules in implantation stage endometrium may be operative in natural conception cycle.

General Procedures
Healthy, mature, and proven fertile female and male rhesus monkeys (Macaca mulatta) housed in a semi-natural condition at the Primate Research Facility of the All India Institute of Medical Sciences (New Delhi, India) were used in the study. The study was conducted in accordance with the guidelines of the Committee for the Purpose of Control and Supervision of Experiments on Animals and was approved by the Ethics Committee for the Use of Primates in Biomedical Research of the All India Institute of Medical Sciences (No.F.1-8/PRF/2004/780922/2021). The details of animal selection, housing, management, monitoring of cycles, mating, laparotomy, endometrial sampling, serum collection, and immunoassays for hormones have been described elsewhere [23].

Animal Groups and Tissue Collection
A total number of 71 proven fertile female monkeys showing at least two consecutive ovulatory menstrual cycles of normal length (26-32 days) were allocated to either of the two groups. One group of females (n = 35) were allowed to cohabit with proven fertile males during days 8 to 16 of their menstrual cycles (mated group), and the second group of females (n = 36) were not allowed to cohabit but were housed in close proximity to proven fertile males (non-mated group). Vaginal smears were checked daily. The day of ovulation was assessed from immunopositive profiles of estradiol-17β and progesterone in peripheral serum samples as described earlier [23]. The females (n = 10) who failed to show clear indication of ovulation were removed from the study.
Endometrial samples were collected on Days 2, 4, and 6 after ovulation from both groups of animals as described elsewhere [32]. Briefly, retrievable endometrial samples were collected on Days 2, 4, and 6 after ovulation from animals following laparotomy and fundal hysterotomy under ketamine (12 mg/kg body weight, Vetlar, Parke-Davis, Mumbai, India) anesthesia. Out of 35 females in the mated group, endometrial samples obtained from 31 females who yielded viable, preimplantation stage-matched embryos on uterine flushing were categorized under the conception cycle group (Group 1), having three sub-groups: Group 1A (Day 2 after ovulation; n = 10), Group 1B (Day 4 after ovulation; n = 11), and Group 1C (Day 6 after ovulation; n = 10). The recovered embryos from this group were immediately assessed for their developmental status under stereozoom microscope, as described earlier [2,23,32], and were employed in a different study. In the non-mated group, endometrial samples obtained from females (n = 30) who showed normal ovulatory cycles were categorized in the non-fecund cycle group (Group 2) with three subgroups: Group 2A (Day 2 after ovulation; n = 10), Groups 2B (Day 4 after ovulation; n = 10), and Group 2C (Day 6 after ovulation; n = 10). A total of 31 tissue samples from both groups were washed in icecold PBS and immediately processed for RNA extraction. Another set of 24 samples from both groups were processed for chemical fixation in 4% (w/v) buffered paraformaldehyde for making their paraffin blocks, which were used for immunohistochemistry [30]. Table 1 shows the details of the tissue samples from both groups used for different experiments.

RNA Extraction
Total RNAs from selected endometrial samples (n = 31) from the conception cycle group (Group 1; n = 15) and the non-fecund cycle group (Group 2; n = 16) were extracted using Trizol and cleaned up with DNase 1 [33]. The yield and purity of the extracted RNA was checked using standard protocols of absorbance ratio between 260 and 280 nm and 1% agarose gel electrophoresis [34]. Furthermore, the RIN score of individual samples was determined using Agilent 2100 Bioanalyzer, RNA 6000 Nano Lab Chip kit, and Agilent 2100 Expert Software (version 02/2000, Agilent Technologies, Inc., Santa Clara, CA, USA) as described elsewhere [35]. Samples (n = 7) that could not yield either adequate amount of RNA or RIN score > 8.0 were discarded from the study (Table 1). Thus, RNA samples obtained from 24 females were obtained from two groups (Group 1, n = 12; Group 2, n = 12) of animals for downstream experiments. All accessory chemicals were purchased from Sigma-Aldrich (St. Louis, MI, USA) and Agilent Technologies (Shung Avenue, Singapore).

