Visualization of preimplantation uterine fluid absorption in mice using Alexa Fluor™ 488 Hydrazide

Abstract

Uterine fluid plays important roles in supporting early pregnancy events and its timely absorption is critical for embryo implantation. In mice, its volume is maximum on day 0.5 post-coitum (D0.5) and approaches minimum upon embryo attachment ~D4.0. Its secretion and absorption in ovariectomized rodents were shown to be promoted by estrogen and progesterone (P4), respectively. The temporal mechanisms in preimplantation uterine fluid absorption remain to be elucidated. We have established an approach using intraluminally injected Alexa Fluor™ 488 Hydrazide (AH) in preimplantation control (RhoA^f/f) and P4-deficient RhoA^f/fPgr^Cre/+ mice. In control mice, bulk entry (seen as smeared cellular staining) via uterine luminal epithelium (LE) decreases from D0.5 to D3.5. In P4-deficient RhoA^f/fPgr^Cre/+ mice, bulk entry on D0.5 and D3.5 is impaired. Exogenous P4 treatment on D1.5 and D2.5 increases bulk entry in D3.5 P4-deficient RhoA^f/fPgr^Cre/+ LE, while progesterone receptor (PR) antagonist RU486 treatment on D1.5 and D2.5 diminishes bulk entry in D3.5 control LE. The abundance of autofluorescent apical fine dots, presumptively endocytic vesicles to reflect endocytosis, in the LE cells is generally increased from D0.5 to D3.5 but its regulation by exogenous P4 or RU486 is not obvious under our experimental setting. In the glandular epithelium (GE), bulk entry is rarely observed and green cellular dots do not show any consistent differences among all the investigated conditions. This study demonstrates the dominant role of LE but not GE, the temporal mechanisms of bulk entry and endocytosis in the LE, and the inhibitory effects of P4-deficiency and RU486 on bulk entry in the LE in preimplantation uterine fluid absorption.

Visualization of the temporal mechanisms of uterine fluid absorption via bulk entry and endocytosis during early pregnancy provides novel insights into cellular and molecular mechanisms in establishing uterine receptivity for embryo implantation.

Graphical abstract

Introduction

It was first reported exactly 100 years ago that the volume of fluid in the uterine cavity/lumen (uterine fluid) is dynamic during estrous cycle in rats, with uterine fluid accumulation at proestrus and estrus [1]. The uterine fluid volume is also dynamic during early pregnancy, as it peaks within the first day post-ovulation/post-coitum and reaches minimum at the time of embryo implantation in rodents (reviewed in [2, 3]). In the physiological condition, the uterine fluid volume is the net result of secretion and absorption. The general observations in ovariectomized rodent models are that ovarian hormones estrogen (E2) induces uterine fluid secretion while P4 promotes uterine fluid absorption [4–10].

Uterine fluid plays important roles in supporting early pregnancy events. Key events in natural early pregnancy include fertilization, embryo development, and embryo transport in the Fallopian tube (human)/oviduct (mouse), as well as continuous embryo development, transport, and implantation in the uterus [2, 3, 11]. Preovulatory E2-induced uterine fluid secretion facilitates semen liquefaction to free the sperm and provides a conducive environment for sperm migration to the Fallopian tube/oviduct for fertilization [12, 13]. Postovulatory P4-induced uterine fluid absorption leads to uterine lumen closure, which enables intimate contact of the implanting embryo with the uterine luminal epithelium (LE) to initiate embryo implantation [2]. Uterine fluid, which is recently revealed in heifers to be metabolically semi-autonomous [14], also relays maternal signals to the preimplantation embryo, and vice versa, to facilitate embryo–maternal communications for embryo implantation [2, 9, 15]. The orchestration of early pregnancy events leading up to embryo implantation is critical for successful pregnancy [2, 16]. In assisted early pregnancy, such as in vitro fertilization-embryo transfer (IVF-ET) that does not involve other early pregnancy events in the female reproductive tract prior to embryo implantation, uterine fluid is still a critical factor for successful embryo implantation. For example, patients with excessive uterine fluid retention at the time of IVF-ET is correlated with reduced rate of embryo implantation, which was confirmed by rising serum β-HCG levels and the presence of gestational sacs [17]. Timely absorption of preimplantation uterine fluid is critical for embryo implantation in both humans and mice [9, 18–23].

Uterine fluid movement is accompanied with ion movement and endocytic activity. In ovariectomized rats, E2 treatment induces secretion of water, sodium ion (Na⁺, the dominant electrolyte in the uterine fluid), and potassium ion (K⁺) into the uterine lumen accompanied with an increased Na⁺/K⁺ ratio, while P4 treatment reabsorbs water, Na⁺, and K⁺, with a reduced Na⁺/K⁺ ratio [4, 5]. Since ligations of the uterine horn to block uterine fluid leakage through oviductal side and cervical side do not affect P4-induced uterine fluid absorption [5], uterine fluid absorption has to pass through the uterine epithelium, which borders the uterine lumen where the uterine fluid stays. Because of the limited paracellular flow in the uterine epithelium [24], transcelluar flow through the apical membrane of uterine epithelium is expected to be the essential passage for uterine fluid absorption. Since there is limited capacity for water passive diffusion through the lipid bilayer of the plasma membrane and osmotic gradient is in general a main driving force for water movement through the plasma membrane, the absorption of the dominant electrolyte Na⁺ in the uterine fluid will generate an osmotic gradient, which is expected to facilitate water absorption via bulk entry. Na⁺ is absorbed through sodium channels, and its entry via the amiloride-sensitive epithelial Na⁺ channel (ENaC, encoded by Scnn1a, Scnn1b, Scnn1g) is the rate-limiting step for Na⁺ absorption across several epithelial tissues [25]; therefore, ENaC is expected to be a key player to transport Na⁺ from the uterine fluid to the uterine epithelial cells. Indeed, a uterine perfusion study in E2 and P4-treated ovariectomized rats reveals that amiloride inhibits P4-induced uterine fluid absorption [8]. A study of electrogenic ion transport across the cultured primary mouse uterine epithelial cells reveals the dominant contribution of apically absorbed Na⁺ for the basal current, which can be inhibited by ENaC inhibitor amiloride [26]. These studies indicate the essential role of ENaC in apical Na⁺ transport into the uterine epithelial cells. To maintain intracellular homeostasis, the absorbed Na⁺ ions are pumped out of the epithelial cell into subepithelial space, a process involving the Na⁺/K⁺ ATPase located on the basolateral membrane [27, 28]. Na⁺/K⁺ ATPase is composed of a catalytic α subunit and an auxiliary β subunit. ATP1A1 is the dominant α subunit in the LE [29]. Water channels/aquaporins (AQPs) have their primary functions in facilitating bidirectional water movement across cell membranes in response to osmotic gradients, among other functions [30–32]. There are 13 AQPs (AQP0–AQP12) in three categories: classical aquaporins (AQP0, AQP1, AQP2, AQP4, AQP5, AQP6, and AQP8) that are considered primarily selective to water; aquaglyceroporins (AQP3, AQP7, AQP9, and AQP10) have larger pore sizes and also permeate glycerol, urea, and other small non-charged solutes; and nonorthodox aquaporins (AQP11 and AQP12) that are intracellular paralogs with functions still under investigation [33]. One study examined the expression of Aqp0–Aqp9 in mouse uterus and showed that only Aqp4 and Aqp5 were mainly detected in the preimplantation mouse uterine epithelium [34]. Another study demonstrated the upregulation of AQP5 and AQP8 by E2 in the preimplantation mouse glandular epithelium (GE) and AQP5 and AQP8 had in vivo function in uterine fluid secretion [9]. Embryo implantation requires a receptive uterus, and the establishment of uterine receptivity is associated with the appearance of pinopodes, the protrusions of the LE apical surface that have endocytic activity, and their appearance is P4-dependent [35–38]. These studies demonstrate that P4 signaling regulates both bulk entry and endocytosis of uterine fluid.

