Embryo implantation is closely associated with dynamic expression of proprotein convertase 5/6 in the rabbit uterus

New Zealand White rabbits aged 3-4 months (2.6-3.2 kg) were maintained at Shanghai Institute of Planned Parenthood Research Animal Facility under controlled conditions (14 h light-10 h dark cycle, 22°C) with free access to food and water. All experimentation was performed in compliance with the laboratory animal care protocols approved by the Institutional Animal Care Committee of the Shanghai Institute of Planned Parenthood Research.

Virgin female rabbits were mated with males of the same strain, with the time of mating designated as day (d) 0 of pregnancy. Animals were euthanased at various time points following mating (d0, 1, 3, 5, 6, 6.5, 7, 8 and 10, n = 3 each) by injecting 20 ml air into the auricular veins. To induce pseudo-pregnancy, rabbits received an i.m. injection of 20 IU pregnant mare serum gonadotrophin, and 72 h later, an i.m. injection of 0.8 μg of gonadotrophin-releasing hormone (GnRH). Ovulation was confirmed in each animal by the presence of the corpus luteum in each ovary.

Total RNA was extracted using Trizol reagent (Invitrogen, Carlsbad, CA) as per the manufacturer's instructions. Contaminating DNA was removed using a DNAse free kit (Ambion, Austin, TX). RNA quality and quantity was assessed by optical densitrometry (Nanodrop, ThermoFisher Scientific, Waltham MA). Reverse transcription was performed using 500 ng RNA with SuperScript VILO cDNA synthesis kit (C#11754, Invitrogen, Carlsbad, CA), and the resulting products were snap-frozen and stored at -80°C.

Quantitative real-time PCR analysis was performed using the Corbett Rotorgene 3000 (Corbett Lifesciences, Sydney, Australia) with the FastStart DNA Master SYBR-Green 1 system (Roche Applied Sciences, Basel, Switzerland). Oligonucleotide sequences and conditions for PC5/6 and 18S amplicons are shown in Table 1. For quantification, a reference cDNA was prepared from a rabbit uterine tissue for quality control and standard. After 38 cycles, a melting curve analysis was performed to monitor PCR product purity. In initial experiments, amplification of a single PCR product was confirmed by agarose gel electrophoresis and DNA sequencing. Data was normalised to 18S and the mean from each triplicate samples from three animals were compared.

Immunohistochemistry was performed using the sheep PC5/6 antibody (n = 3 sites per animals, n = 3 animals for each time point) as previously described [8]. Briefly, formalin-fixed paraffin sections (5 μm) were de-waxed in histosol (Sigma), rehydrated, and subjected to antigen retrieval by microwave (5 min/700W followed by 5 min/140 W) in 0·01 mol/L sodium citrate buffer (pH 6.0). Endogenous peroxidase activity was quenched with 6% H₂O₂ (diluted 1:1 v/v in 100% methanol) for 10 min at room temperature. Nonspecific binding was blocked by incubation with 15% normal rabbit serum and 5% fetal calf serum in TBS containing 0.1% Tween-20. Primary antibody was applied (10 μg/mL) at 37 °C for 2 h. Following washing in TBS, slides were incubated with biotinylated rabbit anti-sheep IgG (1:200 dilution, Vector Laboratories, Burlingame, CA) for 30 min at room temperature. Signal was amplified using Vectastain ABC amplification (Vector Laboratories) for 30 min and positive localization visualized using the peroxidise substrate 3, 3'-diaminobenzidine (DAB, Dako, Kingsgrove, Australia) which produced a brown precipitate. Tissue sections were counterstained with Harris' hematoxylin (Sigma Chemicals, St Louis, MO), dehydrated, and mounted with DPX (ProSciTech, Thuringowa, Australia). Negative controls were included for each section, where pre-immune sheep IgG was substituted for the primary antibody at a matching concentration. The localization of staining was assessed by two investigators who noted expression patterns, for which representative images were recorded.