Whole-Genome Expression Array
The samples having RIN scores > 8.0 were subjected to a whole transcriptome array using the Rhesus Monkey Whole Genome 60-mer 4X44K expression microarray slides purchased from Agilent Technologies, Inc. (Santa Clara, CA, USA) and according to the procedures provided by the manufacturer. Hybridized arrays were scanned using the Agilent G2505B microarray scanner system, and the raw data were imported into GeneSpring v14.9.1 software (Agilent Technologies, Santa Clara, CA, USA) for further analysis. Pearson's correlation coefficients were used to assess the reliability of data obtained from two separate hybridization runs for same RNA preparations to confirm the reproducibility assurance among hybridizations. Analysis of the data retrieved from separate chips with the same RNA samples yielded QC statistics highly concordant with those of the manufacturer, and it revealed more than 95% confidence level. Table 2 provides basic features of the expression microarray chips and data, the details of which are available at GEO repository accession number GSE108760.

Array Data Analysis and Post-Hoc Enrichment Analysis
The array data were subjected to exploratory analysis using unsupervised and supervised hierarchical clustering analysis (HCA) and non-hierarchical K-mean cluster analysis. Further, the data were analyzed using Welch ANOVA followed by pair-wise analysis using Tukey's HSD test with Benjamini-Hochberg multiple testing corrections for false discovery rate to identify differentially (>2-fold at pFDR < 0.05) expressed (DE) genes among various groups and subgroups. All data processing and analysis were performed using the Gene-Spring software v14.9.1. For post-hoc enrichment analysis, candidate genes were matched with known products into the sets of functional ontologies. The probability of a random intersection between a candidate on the target list and ontology entities was estimated in terms of p-values. Thus, a lower p-value meant higher relevance of the entity to the data set due to a higher rating for the entity. Enrichment analyses were performed using a standard cut-off threshold (pFDR = 0.05) for the unique genes perennially expressed only in one of the two groups investigated, as well as, for genes which expressed in both groups. The enrichment analysis for gene ontology (GO) was achieved using a MetaCore bioinformatics platform (GeneGO, St. Joseph, MI, USA) as described elsewhere [35][36][37].

Quantification of Candidate Gene Expression by Real-Time RT-PCR
In order to validate the DE data from microarray experiments, relative expression of fifteen (15) genes (ADCY5, ADIPOR1, AVEN, CHRND, FOXD3, GJD4, MAPK8IP3, MKS1, NNMT, NUP50, PATL1, PIGV, TGFBR2, TOX2, and VWA5B1) was assessed on a real-time RT-PCR platform using SYBR green chemistry and a CFX96 real-time RT-PCR detection system (BioRad, Hercules, CA, USA). The genes were selected based on results from the expression array experiment. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH), betaactin (ACTB) and beta-2-microglobulin (B2M) were selected as endogenous controls based on the observed expressional consistency according to MIQE guidelines [38]. Primers (see Supplementary Materials Table S1 for details) were designed using Beacon Designer software 7 (Labware Scientific Inc., Milipitas, CA, USA) and obtained from Integrated DNA Technologies (Coralville, IA, USA). The ratio of estimated efficiency of the primers for the selected genes and the endogenous control genes was~1.0. An optimized kit (Revert AID First strand cDNA synthesis kit K1612, Thermo Fisher Scientific, Vilnius, Lithuania) was used to synthesize cDNA from respective RNA (2 µg) samples and subsequent amplification was done using SYBR Green master mix obtained from BioRad (Hercules, CA, USA). The procedural details have been described elsewhere [37], and fold changes between groups were calculated using a standard model [39] and statistically evaluated using the Mann-Whitney U-test with Bonferroni correction [38]. A cut-off value of >2-fold change and pFDR < 0.05 were used to identify DE genes.