During early pregnancy, P4 is mainly synthesized from the corpus luteum, which is normally developed from the newly ovulated ovarian follicle [39]. Preimplantation serum P4 levels rise shortly after ovulation and copulation in mice [40–42] and the general consensus is that it reaches a plateau prior to embryo implantation [2, 11], which initiates ~day 4.0 post-coitum (D4.0) in mice [43]. Massive uterine fluid reduction occurs on D0.5 [2] when endocytosis of macromolecules from uterine lumen was reported undetectable [24]. There have been debates on cellular mechanisms of uterine fluid absorption and to this point, the roles of LE and GE in uterine fluid absorption have not been differentiated. We hypothesized that bulk entry and endocytosis are two mechanisms that temporally mediate P4 regulation of preimplantation uterine fluid absorption, and uterine epithelium LE and GE may play different roles in uterine fluid absorption. This hypothesis was tested in control and our RhoA^f/fPgr^Cre/+ mouse model with P4 deficiency [44], coupled with exogenous P4 or RU486 treatments. We have developed a novel method using Alexa Fluor™ 488 Hydrazide (AH) to visualize uterine fluid absorption during early pregnancy in mice.

Materials and methods

Mice and treatments

CD-1 mice were purchased from Charles River (8–10 weeks old) for pilot experiments to evaluate various parameters for establishing the AH injection and detection procedures, including intraluminal injection, AH concentration and volume, post AH injection dissection timing, detection of AH fluorescence, and selection of early pregnancy time points. The established AH injection and detection procedures were subsequently used in RhoA^f/f (control) and RhoA^f/fPgr^Cre/+ female mice, which have P4 deficiency and are infertile, via mating RhoA^f/f females with RhoA^f/fPgr^Cre/+ males that we generated and genotyped as described previously [44]. Virgin adult RhoA^f/f (control) and RhoA^f/fPgr^Cre/+ female mice (~6 months old) were mated with wild-type stud males and checked for the presence of a vaginal plug every morning. The day of plug identification is defined as D0.5 (note: D0.5 is also designated as D1 or occasionally D0 in the literature; when the related papers with different designations are cited, the dating of early pregnancy is converted to our dating system). Same genotype littermates were assigned into different time points/treatment groups. There were seven groups of mice: D0.5 RhoA^f/f, D0.5 RhoA^f/fPgr^Cre/+; D3.5 RhoA^f/f oil, D3.5 RhoA^f/f P4, D3.5 RhoA^f/f RU486; D3.5 RhoA^f/fPgr^Cre/+ oil, and D3.5 RhoA^f/fPgr^Cre/+ P4. N = 3–5/group. Treatments were a single subcutaneous (s.c.) injection of the same chemical each on D1.5 and D2.5 between 11 and 12 h with one of the following: vehicle control (0.1 ml sesame oil, Sigma, S3547-1 L), P4 (2 mg/0.1 ml/mouse, Sigma, P0130-25G), or RU486 (200 μg/0.1 ml/mouse, Fisher, B1511100). All mice were maintained on Labdiet mouse chow 5053. They were housed in polypropylene cages with free access to food and water on a 12 h light/dark cycle (06:00–18:00) at 23 ± 1°C with 30–50% relative humidity. All methods used in this study were approved by the University of Georgia Institutional Animal Care and Use Committee (IACUC) Committee and conform to National Institutes of Health guidelines and federal law.

Alexa hydrazide injection and tissue dissection

Alexa Fluor™ 488 Hydrazide (AH, MW: 570.48) (ThermoFisher, A10436) was dissolved in 1x PBS at 80 μg/ml, which was then 1:1 mixed with 1% Evans blue dye (MP Biomedicals, 151108) solution for tracing AH injection into the uterine lumen, to make the AH working solution at 40 μg/ml for intrauterine luminal injection. Alexa hydrazide injection: At ~11 h on D2.5 (for the pilot experiment), and D0.5 and D3.5, the assigned mice were anesthetized with isoflurane (Patterson Veterinary Supply, 07-893-1389) inhalation for surgery. The uterine horns were located via a small dorsal incision on each side. Alexa hydrazide (~6.5 μl for D0.5 uterine horns and ~4.5 μl for D3.5 uterine horns) was injected into each uterine horn via the uterotubal junction. The blue color would quickly spread to the cervical side of the uterine horn. The left and right uterine horns were dissected at 5 min (min) and 1 h, respectively, after AH injection. Approximately ~2/3 of the uterotubal side of the uterine horn was processed for detecting AH fluorescence and the rest ~1/3 of the cervical side of the uterine horn was flash-frozen.

Detection of uterine fluid absorption via AH fluorescence

The uterotubal side ~2/3 uterine segment was fixed in 4% paraformaldehyde in 1xPBS (pH 7.2) at 4°C for 2 h, rinsed in 1xPBS twice, and then immersed in 25% sucrose in 1xPBS (pH 7.2) at 4°C for overnight. In the following day, the uterine tissue was embedded in Tissue-Tek O.C.T for cross sections (10 μm) in the dark. Sections were counterstained and mounted in 4′,6′-diamino-2-phenylindole (DAPI)-containing Vectashield (Vector Laboratories). Images of AH fluorescence (green) and DAPI staining (blue) were captured.

Detection of uterine autofluorescence

D0.5 and D3.5 mice were processed exactly as for AH injection except that the mice were intraluminally injected with the same volume of vehicle (1xPBS with Evans blue dye). The uterine horns were dissected and processed similarly as those from AH injection.

