Characteristics of the uterine luminal surface epithelium at preovulatory and preimplantation stages in the marmoset monkey


  • Andrea Louise Niklaus,

    Corresponding author
    1. Department of Obstetrics and Gynecology, University of Melbourne, Victoria 3053, Australia
    Current affiliation:
    1. Department of Obstetrics and Gynecology and Reproductive Sciences, University of Maryland, MD 21201
    • The University of Maryland, Department of Obstetrics and Gynecology and Reproductive Sciences, Bressler Research Building 11-019, 655 West Baltimore Street, Baltimore, MD 21201
    Search for more papers by this author
  • Christopher Raymond Murphy,

    1. Department of Anatomy and Histology, University of Sydney, NSW 2006, Australia
    Search for more papers by this author
  • Alexander Lopata

    1. Department of Obstetrics and Gynecology, University of Melbourne, Victoria 3053, Australia
    Search for more papers by this author


Light and electron microscopy was used to examine the apical luminal epithelial surface of the uterus at preovulatory and preimplantation stages in the marmoset monkey. Luminal surface charge, detected by cationic ferritin staining, progressively decreased from preovulation to day 11 of pregnancy. The smooth, regular apical plasma membrane at preovulatory stages was in contrast to the convoluted, irregular surface observed during early pregnancy, especially at 1 day before blastocyst implantation. Profiles of microvilli were also altered, becoming thicker and more irregular during early pregnancy. Within the epithelial cell body, cyclic morphologic changes were seen, largely in association with secretory organelles. Giant phagocytic bodies were prominent at all stages examined, although their composition and intensity of staining varied throughout the cycle. Weak to moderate estrogen alpha and progesterone receptor immunostaining of the luminal epithelium was found during preovulatory and early pregnancy stages. This study describes complex cyclic changes in the morphology and biochemical make-up of the uterine luminal epithelial surface in a New World monkey in preparation for blastocyst attachment. Anat Rec 264:82–92, 2001. © 2001 Wiley-Liss, Inc.

The apical surface of the uterine luminal epithelium is the site of initial contact with the blastocyst, yet the characteristics of receptive and nonreceptive epithelial surfaces remain an enigma. Partially, this finding is related to the emphasis on and examination of a more abundant uterine epithelial source, the glandular epithelium. As early as 1951, Bartelmez et al. (1951) suspected that the threshold of response to ovarian steroids may differ between luminal and glandular epithelia. Also, differential steroid receptor staining has been reported in uterine glandular and luminal epithelia of primates treated with similar doses of steroids (Okulicz et al., 1990, 1993).

Many studies have been carried out on the ultrastructure of the endometrium at various stages of the human menstrual cycle (Wynn, 1977; Dockery and Rogers, 1989; Li et al., 1994). Yet, the uterine epithelial surface, or Zone I of the functionalis, has been overlooked in the majority of cases (Gompel, 1962; Nilsson, 1962, 1974; Dockery et al., 1988). Almost 30 years ago, Ferenczy and Richart (1973) first highlighted this problem, and to the present day, there is little information on the ultrastructure of the uterine luminal epithelium in primates (Li et al., 1994).

In the present study, we aimed to define some of the basic ultrastructural changes in the uterine luminal epithelial cell surface at preovulatory and early pregnancy stages in the marmoset monkey. Apart from morphology, several other changes to the uterine luminal epithelium are known to be important in the apposition and attachment of the blastocyst. We investigated characteristics of the luminal epithelium that have not been examined in primates but are known to be regulated by ovarian hormones in rodents and other mammals (Murphy, 1993). These characteristics included apical plasma membrane regularity, surface charge, and profiles of microvilli (MV). The cyclic expression of estrogen and progesterone receptors was also investigated. Overall, this study provides an insight into the basic cellular make-up of the uterine luminal surface epithelium in a primate at preovulation, days 4–8 of pregnancy and day 11 of pregnancy for the first time. Because day 11 of pregnancy is 1 day before blastocyst attachment in the marmoset monkey (Smith et al., 1987; Enders and Lopata, 1999), features of the uterine luminal epithelium at pre- and peri-implantation stages are elucidated in this study.