Primers specific to the rabbit PC5/6 mRNA (both A and B isoforms) coding region were designed and conditions for RT-PCR optimized (Table 1). Amplification of 18S was used for data normalization. A single band of expected size was amplified for PC5/6 (221 bp) and 18S (187 bp) from rabbit uterine RNA (Figure 1A). Melting curve analysis validated each band as a single DNA product (data not shown). No amplification occurred in the negative controls when either the reverse transcriptase or RNA was omitted (Figure 1A). The PC5/6 product was sequenced and found to be 100% homologous to the rabbit PC5/6 mRNA (Figure 1B), confirming specificity. This established that PC5/6 mRNA was expressed in the rabbit uterus, and that the newly designed primers were specific and suitable for analysing PC5/6 mRNA in the rabbit.

The expression of PC5/6 mRNA in the rabbit uterus during early pregnancy and pseudo-pregnancy was analysed by quantitative RT-PCR. The first 10 days of pregnancy including the period of embryo attachment (d6.5-7) and implantation (d8-10) were examined. For each real-time run, the cycle threshold was determined for replicates of each sample and compared to the linear standard curve, then normalised to 18S expression. The mean and SEM were then determined for each triplicate of animals and are presented in Figure 2. Pseudo-pregnant rabbits showed no significant changes in PC5/6 expression across the 10 days following GnRH administration. In contrast, the pregnant uterus displayed a dynamic pattern of PC5/6 expression during the same time-course (Figure 2). PC5/6 mRNA levels were relatively static during the initial 5 days (except a non-significant increase on d1), but a sustained trend of increased expression in pregnant tissues commenced from d6 (immediately prior to implantation, 1.9 fold greater than d0 level). On d6.5, corresponding to the time of initial embryo attachment, the mean level at the implantation and inter-implantation sites compared to d0 was 2.34 and 1.82 fold respectively. These increased levels were maintained throughout d6.5-8 at both the implantation or inter-implantation sites, and on d10 at the inter-implantation sites (Figure 2).

When pregnant and pseudo-pregnant animals were compared at the equivalent time points, no significant difference in PC5/6 mRNA levels was found during d0-5 (Figure 2). However, from d6 onwards the expression of PC5/6 was found to be higher in the pregnant uterus. On d6, the pregnant uterus was 2.5 fold of the pseudo-pregnant counterpart (Figure 2). On d6.5 (the putative time of embryo attachment), the fold differences were 2.9 and 2.1 for the implantation and inter-implantation sites compared to the pseudo-pregnant uterus, respectively. This trend was maintained between d6.5-8, with a peak difference between the pregnant and pseudo-pregnant uteri on d7 of 3.2- and 3.9-fold in the implantation and inter-implantation sites, respectively (Figure 2). On d10, PC5/6 levels in the inter-implantation sites were 3.0 fold greater than the pseudo-pregnant value (p < 0.001, Figure 2). This data reveals that PC5/6 expression is increased in the rabbit uterus at the time of embryo attachment and implantation, and that this dynamic regulation was specific to pregnancy.

The presence of PC5/6 protein in the rabbit uterus was established by Western blot analysis. Total proteins isolated from implantation and inter-implantation sites on d7-10 were probed with the PC5/6 antibody. The membrane was then re-probed for β-actin as a loading control (Figure 3). Two bands at approximately 66 kDa were detected for PC5/6 in all samples, consistent with previous findings in the mouse [8, 24]. A much higher level of both bands of PC5/6 protein was detected in d8 and d10 isolates than d7 (Figure 3) Furthermore, the expression of PC5/6 protein was more intense in the implantation than the inter-implantation sites at each time point (Figure 3). The prevalence of truncated forms of PC5/6 may represent an additional regulatory mechanism in the rabbit uterus. These data confirmed the specificity of the PC5/6 antibody in the rabbit uterus, and demonstrated that PC5/6 protein was increased during embryo implantation.