Immunofluorescence

E-Cadherin (E-Cad, Cell Signaling, 24E10), lysosomal-associated membrane protein 2 (LAMP2, DSHB, ABL-93) and alpha 1 sodium-potassium ATPase (ATP1A1, Abcam, ab7671) were detected via immunofluorescence as previously described [44]. Briefly, frozen sections were fixed in 4% paraformaldehyde for 20 min and then subjected to antigen retrieval in 0.01 M sodium citrate (pH 6.0) at 95°C for 20 min. Sections were then washed with 1x PBS and permeabilized with 0.15% Triton X-100. The slides were washed and blocked with 10% goat serum (ThermoFisher, 16210064) for 1 h at room temperature; they were subsequently incubated with anti-E-Cad (1:1000), or anti-LAMP2 (1:80), or anti-ATP1A1 (1:1000) in a humid chamber at 4°C for overnight. On the following day, slides were first washed in 1x PBS and then incubated with a secondary antibody (ThermoFisher, A11034 (1:200) for detecting E-Cad, A11007 (1:1000) for detecting LAMP2, and A11031 (1:200) for detecting ATP1A1). The sections were counterstained and mounted with DAPI for examination. Negative controls were processed together without the primary antibody. Uterine sections from different groups of mice were placed on the same slides for processing to detect ATP1A1. Images were captured at the same exposure setting using a Carl Zeiss imaging system with an AxioCam MRc5 digital camera. The presented ATP1A1 images had a resolution of 300 pixels/inch and were only edited for sizes but none of the other parameters, such as exposure levels or contrasts.

Semi-quantification of data

Bulk absorption and endocytosis were semi-quantified by two authors independently. The average of each parameter from each mouse was counted as one data point for statistical analyses. Semi-quantification of bulk absorption was based on the estimation of the percentage of LE segment(s) with discernable smeared cellular AH fluorescence relative to the surrounding area: 5 (>90%, intense), 4 (>50%, intense), 3 (>20%, intense), 2 (>5%, intense), 1 (scattered, intense), 0.5 (scattered, not intense), and 0 (nondiscernable). The abundance of autofluorescent apical fine dots, which were presumptively endocytic vesicles to reflect endocytosis, in the LE cells was semi-quantified based on the estimation of the percentage of LE cells with fine autofluorescent dots forming a string of apical punctae: 3 (>90%), 2 (>50), 1 (>5%), 0.5 (scattered), and 0 (nondiscernable). Bulk absorption in GE was rare, therefore, was not quantified. Since we did not observe any consistent difference in the presence of green dots in the GEs among different groups, endocytosis in GE was not quantified.

Statistical analysis

The semi-quantitative data are presented as mean ± SD. Paired two-tailed Student t-test was used for comparisons of parameters between 5 min and 1 h of the same set of samples. Ranking coupled with two-tailed unequal variance Student t-test was used for other comparisons between two genotypes or between two time points or two treatments within the same genotype. Significance level was set at P < 0.05.

Results

Establishing intraluminal AH injection and AH fluorescence detection procedure

Alexa hydrazide was used in organ culture of intestinal tissues at 20 μg/ml [45]. We made the AH working solution at 40 μg/ml with blue dye to track the spreading of AH in the uterine lumen. The final AH concentrations in the uterine lumen are expected to vary greatly because there is a big difference in the uterine fluid volume during early pregnancy in mice, e.g. from up to 100 μl/uterine horn on D0.5 (Figure 1A, large uterine lumen with extensive LE and LE folding upon drainage of uterine fluid) to possibly a few microliters or even less approaching embryo implantation (Figure 1B, narrow uterine lumen). We adjusted the injected AH volumes to ~6.5 μl in D0.5 uterine horns and ~ 4.5 μl in D2.5 and D3.5 uterine horns and ensured that the spreading of the injected blue dye mixed with AH dye to the cervical side could be visualized. Although the final AH concentrations in the injected uterine horns varied depending on the uterine fluid volume in each uterine horn, the injection regimen was suitable for detecting AH signals in the uterine epithelium.

Figure 1.

E-Cad immunofluorescence (A, B), establishment of uterine AH injection and fluorescence detection procedures (C–E), and outline of experimental design (F–G). (A, B) E-Cad immunofluorescence in D0.5 (A) and D3.5 (B) mouse uteri to show the dramatic reduction of uterine LE, thus uterine lumen and uterine fluid volume, from D0.5 to D3.5. (C–E) D2.5 CD-1 mouse uterine horns were injected with AH and dissected 5 min to 24 h after AH injection. (C) 5 min. Scattered LE cells (red arrow) with intense smeared AH fluorescence in the whole cell; minimal fluorescent dots (dotted yellow arrow) in apical side of some LE cells. (D) 1 h and (E) 24 h. No LE cell with intense smeared AH fluorescence in the whole cell; AH fluorescence dots on apical LE. LE, uterine luminal epithelium; GE, glandular epithelium; Str, stroma; Lu, uterine lumen; red arrow, LE cells with intense smeared AH fluorescence; yellow arrow, clustered AH fluorescence dots on apical LE; dotted yellow arrow, scattered AH fluorescence dots on apical LE; scale bar, 25 μm. (F) D0.5. After the identification of a vaginal plug to confirm mating, the mated/plugged RhoA^f/f (control) and RhoA^f/fPgr^Cre/+ females were undergoing surgery for AH injection into both uterine horns at ~11 h. The left and right uterine horns were dissected 5 min and 1 h after AH injection, respectively. (G) D3.5. The mated/plugged females received one of the treatments (oil, P4, and RU486 (for control mice only)) on both D1.5 and D2.5, with AH injection on D3.5, with the five groups summarized in the table below. AH injection and tissue collection were the same as in D0.5 (F).

To track the time course of AH fluorescence in the uterine LE, we chose a middle early pregnancy day, D2.5, and injected AH fluorescent dye into the uterine horns ~11 h on D2.5. The mice were dissected at different intervals from 5 min to 24 h later. To preserve the AH fluorescence for detection, we first fixed the dissected uterine tissues in 4% paraformaldehyde at 4°C for 2 h, then two uterine segments from the same uterine horn were processed with or without sucrose immersion for overnight. Although AH green fluorescence was detectable in the LE before and after sucrose treatment, the fine dots (endocytic vesicles) on the apical LE cells could only be distinguished in the sections after sucrose treatment. However, some bright dots in the GE were clearly distinguishable with or without sucrose treatment (data not shown). Therefore, the overnight sucrose treatment is necessary to preserve the fine endocytic vesicles with AH fluorescence, especially in the LE (Figure 1C–E).

The time course detection of green fluorescence, presumably AH fluorescence, revealed that although the patterns of green fluorescence in different LE regions of the same D2.5 uterine sections varied, the general trend was that LE cells with intense smeared cellular green fluorescence decreased with time, while distinctive green fine cellular dots in the apical side of LE cells generally increased with time from 5 min to 24 h post AH injection. Images from three time points are shown in Figure 1C–E. At 5 min post AH injection, a few LE cells had bright smeared cellular green fluorescence in the whole cell but LE cells with distinctive apical green fine dots were minimal and the number of such dots was also minimal (Figure 1C). At 1 h (Figure 1D) and 24 h (Figure 1E) post AH injection, no LE cells with bright smeared cellular green fluorescence were observed, most LE cells had varied numbers of distinctive green fine dots mainly in their apical side and much less frequently in the basolateral side, and more such fine dots formed a string of punctae in the apical side of LE at 24 h post AH injection (equivalent to D3.5). This survey reveals that when a D2.5 uterus receives exogenous liquid in the uterine lumen, bulk absorption (seen as bright smeared cellular green fluorescence) diminished with time, while endocytosis (presumably reflected by the distinctive cellular green fine dots) generally increased with time during early pregnancy.