Collection of Endometrial Tissue

Female marmoset monkeys (Callithrix jacchus) over 18 months of age with regular reproductive cycling patterns and of proven fertility were used for this study. As previously described (Niklaus et al., 1999), luteolysis and the initiation of a new reproductive cycle (day 1 of proliferative phase) was induced by an intramuscular injection of 0.5 μg Estrumate (I.C.I., PitmanMoore, Melbourne, Australia). Ovulation (approximately 9–10 days after luteolysis) was determined by a decrease in plasma progesterone levels to less than 10 ng/ml (Hodges et al., 1987). Day one of pregnancy was defined as the day of ovulation in females housed with regularly mating fertile males. Embryos were recovered in approximately 70% of cases at day 11 of pregnancy. At appropriate stages of the reproductive cycle, animals were anaesthetized by administering 0.8 ml of Saffan (Intervet, Melbourne, Australia) intramuscularly and uterine tissue was collected by hysterectomy or curettage as described previously (Niklaus et al., 1999). Three stages of the marmoset reproductive cycle were examined; early-late proliferation (preovulatory), days 4–8 of pregnancy, and day 11 of pregnancy. Collectively, the early embryonic stages without attachment are referred to as preimplantation stages. Day 11 of pregnancy has been estimated to correspond to 1 day before marmoset blastocyst implantation (Smith et al., 1987; Enders and Lopata, 1999) and is also referred to as peri-implantation. Uteri were excised only when an animal had undergone the maximum number of laparotomies (n = 6) approved by the Royal Women's Hospital and University of Melbourne Animal Experimentation Ethics Committees.

Ultrastructural Studies

Transmission electron microscopy was used to investigate the following characteristics of preovulatory and preimplantation staged apical uterine luminal epithelium: (1) general epithelial cell architecture, (2) apical plasma membrane surface regularity, (3) surface profiles of MV, and (4) apical luminal surface negativity.

General epithelial cell architecture and apical plasma membrane surface regularity.

Hysterectomy specimens were obtained from five animals each at the respective stages: preovulatory (n = 5), days 4–8 of pregnancy (n = 5), and day 11 of pregnancy (n = 5). Most of these tissues had been previously used for lectin histochemistry studies of the apical luminal epithelium (Niklaus et al., 1999).

Immediately after excision, uteri were placed in culture medium, by using the alpha modification of Minimum Essential Medium (α-MEM; Trace Biosciences Pty, Ltd., Castle Hill, Australia) and bisected by sagittal or coronal sectioning to expose the endometrium. Excess myometrium was trimmed away to make a border of endometrium that was subsequently cut into 15 to 20 segments, each approximately 2 × 2.5 mm in size. After washing with 0.1 M phosphate buffer at pH 7.4, the uterine segments were immersion fixed in 2.5% glutaraldehyde (in phosphate buffer) for 2 hr.

Tissues were rinsed overnight in phosphate buffer, post-fixed in 2% osmium tetroxide, dehydrated in graded ethanol, and embedded in Spurr's resin (ProSciTech., Thuringowa, Australia). Thick sections (1–2 μm) were stained with 1% Toluidine Blue in 1% borax and surveyed for uterine luminal surface epithelium. Light microscopic images were taken with a Leica DMLB microscope to provide additional information about the uterine luminal surface morphology across the marmoset reproductive cycle.

Subsequent thin sections (pale gold or 90 nm) were cut, mounted onto 150-mesh grids, stained with uranyl acetate and lead citrate, and examined at 80 kv with a Phillips 400 or Joel 101 transmission electron microscope. Representative areas from at least three blocks of tissue were analyzed for each animal.

Apical uterine luminal surface microvilli and their profiles.

Uterine tissue from 12 of the 15 animals processed for general electron microscopic (EM) studies were used for morphometric analysis of profiles of MV. Four animals at each stage were used for MV analysis.


Random photos of surface epithelium were taken at a magnification of 28,000× by using the microscope camera frame as a guide as described previously (Niklaus et al., 1999). A total of 30 electron photomicrographs from at least six blocks of tissue, enlarged to 90,000×, were used per animal for MV analysis.

The parameters assessed were (1) number of MV per 10 cm length of photographed membrane, (2) length (short ≤ 1.5 cm; medium 1.6 to 4.4cm; long ≥ 4.5 cm), (3) thickness (thick ≥ 1.5 cm), and (4) regularity of MV shape. For the last three parameters, only membranes sectioned perpendicular to the luminal surface were considered. Only MV seen in their entirety were counted. Regularity of MV was a parameter analyzed subjectively. Microvilli that were not finger-like or tapered were considered to be irregular in shape.