Immunohistochemical Localisation of Candidate Proteins for Orthogonal Validation
As shown in Table 1, 24 endometrial samples collected from different sets of animals of both groups (Group 1: n = 12 and Group 2: n = 12) were used for immunohistochemical staining for detection of ten (10) proteins (ADCY5, AVEN, CHRND, FOXD3, GJD4, MKS1, NNMT, TGFBR2, TOX2, and VWA5B1), which were selected based on results from post-hoc enrichment analysis. Immunohistochemistry was performed using specific antibodies commercially obtained from Abcam (Cambridge, MA, USA) and according to the methods provided by the manufacturer. Dilutions of stock primary antibodies for incubation were pre-calibrated based on five-point titration and the information provided by the manufacturer. The sources of primary antibodies and the dilutions at which those antibodies were used in this study are provided in Supplementary Materials Table S2. The procedural details of immunohistochemistry has been described elsewhere [30]. Briefly, deparaffinized and hydrated tissue sections were subjected to microwave heating in 0.1 M sodium citrate buffer (pH 6.0) for the retrieval of antigens. The endogenous peroxidase activity was quenched, and non-specific binding was blocked with non-immune sera. Final visualization was achieved using the ABC peroxidase Elite kits (Vector Laboratories, Burlingame, CA, USA) and freshly prepared 3,3 -diaminobenzidine hydrochloride with hydrogen peroxide. All immunostaining procedures for a given antigen were performed in a single run along with primary antibody and secondary antibody controls. The images were (i) viewed and documented using a Leica DMRD microscope (Leica Microsystems, Wetzlar GmbH, Germany), (ii) analyzed by a computer-assisted image analysis system (Leica QWIN DC 200, Cambridge, UK), and (iii) scored for different compartments independently by three trained observers using a simple standardized five scale combinative semi-quantitative scoring system (0 (<5%), 1 (5-25%), 2 (26-50%), 3 (51-75%), 4 (>75%)) and positive controls as references, as described earlier [40,41]. All scores provided by the observers were entered into data sheets on Excel platform and were analyzed using weighted κ-statistics for assessment of inter-observer errors yielding a mean κ-score of 0.65, suggesting a good agreement beyond chance [42]. Statistical analysis was done using (i) a Scheirer-Ray-Hare test, which is a non-parameter alternative to multi-factorial ANOVA for data from samples in different sub-groups, (ii) a Kruskal-Wallis test as a non-parameter alternative to ANOVA for data from samples in different sub-groups in a specific group, and (iii) a Mann-Whitney U-test with Bonferroni correction for post-hoc analysis of data between any two sub-groups [43,44]. Standard statistical packages, viz., SPSS for Windows version 16 (SPSS Inc., Chicago, IL, USA) and Graph Pad version 9 (GraphPad Software Inc., La Jolla, CA, USA) were used.

General
A total of 71 proven fertile female animals were used in two groups. A total of 35 animals were allowed to cohabit with proven fertile males during days 8 to 16 of the cycle. However, four animals failed to exhibit clear indication of ovulation, and two animals failed to yield any embryo despite ovulation, while another two animals yielded delayed embryos; thus 27 mated, ovulated animals yielded age-and-stage matched embryos during Days 2, 4, and 6 after ovulation. This group of animals was categorized under the conception cycle group (Group 1). Total RNAs from endometrial obtained from those 15 animals during peri-implantation period (Days 2, 4, and 6 after ovulation) in proven conception cycles were extracted followed by their quality and quantity assessments. Samples from three (3) animals were not used due to inadequate RNA quality and quantity, thus yielding RNA samples from 12 animals for the whole-genome expression array experiment. The tissue samples obtained from rest of animals in Group 1 (n = 12) were used for immunohistochemistry. Of the 36 animals who were not allowed to cohabit with proven fertile males, six animals failed to exhibit a clear indication of ovulation. Tissue samples obtained from 16 animals in this group categorized in the non-fecund cycle group (Group 2) were used for total RNA extraction followed by their quality and quantity assessments, and four samples could not be used since their RIN scores were less than 8.0. The tissue samples obtained from rest of animals in Group 2 (n = 12) were used for immunohistochemistry. Thus, 24 individual samples in different subgroups in duplicates were used for expression array experiments and a different set of 24 samples for immunohistochemistry towards orthogonal validation (see Table 1 for details).