Since the transition of the patterns of green fluorescence (smeared versus fine dots), reflecting different uterine fluid absorption mechanisms was already detectable from 5 min to 1 h after AH injection, we decided on two time points for detecting AH fluorescence: 5 min and 1 h after AH injection, with one uterine horn for each time point so that the time course effects could be evaluated in the same mouse. We also decided to choose two preimplantation time points: D0.5 when the uterine lumen is filled with uterine fluid (Figure 1A) and D3.5 when the uterine fluid is approaching the minimum to prepare for embryo implantation (Figure 1B) [2], for visualizing temporal mechanisms in preimplantation uterine fluid absorption as well as the function of P4 signaling in preimplantation uterine fluid absorption (Figure 1F and G). P4-deficient RhoA^f/fPgr^Cre/+ mice [44], exogenous P4-injected RhoA^f/fPgr^Cre/+ mice, and PR antagonist RU486-treated RhoA^f/f control mice were employed to investigate P4-PR signaling in the uterine fluid absorption during early pregnancy (Figure 1G).

Detection of AH fluorescence in D0.5 RhoA^f/f (control) and RhoA^f/fPgr^Cre/+ uteri

In the D0.5 control uterus, the uterine lumen is filled with uterine fluid, which is drained during dissection and consequently, the LE layer becomes folded because it does not contract as the outer myometrial layer does (Figure 1A). At 5 min post AH injection, the LE layer was highlighted by green AH fluorescence under a lower magnification (Figure 2A); and under a higher magnification, most LE cells were seen to have bright smeared AH fluorescence in the entire cell (Figure 2A1). However, such bright smeared AH fluorescence in the LE cells was not seen in the GE cells, which had either no visible fluorescence or various sizes and numbers of distinct green dots that could be seen in both apical side and basal side (Figure 2A1), but overall, more appeared to be on the apical side. At 1 h post AH injection, LE cells with intense green fluorescence decreased. There was an individual variation among different mice, two mice had no LE cells with obvious smeared green fluorescence while three mice still had a few segments of the LE layer with intense smeared green fluorescence (Figure 2B). A zoom-in image of an area with transition green fluorescence indicated that the middle part of the LE cells started to show less intensity (Figure 2B1), suggesting that the nucleus had less green fluorescence than that in the cytoplasm. In addition, a few green dots appeared in the apical site of the LE cells that had less intense green fluorescence (Figure 2B1). There was a trend of reduced LE cells with intense smeared green fluorescence from 5 min to 1 h post AH injection in both D0.5 (Figures 2A, B, and4A) and D2.5 uteri (Figure 1C and D). We have noticed that LE cells with intense smeared green fluorescence (bulk absorption) usually do not have noticeable green fine dots (endocytosis) in the apical side, and LE cells with multiple green fine dots in the apical side usually do not have intense smeared green fluorescence. The coexistence of green fine dots in the apical side of LE cells and intense smeared green fluorescence is rare and has only been observed in a few LE cells in two uteri of all seven groups in this study (data not shown). It is possible that the green fine dots are masked by intense smeared green fluorescence.

Figure 2.

AH fluorescence in D0.5 uteri, detection of autofluorescence in D0.5 and D3.5 uteri and LAMP2 in D3.5 uterus. (A-E3) From AH-injected D0.5 uteri. (A) D0.5 RhoA^f/f (control) 5 min. (B) D0.5 RhoA^f/f (control) 1 h. (C) D0.5 RhoA^f/fPgr^Cre/+ 5 min. (D) D0.5 RhoA^f/fPgr^Cre/+ 1 h. (A1–D1) Enlarged from the boxed area in A–D, respectively. (E) D0.5 RhoA^f/fPgr^Cre/+ 1 h to show sperm (white arrows) in the uterine lumen (E1) and a rare observation of bulk entry in some GE (E2). (E3) DAPI staining of the same area in E2. White star in (E), most likely AH-stained semen. (F–K1) From vehicle-injected D0.5 RhoA^f/f control uteri. (F) 5 min post vehicle injection. (G, H) Enlarged from the boxed areas in (F). (I) 1 h post vehicle injection. (J and K) Enlarged from the boxed areas in (I). (L–Q1) From a vehicle-injected D3.5 RhoA^f/fPgr^Cre/+ uterus. (L) 5 min post vehicle injection. (M and N) Enlarged from the boxed areas in (L). (O) 1 h post vehicle injection. (P and Q) Enlarged from the boxed areas in (O). (F1–Q1) DAPI staining of images in (F–Q), respectively. (A–Q) LE, uterine luminal epithelium; GE, glandular epithelium; Str, stroma; Lu, uterine lumen; red arrow, LE cells with intense smeared AH fluorescence; yellow arrow, clustered AH fluorescence dots on apical LE; dotted yellow arrow, scattered AH fluorescence dots on apical LE; black arrow, most likely infiltrated immune cell; white circle, uterine gland. (R–U) Detection of LAMP2 in a D3.5 control uterus. (R) Negative control; (S) LAMP2; pink arrow, LAMP2 localization in LE as apical punctae; pink ^*, because of sectioning angle, single layer LE appeared to be multiple layers. (T) DAPI. (U) Merged image of LAMP2 and DAPI. Scale bar, 200 μm in A–F, F1, I, I1, L, L1, O, O1, and 25 μm in the rest images.

Figure 4.

Semi-quantification of AH fluorescence indicative of bulk absorption and LE apical green fine dots indicative of endocytosis in LE. (A) Bulk absorption in LE. Based on the rough estimation of the percentage of LE segment(s) with discernable smeared cellular AH fluorescence: 5 (>90%, intense), 4 (>50%, intense), 3 (>20%, intense), 2 (>5%, intense), 1 (scattered, intense), 0.5 (scattered, not intense), and 0. (B) LE apical green dots indicative of endocytosis in LE. Based on the estimation of the percentage of LE cells with distinctive cellular AH fluorescent dots forming a string of apical punctae: 3 (>90%), 2 (>50), 1 (>5%), 0.5 (scattered), and 0 (nondiscernable). Error bar, standard deviation; ^*  P < 0.05, 1 h versus 5 min, paired two-tailed Student t-test; and P < 0.05, D3.5 oil versus D0.5; # P < 0.05, D3.5 P4/RU486 versus D3.5 oil; $ P < 0.05, RhoA^f/fPgr^Cre/+ versus RhoA^f/f for the same parameter; &, #, $, ranking coupled with two-tailed unequal variance Student t-test; N = 3–5.