Statistical analysis.

Statistical analysis of the number and length parameters was evaluated by one-way analysis of variance (ANOVA) and Tukey's pairwise comparisons. MV shape and length were analysed by Kruskal-Wallis and Mann-Whitney statistical tests. Probability values of less than 0.05 (P < 0.05) were considered significant. In addition, to determine whether MV were similar in size, shape, and distribution on anterior and posterior uterine surfaces, a paired t-test and Wilcoxon signed rank test were performed.

Apical Luminal Surface Negativity

Cationic ferritin was used to examine the distribution of anionic sites across the marmoset uterine luminal surface epithelium. Twelve animals were used for this study, four each at preovulatory, days 4–8 of pregnancy, and day 11 of pregnancy stages.

Detection of anionic sites.

Curettage biopsy specimens were rinsed in Veronal HCl (Michaelis) buffer, pH 7.2, and then immersion fixed in 2.5% glutaraldehyde for approximately 60 min. Once washed in buffer, the epithelial sheets were cut into smaller fragments to optimize probe penetration. Tissues were then exposed to cationic ferritin (CF) (ProSciTech., Thuringowa, Australia) diluted to 0.5 mg/ml with Michaelis buffer for 30 min with agitation, rinsed thoroughly in buffer, and prepared for EM as described above. However, thin sections were examined unstained at 80 kV with a Phillips CM12 transmission electron microscope. Staining with CF was compared with incubations in avidin-ferritin conjugate (Part A) at 1 mg/ml to rule out nonspecific ferritin staining.

Analysis of staining density.

Within each grid-bar region viewed under the electron microscope, an estimate of the percentage coverage of CF staining along the luminal epithelium was estimated. An overall percentage of the coverage of CF staining of the luminal epithelium per animal was determined. Observations of the pattern of CF staining with respect to MV and inter-MV membranes were also recorded.

Estrogen and Progesterone Receptor Expression

The cyclic protein expression of estrogen receptor alpha (ERα) and progesterone receptors (PR) in the uterine luminal epithelium was determined by immunocytochemical means. The number of animal studied at preovulatory, days 4–8 of pregnancy, and day 11 of pregnancy stages were three, two, and three, respectively.

Immunocytochemical procedure.

Uteri were bisected and immersion fixed in 4% paraformaldehyde in phosphate buffer for 2 to 3 hr. After thorough rinsing, specimens were processed for routine histology by using a Tissue Tek VIP E150 and embedded in paraffin wax. Sections 3 μm in thickness were mounted on 2% 3-aminopropyltriethoxy-silane (Sigma Chemical Company, St. Louis, MO) -coated slides, dried overnight at 37°C, and then processed for immunocytochemistry.

All incubations required for the immunocytochemical procedure were performed in a humidified chamber at room temperature (RT). After deparaffinization and rehydration, sections were treated with 3% hydrogen peroxide (Sigma) to quench endogenous peroxidase activity. After three rinses in phosphate buffer, sections were further subjected to a 10-min heat-induced antigen retrieval procedure by using 0.01 M citrate buffer. After the slides were cooled for 30 min at RT, sections were subsequently rinsed and incubated with normal rabbit serum (Dako Corporation, Carpinteria, CA) for 20 min to block nonspecific antibody binding. Sections were then incubated for 60 min with monoclonal mouse anti-human ER (Zymed Laboratories, Inc., San Francisco, CA) or monoclonal mouse anti-human PR (Zymed). The ERα and PR antibodies were specific for peptides representing the recombinant estrogen receptor alpha and the N-terminal of human progesterone receptor respectively.

After thorough washes, the secondary antibody, biotinylated rabbit anti-mouse immunoglobulins (Dako) was applied for 30 min, followed by an avidin-biotin-peroxidase complex for the same length of time. The staining was developed with 3,3′-diaminobenzidine tetrachloride solution (Dako) for 5 min, counterstained with Harris hematoxylin, and cleared with Xylene. Negative controls were run in parallel to the uterine specimens by replacing the primary antibody with mouse primary antibody isotype control (Zymed).

Analysis of receptor staining.

The intensity of the immunocytochemistry staining for both steroid receptors was analyzed by using an objective of 40× at the light microscope level by two independent observers. Negative, weak, moderate, and strong staining was scored, respectively, on a scale from 0 to 3.