Descriptive Data Analysis
A distribution analysis of the number of probes and genes for different ranges of microarray expression obtained from 24 proven fertile female rhesus with confirmed indication of ovulation revealed normal distribution individually and in group-wise analyses with acceptable and consistent background and signal-to-noise ratio. For total numbers and percent estimates of probes and genes expressed and estimated in an optimized scale for samples from different groups, see Table 2. On average,~60% and~68% of total genes were expressed in endometrial samples obtained from the conception cycle group (Group 1) and non-fecund cycle group (Group 2), respectively.
Group-wise HCA yielded highest segregation between the peri-implantation stage (Groups 1b and 1c) and other (Groups 2a-c and Group 1a) samples with cluster distance (cd) 0.9, when cd = 1.0 denotes complete segregation. On the other hand, least segregation was seen between Groups 1a and 2a with cd 0.2, when cd = 0 denotes complete aggregation. The results obtained from K-means cluster analysis also revealed two distinctive clusters: Cluster 1 comprised Groups 1a and 2a-c, while Cluster 2 comprised Groups 1b and 1c. Supplementary Materials Figure S1 shows the results of two-way HCA and K-means cluster analysis. Figure 1 shows the synopsis of comparative profiles of genes expressed in different subgroups as revealed from Venn analysis of expression data with the details provided in Supplementary Materials Tables S3-S5. There were a few notable observations. Firstly, between-groups analysis revealed that the numbers of expressed genes matching between groups displayed a clear declining trend with increasing number of expressed non-matching genes with time, maximal (~50%) non-matching genes (e.g., 3006 out of total 6617 expressed genes in Groups 1c and 2c) being observed on Day 6 after ovulation in conception cycle, i.e., Group 1c ( Figure 1D). Secondly, within-group analysis revealed that about one out of four expressed genes were matching on different days after ovulation in the conception cycle group (viz. 1159 out of total 4176), while about one out of two expressed genes were matching on different days after ovulation in the non-fecund cycle group (viz. 2191 out of total 4866). Thirdly, within-group analysis also revealed that the numbers of expressed genes common on different days in Group 1 (i.e., conception cycle group) showed a declining trend with days (see values in green italics in Figure 1B) viz., 3.2 × 10 3 , 2.5 × 10 3 and 1.9 × 10 3 genes on Days 2, 4, and 6 after ovulation, respectively. This feature was not observed in Group 2, i.e., the non-fecund cycle group (see values in red italics in Figure 1C), viz., 3.9 × 10 3 , 2.6 × 10 3 and 2.8 × 10 3 genes on Days 2, 4, and 6 after ovulation, respectively.

Differentially Expressed (DE) Genes
The analysis for DE matching genes in all subgroups based on ANOVA of data sets obtained from whole-genome expression arrays identified fifteen (15) DE genes. Of these 15 genes displaying differential expression, in Group 1c, the expression of six (6) genes (CHRND, FOXD3, GJD4, MAPK8IP3, MKS1, and NUP50) was upregulated as compared to Group 1a, and one (1) gene (AVEN) as compared to Group 2c. In Group 1c, on the other hand, eight (8) genes (ADCY5, ADIPOR1, NNMT, PATL1, PIGV, TGFBR2, TOX2, and VWA5B1) displayed downregulation as compared to Group 1a; of those genes, four (4), e.g., ADCY5, NNMT, TOX2, and VWA5B1, were also seen to be down regulated as compared to 2c. The results obtained from whole-genome expression arrays were validated by real-time RT-PCR. Table 3 provides the summary of the results.  Table 4 and Supplementary Materials Table S6 details the results of the GO analysis of group-specific unique expression data sets. It revealed that (i) all ten top-scored GO processes in the conception cycle group (Group 1) were aligned to the processes related to tolerance induction to non-self antigen involving T-cells, and (ii) all ten top-scored GO processes in the non-fecund cycle group (Group 2) were aligned to the metabolic processes, both with very high level of significance. Further, the genes found to be expressed in both groups aligned with GO processes such as protein targeting to membrane (p-value = 1.1 × 10 −21 ) and endoplasmic reticulum (p-value = 3.7 × 10 −21 ), as well as metabolic process (p-value = 4.0 × 10 −21 ), however, at much lower levels of significance compared to those uniquely expressed in Groups 1 and 2. It is notable that these processes could not be identified in top ten-scored GO processes in the conception cycle group (Group 1), while the metabolic process was identified at a much higher significance (p-value = 5.6 × 10 −55 ) in the non-fecund cycle group (Group 2).   Table S7 record the results of the semi-quantitative scores of immunopositivity for target antigens in epithelial and stromal compartments of monkey endometrium. Collectively, it appears that the degree of immunoprecipitation for AVEN, CHRND, FOXD3, GJD4, and MKS1 was markedly higher, while that for ADCY5, NNMT, TGFBR2, TOX2, and vWA5B1 was markedly lower in Group 1c compared with other sub-groups, particularly Group 2c. Thus, the protein expression profile of target genomic products validated the primary transcriptomics data. Some of these proteins were predominantly detected in epithelial cells (viz. FOXD3, GJD4, and MKS1), while a few of them were detected in both epithelium and stroma (viz. AVEN and CHRND). Endothelial cells of spiral arterioles showed a marked immunopositive stain for ADCY5 (see Figure 3) and VWA5B1 (see Figure 3) in the Group 2c (i.e., non-fecund cycle group, Day 6 after ovulation) samples, and for AVEN (see Figure 2) in the Group 1c (i.e., conception cycle group, Day 6 after ovulation) samples.