In the D0.5 RhoA^f/fPgr^Cre/+ uterus, the uterine lumen was also filled with uterine fluid. At 5 min post AH injection, the green fluorescence in the LE layer (Figure 2C) did not have the overall intensity seen in the control LE layer (Figure 2A) under a lower magnification. A few LE cells were brighter with smeared green fluorescence than the rest of the LE cells (Figure 2C1). Limited green fine dots were visible in the apical site of some LE cells (data not shown). At 1 h post AH injection, most uterine sections did not show LE cells with bright smeared AH fluorescence (Figure 2D) except sections from one uterus that had a few scattered LE cells with less intense smeared green fluorescence (data not shown). A few green dots were present in the apical side of a few LE cells (Figure 2D1), and some could be seen throughout a few LE cells (Figure 2E2). These observations reveal that bulk uterine fluid absorption through the LE was reduced in the D0.5 RhoA^f/fPgr^Cre/+ uterus but the presence of green fine dots in the apical side of LE appeared to be comparable to those in the control (Figures 2A–E and 4). Interestingly, we observed a rare phenomenon that a few GE cells had bright smeared AH fluorescence (Figure 2E2 and E3), while the LE cells in the same section only had green dots (Figure 2E1–E3). This phenomenon was observed in only one D0.5 RhoA^f/fPgr^Cre/+ uterus among all seven groups of mice in this study. Since we only examined two sections per uterine horn, it is highly possible that we missed detection of this phenomenon in other uterine areas. Regardless, it is reasonable to conclude that bulk absorption in the GE is not common and is not a major contributing factor to uterine fluid absorption.

In addition to the intense smeared green fluorescence and green fine dots in the uterine epithelium described above, we occasionally observed scattered large and bright dots, which could be infiltrated immune cells, in the LE layer (data not shown); however, we consistently observed green fluorescence, most as clusters of dots with various sizes, in the stromal layer and less in the myometrial layer of all the uterine sections examined and it did not appear to correlate with AH entry into the uterine epithelium, especially bulk entry into the LE. It has been reported that cellular components, such as NADPH and avins, and intracellular organelles including mitochondria and lysosomes, in mammalian cells are autofluorescent (e.g. at 488 nm, the same wavelength for detecting AH fluorescence), and different cells (e.g. immune cells) have different autofluorescence [46–49]. Autofluorescence in preimplantation uteri was examined below.

Detection of uterine autofluorescence

In uterine sections from two intraluminally vehicle-injected D0.5 RhoA^f/f control mice, we detected similar green fluorescence in the stroma, GE, and myometrium as seen in the AH-injected D0.5 uterine sections (Figure 2F–K1), indicating autofluorescence in these uterine compartments. On the other hand, none of the LE cells in the vehicle-injected D0.5 control uterine sections had intense smeared green fluorescence (Figure 2F–K1), confirming that the intense smeared green fluorescence was indeed AH fluorescence (Figure 2A–C1). However, we also detected autofluorescence in the LE as fine dots that were similar as seen in AH-injected uterine sections (Figure 2G and J, versus Figure 2B1 and E2). We also detected scattered large and bright dots (most likely immune cells) in the LE layer as seen in the AH-injected uterus (data not shown, but similar as in Figure 2M and P). We also examined two D3.5 RhoA^f/fPgr^Cre/+ uteri and found similar pattern of autofluorescence as seen in D0.5 control uterine sections (Figure 2F–K1) except that there appeared to be more green dots on the LE apical side (Figure 2L–Q1), and 1 h after vehicle injection, a string of apical punctae could be observed in segments of LE layer (Figure 2P), which were also seen in the AH-injected D3.5 RhoA^f/fPgr^Cre/+ uterine sections (Figure 3). In addition, scattered large and bright dots, which were most likely immune cells, were detected in the LE layer (Figure 2M and P). These data demonstrate extensive autofluorescence in the preimplantation uterine compartments and bulk absorption could be specifically detected by AH dye. It was reported that mitochondria and lysosomes can produce autofluorescence [46], intraluminal markers taken up by endocytosis were channeled into lysosomes in ~D4.5 rat LE [50], and it was apical endocytosis but not apocrine secretion for the accumulation of vesicles accumulated in the apical side of D4.5 rat LE (embryo penetration through LE occurs ~D4 20 h in mice and ~ D5 9 h in rats [2]) [51]. We detected LAMP2, a marker of lysosome-associated membranes, as apical punctae in the D3.5 control LE (Figure 2R–U). These observations support that the autofluorescent green fine dots in the apical LE cells are an indication of apical endocytosis in the LE cells.

Figure 3.

AH fluorescence in D3.5 uterus. (A) Oil (vehicle)-treated D3.5 RhoA^f/f, 5 min. (B) Oil-treated D3.5 RhoA^f/f, 1 h. (C) Oil-treated D3.5 RhoA^f/fPgr^Cre/+, 5 min. (D) Oil-treated D3.5 RhoA^f/fPgr^Cre/+, 1 h. (E) P4-treated D3.5 RhoA^f/f, 5 min. (F) P4-treated D3.5 RhoA^f/f, 1 h. (G) P4-treated D3.5 RhoA^f/fPgr^Cre/+, 5 min. (H) P4-treated D3.5 RhoA^f/fPgr^Cre/+, 1 h. (I) RU486-treated D3.5 RhoA^f/f, 5 min. (J) RU486-treated D3.5 RhoA^f/f, 1 h. (A1–J1) Enlarged from the boxed area in (A–J), respectively. (K) Another RU486-treated D3.5 RhoA^f/f, 1 h to show the varied AH fluorescent dots in a cluster of uterine glands (K1). (K1 and K2) Enlarged from the boxed areas in (K). (K3) DAPI staining covering the areas in (K1) and (K2). LE, uterine luminal epithelium; Str, stroma; Lu, uterine lumen; red arrow, LE cells with intense smeared AH fluorescence; yellow arrow, clustered AH fluorescence dots on apical LE; dotted yellow arrow, scattered AH fluorescence dots on apical LE; black arrow, most likely infiltrated immune cell; white circle, uterine gland; scale bar, 200 μm in A–K, 50 μm in K3, and 25 μm in A1–K2.

Detection of AH fluorescence in D3.5 RhoA^f/f (control) and RhoA^f/fPgr^Cre/+ uteri

Unlike D0.5 mouse uterus that is normally filled with uterine fluid (Figure 1A), D3.5 mouse uterus normally has minimal uterine fluid and a narrow uterine lumen (Figure 1B); therefore, the injected AH solution may slightly enlarge the uterine lumen in certain areas. We used oil-injected mice as the vehicle control for the P4-treated and RU486-treated mice in the set of mice on D3.5, which were injected with oil on D1.5 and D2.5 (Figure 1G). In the D3.5 RhoA^f/f (control) mice, there were patches of and/or scattered LE cells with bright smeared AH fluorescence (Figure 3A and A1) 5 min post AH injection. At 1 h post AH injection, LE cells with bright smeared AH fluorescence were decreased and most LE cells had AH fluorescent dots in the apical side (Figure 3B and B1). Glandular epithelium had green fluorescent dots in various sizes and numbers (data not shown).