General Morphologic Assessment

Columnar uterine luminal epithelial cells of regular thickness were characteristic of the marmoset proliferative phase (Fig. 1A). At this stage, nuclei with regular outlines were located centrally and basally. In comparison, during preimplantation stages, the epithelial cells were taller and more irregularly shaped with nuclei in varied positions, producing a pseudostratified-like appearance (Fig. 1B).

Figure 1.

At the light microscope level, the regularity of the uterine surface at (A) preovulation contrasted to the convoluted surface observed at (B) day 11 of pregnancy. Several cells had apical protrusions (asterisks) of varying size at day 11 of pregnancy, and the epithelium appeared pseudostratified. The electron dense oval cell bodies (arrows), found within every few cells, were prominent at preovulatory stages.

In epithelium from the proliferative phase, mitochondria appeared small, concentrated mainly in the basal region of the cell. The major secretory components such as endoplasmic reticulum and Golgi apparatus, varied in length but were usually compact and not dilated (Fig. 2A). Nuclei contained predominately pale chromatin (euchromatin).

Figure 2.

Uterine luminal epithelial cell secretory organelles such as Golgi apparatus (GA) were smaller and less extensive at (A) preovulation compared with (B) those at day 11 of pregnancy. Cell membranes are labeled (CM). N, nucleus.

During early pregnancy, the uterine luminal epithelial cell cytoplasm was well differentiated with larger mitochondria and more extensive secretory organelles. Dilated endoplasmic reticulum, particularly the rough endoplasmic reticulum, became more prominent as pregnancy advanced. Golgi complexes were extensive with numerous Golgi vesicles of varying diameters occupying the cell cytoplasm at this time (Fig. 2B). Located in the cell apex, the terminal web microfilaments also became more apparent at day 11 of pregnancy (Fig. 3). Granular forms of chromatin (heterochromatin) were more typical of uterine epithelial cell nuclei of pregnant rather than nonpregnant animals.

Figure 3.

An electron photomicrograph of a surface epithelial protrusion at day 11 of pregnancy, illustrating in the apical cell cytoplasm, the presence of multivesicular body(s) (MVB), and the prominent distribution of microfilaments (MF).

Most conspicuous within the preovulatory epithelial cell cytoplasm at both light and electron microscopic levels was the presence of large, oval, usually intensely stained bodies. Termed oval cell bodies (OCBs), these structures appeared vacuolated or swollen and were common to most luminal epithelial cells (Fig. 1A). These OCBs appeared as giant membrane-bound vacuoles composed of other structures such as (1) smaller circular bodies (CBs) that varied in density, (2) smaller circular membrane whorls that were mostly of intermediate electron density, and (3) medium to large aggregations of electron dense granules (DGs).

Before ovulation and during preimplantation stages, changes in the composition of OCBs varied but the major changes were associated with size, number, and staining intensity. CBs were the most predominant and DGs were the least common component of OCBs.

At preovulation, OCBs appeared densely stained due to the prevalence of electron dense CBs and DGs (Fig. 4A). The large size of the DGs contributed greatly to the overall enormity of these structures that ranged in diameter from approximately 1 to 8 μm (Fig. 4A). These giant OCBs were located primarily in a supranuclear position (Fig. 1A). Additionally, OCBs were seen in close association with secretory organelles and mitochondria (Fig. 4A).

Figure 4.

The ultrastructural features of oval cell bodies (OCBs; dotted circles) within the cytoplasm of the luminal epithelium at (A) preovulation and (B) day 11 of pregnancy. At preovulatory stages, the OCBs that appeared in close association with mitochondria (M) and endoplasmic reticulum (ER) were composed of electron-dense structures such as circular bodies (CB) and dense granules (DG). OCBs, were larger and electron lucent at day 11 of pregnancy and contained structures such as membrane whorls (arrows). Nuclei (N) are labeled.

Generally OCBs were smaller (0.5 μm to 1.0 μm in diameter) and less abundant at days 4–8 of pregnancy than at preovulatory stages. They were distributed similarly as within preovulatory epithelium but were generally less intensely stained.

At day 11 of pregnancy, 1 day before blastocyst attachment, the OCBs were numerous and varied in size and location. Similar to preovulatory epithelium, the OCBs were mostly large but generally lacked DGs. However, a striking difference to preovulatory and days 4–8 of pregnancy stages was their electron-lucent rather than dense appearance (Fig. 4B). At this stage, OCBs were more randomly distributed within the cell cytoplasm.