Discussion
Blastocyst implantation in several primates including the rhesus monkey and the human proceeds through timed attachment and adhesion to the uterine luminal epithelial wall in the receptive phase of the pregnancy cycle followed by disruption of junctional complexes between luminal epithelial cells by invasive trophoblasts [25]. In the period preceding embryo attachment, maternal endometrium in the natural conception cycle around Days 5-6 after ovulation when preimplantation embryos (PIEs) are mostly zonaencased reportedly display differential physiology as compared with that from a normal ovulatory cycle [2,7,19]. In the following stage, zona shedding and attachment-adhesion of the embryo occurs around Days 7-8 of gestation [2]. The possibility that genomic expression of receptive endometrium in such natural conception cycles exhibit differential

Discussion
Blastocyst implantation in several primates including the rhesus monkey and the human proceeds through timed attachment and adhesion to the uterine luminal epithelial wall in the receptive phase of the pregnancy cycle followed by disruption of junctional complexes between luminal epithelial cells by invasive trophoblasts [25]. In the period preceding embryo attachment, maternal endometrium in the natural conception cycle around Days 5-6 after ovulation when preimplantation embryos (PIEs) are mostly zonaencased reportedly display differential physiology as compared with that from a normal ovulatory cycle [2,7,19]. In the following stage, zona shedding and attachment-adhesion of the embryo occurs around Days 7-8 of gestation [2]. The possibility that genomic expression of receptive endometrium in such natural conception cycles exhibit differential display as compared to non-fecund cycles remained unexplored. In the present study, we have examined this possibility using rhesus monkey whole-genome expression in endometrium collected from mated monkeys yielding viable, synchronous preimplantation embryos and non-mated female monkeys with normal ovulatory cycles on Days 2, 4, and 6 after ovulation.

Genes Expressed in Both Groups and Involved in Protein Synthesis and Secretion along with Metabolic Processes Might Be Regulated by Progesterone
The GO analysis of endometrial transcriptomics in the present study revealed that the secretory phase endometrium of both conception and non-fecund cycles was perennially engaged in the protein synthesis and secretion along with marked involvement of metabolic processes. Table 4 and Supplementary Materials Table S6 substantiate the notion. A large number of genes related to protein synthesis and metabolism in endometrium are regulated in the estrogen-primed, progesterone-dominated secretory phase in several eutherian mammals including primates [2,[45][46][47][48]. Progesterone-dependent differentially regulated expressions in endometrium take place presumably to support embryo growth and endometrial preparation for blastocyst implantation [2,49].

Genes Uniquely Expressed in Natural Conception Cycle and Involved in Induction of Immunotolerance to Non-Self Antigens Might Be Regulated by Factors in Seminal Plasma
Post-hoc analysis of the genes markedly expressed only in the preimplantation stage endometrium of mated animals, resulting in viable conception potentially reflect a characteristic bias towards processes involved in the major histocompatibility factor (MHC)-1. (i.e., MAMU-A and MAMU-E in the rhesus monkey) dependent induction of immunotolerance to non-self antigens [50,51]. This observation leads us to propose the following points in relation to the phenomenon of non-rejection of embryonic allograft by the receptive stage maternal endometrium during blastocyst implantation and establishment of pregnancy. i.
Immunotolerance to ovo-implantation in a natural pregnancy cycle is a primary process triggered in receptive endometrium, and it is not exclusively dependent on trophoblastic input, although it has-in all likelihood-reinforcing action to the process for successful establishment of pregnancy [52,53]. ii.
The regulation of underlying genomic expression of this process is not primarily dependent on progesterone, as the levels of progesterone in maternal circulation during the preimplantation period generally do not show any difference between fecund and non-fecund groups of animals [54,55]. iii.
The underlying genomic expression of the immunotolerance process in secretory phase endometrium is unlikely to be regulated by the preimplantation stage embryo because, in primates including the rhesus macaque, embryos seldom reach uterine lumen on Day 2 after ovulation [56]. Still, in the present study, we observed from transcriptomic expressions, endometrium geared towards immune tolerance to non-self antigens on Day 2 post-fertilization onwards.