In the D3.5 RhoA^f/fPgr^Cre/+ uterus (Figure 3C–D1), the LE cells with bright smeared AH fluorescence were rarely seen at both 5 min and 1 h post AH injection. Among all sections from five mice in this group, only a few scattered LE cells in the uterine sections from three mice at 5 min and one mouse at 1 h post AH injection showed identifiable but not intense smeared AH fluorescence (data not shown). These observations indicate less activity of bulk AH absorption in the LE from the D3.5 RhoA^f/fPgr^Cre/+ uterus compared to that in the D3.5 control uterus at the examined time points. However, the LE apical green fine dots in different mice varied and there was no significant difference in the overall level between this group and the control group (Figure 4B).

Effects of P4 on AH fluorescence in D3.5 RhoA^f/f (control) and RhoA^f/fPgr^Cre/+ uteri

Since D3.5 RhoA^f/fPgr^Cre/+ mice have P4 deficiency [44], we gave exogeneous P4 injections to both RhoA^f/f (control) and RhoA^f/fPgr^Cre/+ mice on D1.5 and D2.5 and dissected them on D3.5 (Figure 1G). In the P4-treated D3.5 control mouse uterus, the AH fluorescence detected as bright smeared cellular signal or dot signal at 5 min (Figure 3E and E1) and 1 h (Figure 3F and F1) post AH injection was generally comparable to that seen in the oil-injected D3.5 control mouse uterus (Figure 3A–B1). However, compared to oil-injected D3.5 RhoA^f/fPgr^Cre/+ mice, P4-treated D3.5 RhoA^f/fPgr^Cre/+ mice had a narrower uterine lumen and patches of LE cells with bright smeared AH fluorescence in uterine sections from all five mice in this group at 5 min post AH injection and two mice at 1 h post AH injection (Figure 3G–H1 and data not shown). The appearances of uterine lumen and the patterns of green fluorescence signals in the LE are generally similar to those in the RhoA^f/f control mice (Figures 3A–B1, E–F1, and 4B) except an increased percentage of LE cells with bright smeared AH fluorescence 5 min post AH injection (Figure 4A). These results demonstrate that exogenous P4 facilitates uterine fluid absorption in the P4-deficient D3.5 RhoA^f/fPgr^Cre/+ mice, indicated by the narrower uterine lumen and enhanced bulk absorption. P4 treatment did not seem to have any obvious effect on green fluorescence in the GE cells (data not shown).

Effects of RU486 on AH fluorescence in D3.5 RhoA^f/f (control) uterus

Since P4 mainly acts through progesterone receptor (PR) in the uterus, we also treated RhoA^f/f control mice with PR antagonist RU486 on D1.5 and D2.5 as a pharmacological loss of function model (Figure 1G). RU486-treated RhoA^f/f control mice have an enlarged uterine lumen, an indication of uterine fluid retention (Figure 3I–K). The LE cells lack bright smeared AH fluorescence (Figure 3I1, J1, and K2), a similar phenotype as seen in oil-treated D3.5 RhoA^f/fPgr^Cre/+ mice (Figures 3C1, D1, and 4A). The GE had various numbers of green fine dots, which were on the apical side of some GEs (Figure 3I1–J1) but not other GEs (Figure 3K1).

General observations of endocytosis in LE and GE

Overall, there is increased endocytosis, reflected by a string of apical punctae in the LE from 5 min to 1 h post AH injection on D3.5 and from D0.5 to D3.5 LE (Figure 4B). Exogenous P4 and RU486 do not seem to affect this trend (Figure 4B). Among the uterine glands in all uterine sections examined, the numbers and sizes of green dots were varied in different GE cells of the same uterine gland, in different uterine glands of the same uterine sections, and among different uterine samples in the same group. Unlike the green fine dots mainly detected in the apical LE, the green dots in the GE cells were often detected in the basal side also, although overall more seemed to be in the apical side. There was a lack of consistent difference in the green dots, in the GE cells among all seven groups, suggesting that endocytosis in the GE is not sensitive to the changes in early pregnancy environment and P4-PR signaling. Therefore, the contribution of GE in preimplantation uterine fluid absorption is minimal.

Expression of ATP1A1 in preimplantation uterus

Because of the expected role of ENaC in Na⁺ absorption from the uterine lumen to generate an osmotic gradient for uterine fluid absorption [4, 5, 8], and the expected role of aquaporins to facilitate water absorption in response to the osmotic gradient generated by Na⁺ absorption, we attempted to immunodetect ENaCα, ENaCβ, and ENaCγ, the three subunits for ENaC, as well as AQP4 and AQP5, the two aquaporins that were convincingly detected using in situ hybridization in the preimplantation mouse uterine epithelium [34], in the uterine tissues from all seven groups to correlate their localizations in the uterine epithelium with uterine fluid volumes and uterine fluid absorption. After testing multiple commercially available antibodies for ENaC subunits, only one anti-ENaCγ antibody gave seemingly specific signals, which showed largely similar uterine expression patterns to those revealed by anti-AQP4 and anti-AQP5 antibodies in serial uterine sections from the seven groups (data not shown). However, based on the limited reliable information of uterine expression and regulation of these three genes (mRNA/protein) in the literature and our unpublished mRNA profiling data, ENaCγ, AQP4, and AQP5 are expected to have different spatiotemporal expression patterns in the preimplantation mouse uterus. We are not confident on the immunohistochemistry data of ENaCγ, AQP4, and AQP5, and only present the immunofluorescence data of ATP1A1.

Our microarray data revealed ATP1A1 as the dominant α subunit of Na⁺/K⁺ ATPase in the LE (GSE44451) [29]. Na⁺/K⁺ ATPase localized on the basolateral membrane pumps apically absorbed Na⁺ ions out of the epithelial cell into subepithelial space to maintain intracellular Na⁺ homeostasis [27, 28]. ATP1A1 was mainly detected in the basolateral membrane of LE and GE (Figure 5). There seemed to be no obvious difference in the overall ATP1A1 staining between control and RhoA^f/fPgr^Cre/+ LE and GE on D0.5 (Figure 5A–B2). In the D3.5 control uterus, P4 treatment did not have an obvious effect on ATP1A1 staining in the LE and GE (Figure 5C–C2,5E–E2). However, RU486 treatment reduced ATP1A1 staining, mainly in the LE, especially the lateral membrane, but not obvious in the GE, of the D3.5 control uterus (Figure 5G–G2). In the D3.5 RhoA^f/fPgr^Cre/+ uterus, P4 treatment increased ATP1A1 expression in the LE but not in the GE (Figure 5D–D2,5F–F2), which was consistent with increased bulk absorption in the LE upon P4 treatment (Figure 3G and G1) and with a report of upregulation of ATP1A1 in the LE of P4-treated ovariectomized rats [52].

Figure 5.

Detection of ATP1A1 using immunofluorescence. (A) D0.5 RhoA^f/f (control). (B) D0.5 RhoA^f/fPgr^Cre/+. (C) Oil (vehicle)-treated D3.5 RhoA^f/f. (D) Oil-treated D3.5 RhoA^f/fPgr^Cre/+. (E) P4-treated D3.5 RhoA^f/f. (F) P4-treated D3.5 RhoA^f/fPgr^Cre/+. (G) RU486-treated D3.5 D3.5 RhoA^f/f. (H) Negative control, RU486-treated D3.5 RhoA^f/f. (I) DAPI staining of the section in (H). (A1–I1) with focus on LE and (A2–I2) with focus on GE, enlarged from the boxed area in (A–I), respectively. LE, uterine luminal epithelium; GE, glandular epithelium; Str, stroma; scale bar, 200 μm in A–I, and 25 μm in A1–I2.