From over 50 blocks of apical uterine luminal epithelium examined, only 4 contained ciliated cells, which were usually confined to a small region of the surface. Four, 6, and 35 ciliated cells were located within epithelium from preovulatory, days 4–8 of pregnancy and day 11 of pregnancy groups, respectively.

Apical Plasma Membrane

Surface morphology.

Throughout the marmoset reproductive cycle, variation in the contours of the apical surface was evident at the light and electron microscopic levels. Before ovulation, the apical uterine epithelial surface appeared smooth and regular (Fig. 1A). This finding contrasted to preimplantation stages when a greater variation and irregularity of the luminal surface was demonstrated (Fig. 1B). Apical protrusions ranging in diameter from 1–10 μm were only seen during early pregnancy (Figs. 1B, 5A,B). Either dome shaped or highly irregular, these bleb-like structures were found regularly along most regions of the uterine surface at days 4–8 and day 11 of pregnancy (Fig. 1B). Despite some variability, the outline of surface protrusions generally became increasingly irregular with advanced pregnancy (Fig. 5A,B). Often, small, round membrane bound bodies containing numerous tiny vacuoles ranging in diameter from 0.35 μm to 1.5 μm were seen within the body of the apical protrusions (Fig. 5B).

Figure 5.

Apical cell protrusions or uterodomes (thick arrows), extending out from the cell surface membrane, varied in size during (A) days 4–8 of and (B) day 11 of pregnancy and generally became more irregular as implantation approached (day 11 of pregnancy). These structures often contained vesicles (Ve) and various irregular profiles of microvilli were observed across their surface (thin arrows). Cell membranes (CM) are labeled.

Profiles of microvilli.

There were no significant differences in MV length (Fig. 6A) or number (Fig. 6B) per unit length of membrane across the marmoset reproductive cycle. However, MV became significantly more irregular in shape (P < 0.05) between the preovulatory stage and days 4–8 of pregnancy as well as between the preovulatory stage and day 11 of pregnancy (Fig. 6C). Also significant differences were found in the MV thickness when preovulatory stages and days 4–8 of pregnancy were compared (P < 0.01) and from when preovulatory and day 11 of pregnancy (P < 0.05) were compared (Fig. 6D). Thus, MV became thicker and more irregularly shaped during the early stages of pregnancy (Fig. 7). No significant differences were observed in shape (Fig. 6C) or thickness (Fig. 6D) between early pregnancy groups. Furthermore, there were no significant differences in any of the MV parameters between epithelia from the anterior and the posterior walls of the uterus.

Figure 6.

There was no difference in the (A) length and (B) number of microvilli (MV) on the uterine luminal surface at various stages of the marmoset reproductive cycle. However, compared with preovulatory stages, there was a significant increase (P < 0.05) in the (C) MV irregularity and (D) thickness during early pregnancy.

Figure 7.

Regular and thin microvilli at (A) preovulatory stages contrasted to the significantly thicker and more irregular profiles observed on the uterine surface during (B) early pregnancy (day 11 of pregnancy shown here). Ferritin particles represent staining with lectins and not cationic ferritin.

Apical Luminal Surface Negativity

Control studies.

Incubations of marmoset luminal epithelium in avidin-ferritin conjugate demonstrated an absence of nonspecific ferritin staining. This finding suggests that neutral ferritin does not bind to the apical surface and supports the idea that cationic ferritin (CF) is likely to reflect specific binding to negatively charged surfaces.

Uterine epithelial anionic sites.

At each of the stages examined, there were many regions that were devoid of CF staining. For this reason, the percentage of ferritin covering epithelial areas was estimated. Overall, maximum CF staining was seen at the preovulatory stage (Fig. 8A), with a progressive decrease in the distribution of anionic sites as pregnancy advanced (Fig. 8B,C). Although regions of uterine epithelium displayed patchy staining (areas of both negative and positive reactivity), the overall ferritin binding pattern within groups was reproducible.

Figure 8.

There was a decrease in the abundance of cationic ferritin staining across the apical plasma membrane from (A) preovulatory, (B) days 4–8 of pregnancy, to (C) day 11 of pregnancy.