iv.
A plausible explanation for the observed differentiation towards immunotolerance in the secretory phase endometrium of mated and proven fecund group of animals could be due to the presence of seminal plasma inducing the observed genomic expression. In fact, evidence based on a series of elegant experiments is now available across various mammalian species to indicate that seminal plasma exerts substantial modifying action on endometrium towards receptivity, ovo-implantation, immunotolerance, embryo development, and establishment of pregnancy [38,52,[57][58][59][60][61][62]. To our knowledge, this is the first report supporting the notion that seminal plasma could induce at the genomic expression level a process bias for immunotolerance to non-self antigens in the secretory phase endometrium of natural conception cycle in the rhesus macaque.
The questions related to how and why seminal plasma might influence immunetolerance in the endometrium of a pregnancy cycle appear challenging. Based on observed enrichment analysis of input data for GO processes in the present study, our hunch is that a shift in the reproductive tract cytokine milieu by factors present in seminal plasma induced peripheral anergy engaging MHC molecules along with suppression of co-stimulatory signals [52,63]. While understanding such proximate causation shall indeed provide direct leads for IVF-ET technology [47], the studies on supportive role of seminal plasma on endometrial receptivity in eutherian mammals from the viewpoints of reproductive ecology and evolution may yield knowledge of far-reaching significance [64,65]. These issues deserve attention in the future.

Genes Involved in Cell Polarity, Differentiation, Signaling, Fate, and Immunomodulation Might Be Regulated by Embryo-Derived Signals
The observation that a specific group of genes (AVEN, CHRND, FOXD3, GJD4, MAPK8IP3, MKS1, and NUP50) displayed upregulation, while another set of genes (ADCY5, NNMT, PATL1, PIGV, TGFBR2, TOX2, and VWA5B1) displayed down regulation in the endometrium on Day 6 after ovulation in the conception cycle only in the presence of retrievable, stage-matched embryos appears intriguing. Indeed, quantitative RT-PCR and immunohistochemical analysis of the candidate markers substantiated the observations of their differential regulation in steady state transcript expression on Day 6 after ovulation of conception cycle. These factors are reportedly involved in regulation of epithelial cell polarity (GJD4 and MKS1), stromal cell differentiation (ADCY5 and TOX2), cell fate (AVEN, FOXD3 and NNMT), signal transduction and transfer (GJD4 and MAPK8IP3), and immunomodulation (CHRND and TGFBR2). Table 5 enumerates the physiological functions of some of these proteins. Many of these factors are being reported for the first time to be differentially regulated in the implantation-stage endometrium in the fecund cycle of the rhesus macaque. The physiological modules to which these factors belong are putatively involved in the process of implantation [2,19,[66][67][68][69][70][71][72][73][74]. Collectively, the results of the present study corroborate well with the notion that a preimplantation stage blastocyst on Day 6 of gestation imparts regulatory action, possibly via chemical mediators, on the physiology of receptive stage endometrium towards blastocyst implantation in the rhesus monkey [2,19].
Finally, based on the results obtained from the present study and available bioinformatics knowledge, we employed a causality clustering approach by grouping nodes into clusters according to their similarity in the overall information dynamics to obtain a simplified networking model [75]. This process yielded a functional network involving CREB, NF-kB, and STAT regulatory hubs as seen in Figure 6, and it appears to be operative in the implantation-stage endometrium. Previously reported several studies indeed have documented the critical importance of these regulatory clusters in the endometrium during blastocyst implantation [76][77][78][79]. Figure 6. Proposed functional networks involving CREB, NF-kB, and STAT regulatory clusters in implantation-stage endometrium. A causality clustering approach by grouping nodes into clusters according to their similarity in the overall information dynamics was employed to obtain a simplified networking model. Upregulated genes reportedly under progesterone regulation are shown in blue fonts in grey boxes, while up-regulated genes putatively under regulation of embryo-derived signals are shown in blue italic fonts in blue boxes, and downregulated genes under progesterone regulation are shown in red fonts in grey boxes.
Despite yielding novel leads regarding the transcriptomic expression of preimplantation stage maternal endometrium in natural conception cycle in a non-human primate species, the present study has limitations. Differential display analysis of expression data and histological features in immunohistochemistry in the present study did not indicate any marked variations in relative proportions of various cell types (viz., epithelial cells, endothelial cells, mesenchymal cells, smooth muscle cells, etc.) in the test samples. Nevertheless, quantitative immunopositivity scoring using specific antibodies against the repertoire of phenotypic markers, which would further validate the prediction, was missing in the present study [80][81][82]. Furthermore, the present study is a descriptive study conducted in semi-natural conditions and does not provide any direct proof of the concepts forwarded.  Despite yielding novel leads regarding the transcriptomic expression of preimplantation stage maternal endometrium in natural conception cycle in a non-human primate species, the present study has limitations. Differential display analysis of expression data and histological features in immunohistochemistry in the present study did not indicate any marked variations in relative proportions of various cell types (viz., epithelial cells, endothelial cells, mesenchymal cells, smooth muscle cells, etc.) in the test samples. Nevertheless, quantitative immunopositivity scoring using specific antibodies against the repertoire of phenotypic markers, which would further validate the prediction, was missing in the present study [80][81][82]. Furthermore, the present study is a descriptive study conducted in semi-natural conditions and does not provide any direct proof of the concepts forwarded. Table 5. Known physiological functions of proteins differentially expressed in endometrium in presence of zonal blastocyst on Day 6 of gestation.