Discussion

We initially intended to use Alexa Fluor™ 488 Hydrazide (AH) for tracing uterine fluid absorption via endocytosis during early pregnancy. However, when we tested AH in a D0.5 mouse uterus, we were surprised to see a bright LE layer with smeared green fluorescence. With more pilot experiments, we concluded that it was not an artifact. Considering AH being a small-sized and water-soluble molecule and the osmotic gradient generated by the absorption of Na⁺, the dominant electrolyte in the uterine fluid promoted by P4 treatment [4, 5], it is reasonable to speculate that AH can enter LE cells from the uterine lumen with water through the osmotic gradient. Therefore, AH can be visualized as smeared green fluorescence in the cytoplasm, which is in contrast to AH entry via endocytosis expected to be seen as green fine dots. Although the molecular mechanisms for bulk entry of AH into the LE cells remain to be investigated, mechanisms for water movement through mammalian plasma membrane driven by osmotic gradients could involve: (1) water diffusion through the lipid bilayer with limited capacity, (2) water passage through AQPs, and (3) passive water transport through cotransporters. In addition, some cotransporters could also conduct active water transport along with their transported solutes independent of the osmotic gradient [53–55].

Since the first report of dynamic uterine fluid volume during estrous cycle in rats [1], seminal studies have established the dominant roles of E2 in promoting uterine fluid accumulation and P4 in promoting uterine fluid absorption [4–10]. There were controversial reports about how uterine fluid disappeared upon P4 treatment, via leakage through the cervix or absorption through the uterine epithelium (reviewed in [5]). One well-controlled study firmly established uterine epithelium as the surface for uterine fluid absorption, in which the ligated (at both the ovarian and cervical ends to prevent uterine fluid leakage) and nonligated uterine horns in the same ovariectomized rats had similar reduction of uterine fluid upon P4 treatment [5].

Uterine epithelium includes LE and GE, which extends from LE into the underlying stromal layer. Because of the secretory nature of a gland, it is generally expected that GE may have a preferred role in uterine fluid secretion and LE may play a more important role in uterine fluid absorption. Studies from cultured mouse uterine epithelium (most likely dominated by LE based on the isolation procedure) demonstrated the ability of these epithelial cells in absorption of Na⁺ and secretion of Cl⁻ [56, 57]. Studies from cultured human glandular epithelial cells demonstrated the abilities of ion transport and amiloride-sensitive Na⁺ conductance in these cells [58–60]. Avian uterine (shell gland) epithelium is also capable of transporting Na⁺ and anions and amiloride could inhibit Na⁺ absorption, indicating the involvement of ENaC [61]. These in vitro and ex vivo studies support uterine epithelium as the uterine fluid absorption surface involving the movements of electrolytes. However, the contributions of LE and GE in uterine fluid absorption in vivo during early pregnancy remain unclear.

Our study using intraluminally injected AH clearly demonstrates that, in the control mice, the LE layer plays a dominant role in bulk absorption of uterine fluid, which is reflected by bright smeared AH fluorescence in the LE cells 5 min post AH injection. On D0.5, the uterine fluid volume is maximum under the influence of preovulatory E2. The uterine fluid volume reduces dramatically from D0.5 to D1.5 based on the reduction of LE folding and drops to minimum approaching embryo attachment on ~D4.0 [2]. On D0.5, LE cells have extensive bulk absorption 5 min post AH injection, which is reduced on D3.5. Interestingly, intense bulk absorption on D3.5 is only observed in a segment of LE cells or scattered LE cells. One possible explanation is that different LE cells have different expression levels of the genes critical for bulk absorption, such as ENaC subunits and aquaporins, and the LE cells with stronger molecular equipment for generating an osmotic gradient will be selected for taking the role of bulk absorption when the demand for bulk absorption is limited due to the minimal uterine fluid on D3.5.

Since we evaluate bulk absorption based on the intensity of smeared AH green fluorescence relative to its surrounding, we can be sure of bulk absorption when LE cells have bright smeared AH fluorescence. However, we cannot be sure of lack of bulk absorption if the LE cells do not have bright smeared AH fluorescence 5 min post AH injection, because it is most likely that bulk absorption in the LE is not strong or efficient enough to take up sufficient AH dye to make it much brighter than the surrounding. If there is a significant reduction of LE cells with bright smeared AH fluorescence 1 h post AH injection compared to 5 min post AH injection in the same mouse, one possible explanation is that bulk absorption via LE cells is still strong but there is insufficient AH dye available in the uterine lumen to be absorbed and to subsequently fluoresce the LE cells. However, this might not be a main contributing factor because strong bulk absorption remains detectable at 1 h in a few D0.5 control uteri. Another potential explanation is that amiloride-sensitive ENaC, which is shown to be essential for P4-induced uterine fluid absorption in a uterine perfusion study [8], can be activated by shear flow [62–64], which is expected to be enhanced by intraluminal AH injection and diminishes with time, e.g. stronger at 5 min than 1 h post AH injection.

In preimplantation mice, P4 levels rise shortly after copulation [40–42]. Two studies reported that plasma P4 levels quickly reached the early pregnancy plateau level ~ D2.5 [40, 41], one study indicated plateauing time at ~D3.5 [42], and another study showed it at ~D4.5 [65]. This last report was questionable because plasm P4 levels were shown a trend of decreasing from D0.5 to D1.5 before increasing thereafter [65]. Our study demonstrates that P4 signaling promotes bulk absorption in the preimplantation LE even though the extent of bulk absorption in the LE in the control mice is seemingly inversely correlated with the rising of preimplantation P4 levels. We previously demonstrated that RhoA^f/fPgr^Cre/+ mice were infertile with P4 deficiency [44]. Reduced bulk absorption is observed in the RhoA^f/fPgr^Cre/+ mice both on D0.5 and D3.5. RhoA could activate ENaC [66, 67]. The reduced bulk absorption in the RhoA^f/fPgr^Cre/+ LE might be contributed by RhoA deficiency in the LE and impaired corpus luteum function in P4 synthesis in the RhoA^f/fPgr^Cre/+ mice. The role of P4 signaling in promoting bulk absorption of uterine fluid is confirmed in P4-treated RhoA^f/fPgr^Cre/+ mice, which have increased LE cells with bright smeared AH fluorescence compared to oil-treated RhoA^f/fPgr^Cre/+ mice. On the other hand, since P4 function is mainly mediated by PR, PR antagonist RU486 inhibits bulk absorption of uterine fluid in the LE of D3.5 control mice, further supporting a positive role of P4 signaling in promoting uterine fluid absorption. However, bulk absorption is decreased from D0.5 to D3.5. This finding seemingly contradicts the observations that P4 levels increase from D0.5 to D3.5 and P4 induces uterine fluid absorption. The difference between D0.5 and D3.5 in bulk absorption via LE may reflect the need for bulk absorption because the D0.5 uterus is filled up with uterine fluid while the D3.5 uterus has minimal uterine fluid. On the other hand, the large volume of uterine fluid on D0.5 will create a hydraulic pressure on the apical membrane of the LE. Hydraulic pressure and osmotic pressure are driving forces for water movement in the cell [68]. The flow of uterine fluid can induce shear flow to activate ENaC on epithelial apical membrane [62–64]. Therefore, it is reasonable to speculate that the hydraulic pressure and osmotic pressure could both contribute to the extensive bulk entry of uterine fluid on D0.5. Our AH tracking approach in this study could not trace the exit mechanisms of the absorbed uterine fluid in the LE. Based on bulk absorption of water along the osmotic gradient from the uterine lumen, the expected exit of Na⁺ to the subepithelial area via Na⁺/K⁺ ATPase will generate an osmotic gradient across LE basal membrane to facilitate water exit to the subepithelial area.