The cellular localization and expression pattern of PC5/6 protein in the rabbit uterus across early pregnancy and pseudo-pregnancy was then determined by immunohistochemistry using formalin-fixed tissues and the PC5/6 antibody previously validated by Western blot. In the pregnant uterus, between d0-1 of pregnancy (Figure 4A-C), endometrial stroma and luminal epithelium demonstrated low levels of immuno-reactive PC5/6. The basal regions of luminal epithelium folds were also PC5/6 positive (Figure 4B-C), and to a lesser extent, positivity was also observed in the endothelial cells (Figure 4B-C). Between d3-5 (Figure 4D-F), the endometrial cells proliferate extensively, leading to the formation of mucosal folds. During this period, little immuno-reactive PC5/6 was detected in the endometrial folds (Figure 4E). In contrast, the deep glands located in the crypts against the myometrium were clearly positive (Figure 4F). At the time of embryo attachment on day 6.5 (Figure 4G-I), PC5/6 positivity in the deep basal glands became more prominent (Figure 4I). At this point of uterine remodelling, the mucosal folds were further extended compared with the early stages and were positive for PC5/6 (Figure 4G-H).

PC5/6 levels and localization transformed significantly during the time of embryo attachment and implantation (Figure 5). Immediately after initial embryo attachment on d7 (Figure 5A-C), PC5/6 was differentially expressed between the attachment (antimesometrial side) and non-attachment regions (mesometrial side). Whilst the mesometrial folds maintained PC5/6 immuno-negativity (Figure 5C), the antimesometrial luminal epithelium closely apposed to the blastocyst, now fusing, becoming multinucleated and transforming to a structure termed symplasma, became PC5/6 immuno-positive (Figure 5B). This pattern of localization persisted with gestation, and the intensity of immuno-reactive PC5/6 increased progressively (Figure 5). By d8 (Figure 5D-F), PC5/6 staining was much stronger than on d7 in the symplasma (Figure 5E). Noticeably, the staining was polarized within the multinucleated symplasma, with a prominent accumulation of PC5/6 protein at the basal surface (Figure 5E). On d10, very strong PC5/6 staining was detected at the site of embryo implantation (Figure 5G-5H). Strong immuno-reactive PC5/6 was seen in the multinucleated luminal epithelium cells (Figure 5H), which now appeared to be evenly distributed, rather than polarised. In addition, strong PC5/6 staining was evident in the peri-vascular decidual cells (Figure 5H, arrow head and insert). The luminal epithelium and glandular epithelium away from the implantation site maintained PC5/6 negativity (Figure 5I), but positivity was evident in the stroma of this region.

The uteri of pseudo-pregnant rabbits (d0-d10) were also examined by immunohistochemistry. The morphological changes during pseudo-pregnant d0-5 resembled those of pregnant rabbits on the equivalent days (data not shown). Likewise, uterine PC5/6 expression and localization were similar between pregnant and pseudo-pregnant rabbits during this period (d0-5, data not shown). However, unlike the pregnant uterus, the pseudo-pregnant uterus showed no further uterine remodelling with increasing gestation after d5. While a significant increase in PC5/6 staining was detected specifically at implantation sites in the pregnant uterus, no such changes were observed in the epithelial cells during pseudo-pregnancy, and PC5/6 levels and localization in the pseudo-pregnant uterus on d8-10 were similar to that of d5 pregnant uterus. Example images of PC5/6 immunolocalization in d8 pseudo-pregnant uterus are shown in Figure 5J-L.

Collectively, this data establishes that PC5/6 protein is dynamically regulated during early pregnancy in the rabbit uterus. PC5/6 was up-regulated in the endometrial deep glands just prior to embryo attachment. PC5/6 localization then strongly intensified in the multinucleated luminal symplasma and decidual cells at the site of embryo attachment during the time of implantation. PC5/6 up-regulation was both pregnancy, and implantation-site specific.