Marker Proteins (Alias) Physiological Functions Reference
Adenylate cyclase 5 (ADCY5) Class III adenylate cyclase gene family expressed as transmembrane protein catalyzing the formation of cAMP in response to G-protein signaling. Required for stromal cell decidualization in mammals.
Thandapani et al. [84] Cholinergic receptor nicotinic, delta (CHRND) Family of pentameric acetylcholine-gated ion channels expressed in both neuronal, non-neuronal cells. Mediates signal transduction independent of generating ion currents in non-neuronal cells. Suppresses pro-inflammatory cytokines in the cholinergic anti-inflammatory pathway (CAP) in epithelial and endothelial cells.
Stokes et al. [85] Connexin 40.1 (GJD4) Involved in formation of gap junctional intercellular of communication (GJIC) for direct intercellular transfer ions and small molecules, cAMP, and cGMP and propagation of inositol trisphosphate-dependent calcium waves, involved in growth control, differentiation, and apoptosis of normal cells.
Saez et al. [86] Foxhead box D3 (FOXD3) Tumor suppressor. Member of forkhead box (FOX) family of transcription factors bearing helix-turn-helix DNA binding motif that interacts with DNA consensus sequences for nuclear localization and transcriptional regulation. Required for maintenance of cellular pluripotency and epigenetic priming of enhancers.
Krishnakumar et al. [87] Meckel syndrome1 (MKS1) Structural protein required for ciliogenesis. Also involved in planar cell polarity of epithelium and directional cell migration.
Jones et al. [64] TOX high mobility group, box family member 2 (TOX2) Transcription factor belonging to TOX family domain, involved in differentiation and maturation of different cell types.

Conclusions
In conclusion, we infer that a triad of timed-actions of progesterone, seminal plasma, and preimplantation stage embryo on the genomic expression of secretory stage endometrium potentially involving a functional network of CREB, NF-kB, and STAT regulatory modules is integral to the evolution of its receptive stage for blastocyst implantation in the rhesus monkey. Further studies with more focused questions shall unravel the underlying core mechanism and the combinative impact of actions of individual elements of the proposed triad on secretory phase endometrium to gather usable knowledge towards endometrial approach to fertility regulation and control.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/reprodmed3010003/s1, Figure S1: Group-wise two-way hierarchical cluster analysis of expression array data yielding highest segregation between peri-implantation stage (groups 1b and 1c) and other (groups 2a-c and group 1a) samples (A), while K-means cluster analysis revealing two distinctive clusters: cluster 1 comprised of groups 1a and 2a-c, while cluster 2 comprised of groups 1b and 1c (B). Note cluster distance (cd) 1.0 denotes complete segregation, while cd = 0 denotes complete aggregation. Truncated expression data are shown.; Table S1: List of genes and their primers used for quantitative real time RT-PCR; Table S2: Details of primary antibodies against target antigens, Table S3: List of expressed genes in the endometrium of Conception cycle, Table S4: List of expressed genes in the endometrium of Non-fecund cycle, Table S5: List of expressed genes in the endometrium, Table S6, List of input genes in network objects of top-scored GO processes for two groups, Table S7: Scores a of immunohistochemical staining in epithelial and stromal compartments of endometrial functionalis of different sub-groups.