Our study using intraluminally injected AH seemingly reveals endocytosis in D3.5 LE, visualized as green fine dots in the apical LE. However, our further study demonstrates that these green fine dots in the apical LE are autofluorescent. Therefore, the intraluminally injected AH dye could not specifically trace active apical endocytic activity in the LE. One study used intraluminally injected horseradish peroxidase (HRP) to trace apical endocytosis in the D0.5 mouse uterus but failed to detect any endocytosis of HRP [24]. We could detect scattered green fine dots in the apical side of a few LE cells of some D0.5 uterine samples but not all, however, we could not be sure if they were vesicles with active endocytic activity. In the D3.5 LE, there is a trend of more dots from 5 min to 1 h post AH injection, which is opposite of bulk absorption. Based on the literature that rat LE cells approaching embryo implantation time are active in apical endocytosis (but not apocrine secretion) that channels into lysosomes [50, 51] and our detection of LAMP2, a marker of lysosome-associated membranes, as a string of apical punctae in the D3.5 LE right prior to embryo implantation initiation in mice [2], we are confident that the strings of green fine dots in the D3.5 apical LE are reflecting increased LE apical endocytic activity from that on D0.5.

Luminal epithelium apical endocytosis is minimal on D0.5 when the maximal uterine fluid reduction occurs accompanied with maximal bulk absorption; this unequivocally places bulk absorption via LE as the dominant mechanism for uterine fluid reduction on D0.5. Endocytosis in LE is increased from D0.5 to D3.5. This change may reflect the functional needs on these respective early pregnancy days. On early D0.5 (shortly post coitum), the uterine fluid serves as a passage to facilitate sperm transport to the oviduct for fertilization. On D3.5, there is minimal uterine fluid; therefore, the requirement for bulk absorption of uterine fluid is reduced; on the other hand, the embryo in the D3.5 uterus is preparing for embryo implantation, which initiates ~D4.0 in mice. One prerequisite for embryo implantation is the mutual communications between the embryo and the uterus to ensure the synchronized readiness of both the embryo and the uterus for embryo implantation [2, 11]. Endocytosis can internalize various molecules, including messenger molecules, which are expected to facilitate embryo–uterus communications for embryo implantation. Prior to embryo attachment (~D4.0 in mice) [43], there are protrusions (pinopods/uterodomes) appearing on apical LE, coincident with a brief period of intense endocytic activity at the time of embryo implantation [35, 51, 69]. In ovariectomized rats, P4 promotes uterine lumen closure and endocytosis [70]. These previous reports consistently demonstrated the critical role of P4 in apical LE endocytosis. Our study using intraluminally injected AH shows that LE apical endocytosis is increased from D0.5 to D3.5, which is correlated with the increased P4 levels during preimplantation. However, P4 treatment and RU486 treatment (on D1.5 and D2.5) do not seem to have obvious and consistent effects on the LE apical endocytosis detected on D3.5. It is unclear if it is because of the 24+ h between the last treatment and AH detection that missed the best time for detecting the effect on endocytosis, or unlike ovariectomy, our experimental setting with intact ovaries during early pregnancy was not sensitive enough to detect the effects of exogenous P4 treatment and RU486 treatment on LE apical endocytosis.

Based on our observations, the contribution of GE on uterine fluid absorption is minimal for the following reasons: First, the GEs are folds from LE into the stromal layer. Unless there is a suction system to draw the uterine fluid from the uterine lumen to the glandular lumen, the direct contact surface of GE to the bulk of uterine fluid is limited. Second, bulk absorption of uterine fluid in GE is rare on both D0.5 and D3.5. One potential reason could be that the volume of uterine fluid in the lumen of the uterine gland is insufficient for a need of bulk absorption. Third, although the presence of green dots varies greatly among different GEs in the same uterine sections or in the uterine sections from different mice in the same treatment groups, the green dots are present in some GEs of all the uterine samples examined. Since both mitochondria and lysosomes are autofluorescent [46], and the green dots in the GEs vary in sizes and cellular localization, we could not differentiate any of the green dots in GE cells to be endocytic vesicle. However, there is no obvious and consistent effect/evidence of P4 signaling on the abundance and cellular distribution of the green dots in the GEs despite the observations of P4-induced uterine fluid absorption in the RhoA^f/fPgr^Cre/+ mice with P4 deficiency and RU486-induced uterine fluid accumulation in the RhoA^f/f control mice. Interestingly, it was reported that in ovariectomized rat GE cells, the number of basal pinocytotic invaginations was relatively small and unresponsive to P4 stimulation [6]. These observations suggest that apical and basal endocytosis (including pinocytosis) in GE cells might be not regulated by P4 signaling, further demonstrating that LE and GE have different functions and regulatory mechanisms in uterine fluid absorption.

In summary, this study establishes a straightforward approach to simultaneously visualize bulk entry (directly) and endocytosis (somewhat indirectly) of uterine fluid absorption during early pregnancy. It reveals the temporal mechanisms of preimplantation uterine fluid absorption, with bulk absorption in the LE reduces from D0.5 to D3.5 while endocytosis in the LE increases from D0.5 to D3.5, as well as the suppressive effects of P4 deficiency and RU486 on bulk absorption in the LE during early pregnancy. This study also demonstrates the dominant role of LE but not GE in preimplantation uterine fluid absorption. The molecular mechanisms for preimplantation uterine fluid absorption remain to be elucidated.

Data availability

Data available on request.

Authors’ contributions

YL and XY conceived the project and designed the experiments. YL performed the majority of the experiments. TEM, JMH, RL, and SV assisted on maintaining mouse colony and data collection. YZ contributed RhoA^f/f mice and JPL contributed Pgr^Cre/+ mice for generating RhoA^f/fPgr^Cre/+ mice used in this study. XY supervised the entire project. YL and XY analyzed the data and wrote the manuscript; all authors reviewed /revised the manuscript.

Conflicts of interest

The authors have declared that no conflict of interest exists.