At the preovulatory stage (Fig. 8A), CF bound to 40–80% of epithelial surfaces analyzed in comparison to 10–20% coverage at days 4–8 (Fig. 8B) and 5–10% coverage at day 11 of pregnancy (Fig. 8C). Cationic ferritin bound to both MV and the inter-MV regions of the apical membranes at the preovulatory stage (Fig. 8A). Conversely, at both stages of pregnancy, cationic staining was largely confined to the membranes of MV (Fig. 8B,C). In particular, sparse staining and small random ferritin aggregates were observed at day 11 of pregnancy (Fig. 8C).

Estrogen and Progesterone Receptor Expression

There was no difference in the ERα pattern of the apical uterine luminal epithelium at the three stages of the reproductive cycle examined. A minority (10–30%) of cells were weakly stained for ERα, whereas the remaining showed negative reactivity at the three stages examined (Fig. 9A). PR immunostaining of luminal epithelial cells was also similar at preovulatory, days 4–8 of pregnancy, and day 11 of pregnancy stages. The majority of cells (50–70%) were unstained for PR, although the immunoreactivity in the stained cells ranged from weak to moderate (Fig. 9B). Irrespective of the stage examined, there was a marked increase in the intensity of ERα staining and PR staining of the glandular epithelium compared with the luminal surface (Fig. 9A,B).

Figure 9.

A: Weak estrogen alpha and (B) weak to moderate progesterone receptor staining of the uterine luminal epithelium detected by immunocytochemistry was similar in pregnant and nonpregnant groups. However, receptor staining differed between the luminal surface (LE) and glandular epithelium (GE) at all stages. Endometrium from (A) days 4–8 of pregnancy and (B) preovulatory stages are shown.


Because the luminal surface is the site of initial contact of the blastocyst at implantation, an understanding of the ultrastructural changes associated with such early interaction is important. The three-dimensional microarchitecture of the uterine luminal epithelium has been extensively studied by using scanning electron microscopy (Ferenczy et al., 1972; Martel et al., 1981; Nikas et al., 1995). However, little information is available on the ultrastructural detail of these surface cells in humans (Wynn, 1977; Cornillie et al., 1985; Dockery and Rogers, 1989) and nonhuman primates (MacLennan et al., 1971; Rune et al., 1992; Ghosh et al., 1996).

General Morphologic Assessment

The general ultrastructural features of uterine luminal epithelial cells of the marmoset resemble those described for human endometrial glandular cells (Wynn, 1977; Dockery et al., 1988). During the proliferative stage of the cycle corresponding to the preovulatory epithelial specimens studied, the endometrium was mainly under the influence of estrogen. At this stage, the relatively undifferentiated uterine epithelial cells proliferated and appeared to develop luminal cell organelles required for later phases. The large amount of euchromatin contained within uterine luminal epithelium during the proliferative phase suggested that gene transcription activity was occurring (Armstrong et al., 1973; Verma, 1983; Cornillie et al., 1985).

During the mid to late secretory phase of the cycle of the marmoset, uterine epithelial cells were predominantly under the influence of progesterone. At day 11 of pregnancy, the uterine luminal epithelial cells were well differentiated and appeared to be at maximal secretory activity. The major changes in nuclei, endoplasmic reticulum, mitochondria, and Golgi apparatus during early pregnancy suggested enhanced protein synthesis and transport by the uterine luminal epithelium.

Oval cell bodies.

Of particular interest in this study was the prominence of uterine luminal epithelial OCBs and their products, which were distributed throughout the cell cytoplasm at the three stages of the marmoset reproductive cycle that were studied. These large, complex structures have not been described previously in the uterine luminal epithelium of primates. Bearing some resemblance to OCBs, lysosome-like structures, given a variety of names such as “phagolysosomes,” “autophagosomes,” and “granulated bodies” have been identified in the primate glandular epithelium (Armstrong et al., 1973; Verma, 1983; Smith et al., 1987; Kaiserman-Abramof, 1989; Rune et al., 1992). The OCBs described here differed from these lysosome-like structures in abundance and size.

Given the similarities between the compositions of OCBs and the lysosome-like structures, it is likely that these giant intracellular bodies may be involved in cytoplasmic turnover and cell breakdown (Armstrong et al., 1973). Because the marmoset is a nonmenstruating primate, these OCBs may be involved in the removal and breakdown of older cells that facilitates regeneration of a new epithelium.

Surface irregularity.

During preimplantation stages in the marmoset, the uterine epithelial surface was convoluted with large cytoplasmic projections. In rats and mice, similar structures have been found (Enders and Nelson, 1973; Parr and Parr, 1977). Experiments in which ferritin tracer was injected into the uterine lumen proved that these apical structures had pinocytotic activity (Enders and Nelson, 1973); hence, they became known as pinopods. This term has since been widely applied to describe uterine surface projections in many animals, including humans, despite lack of evidence of any pinocytotic function in these other species and indeed some evidence to the contrary (Parr and Parr, 1982; Guillomot et al., 1986; Psychoyos and Nikas, 1994; Nikas and Psychoyos, 1997; Murphy, 2000).

The true pinopods on the apical surface of rats and mice appear to be somewhat different in structure to those structures seen in many other animals (Murphy, 2000). In rats and mice, pinopods do not occupy the majority of the cell surface and are only found on a minority of cells (Nilsson, 1972). Uterine surface protrusions in other animals, particularly in humans (Psychoyos and Nikas, 1994), generally arise from the entire cell surface. We also find this to be the case in the marmoset. Additionally, as in other animals (Abd-Elnaeim et al., 1999) including primates (Bhartiya and Bajpai, 1995), the apical protrusions of the marmoset luminal epithelium appeared on the majority of cells. Although the marmoset apical protrusions contained vesicles somewhat like that observed in pinopods of rats and mice (Enders and Nelson, 1973; Parr and Parr, 1974), we did not observe the characteristic large vacuoles and have no evidence of their function. Therefore, we prefer to use the term “uterodome” to describe them as recently proposed (Murphy, 2000). Function aside, uterodomes may be potential markers for the window of receptivity in the marmoset as are the so-called “pinopods” in the human (Nikas et al., 1995). However, we have not determined in this study whether marmoset uterodomes are reduced at the time of implantation as true pinopods are in rodents (Parr and Parr, 1974; Singh et al., 1996).

Profiles of microvilli.

It is well known that ovarian hormones influence the MV of uterine luminal epithelium in a variety of mammals (Murphy, 1993), and it is thought that these changes are important in the apposition and attachment of the blastocyst. In the rat, MV changes as part of the plasma membrane transformation are observed as early as day 3 of pregnancy and occur across the entire luminal surface (Murphy and Shaw, 1994; Murphy, 1995). Studies of human uterine luminal epithelium by scanning electron microscopy also suggest that shorter and more irregular profiles of MV are present at the luteal phase compared with the proliferative phase (Nilsson et al., 1980; Martel et al., 1981). The present study did not observe flattening of the apical surface as described in other species, but it is possible that localized changes occur at the implantation site. Because we did not focus on this region, flattening of the MV may have gone undetected. However, the present study in the marmoset is consistent with many others in describing major changes in profiles of MV on uterine epithelial cells during early pregnancy, especially the increasing irregularity of these apical structures (Nilsson, 1966, 1974; Ljungkvist, 1972; Salazar-Rubio et al., 1980; Murphy, 1993; Png and Murphy, 1997).

Ciliated epithelial cells.

The present study found limited populations of ciliated cells in the uterine luminal epithelium at all stages examined. Similar findings have been reported in human (Gompel, 1962; Colville, 1968) and nonhuman primates (Enders et al., 1983). The exact role of cilia in reproduction is not known, but it is generally thought that they facilitate the movement of secretions and embryos across the uterine surface (Ferenczy et al., 1972; Johannisson and Nilsson, 1972).

Surface Negativity

Electronegativity of the marmoset uterine epithelial surface was shown to progressively decrease from the preovulatory stage up to day 11 of pregnancy. A similar decrease in uterine epithelial surface negativity as implantation approaches, illustrated by CF staining, has also been described in rodents (Hewitt et al., 1979; Hosie and Murphy, 1989; Chavez, 1990), rabbits (Anderson and Hoffman, 1984), and the cynomolgus monkey (Anderson et al., 1990). In addition, these results are supported by studies that used other reagents to detect surface negativity (Jansen et al., 1985; Hosie and Murphy, 1989; Anderson et al., 1990).

The characteristic patchiness of the uterine surface that reacted with CF has been also described in the rabbit (Hewitt et al., 1979). Chavez (1990) also mentioned that the uterine luminal epithelial surface of mice stained uniformly with CF and certain lectins but that individual cells showed heterogeneity. It is also interesting that the uterine luminal surface of the marmoset, reacted variably from cell to cell to a range of lectins, despite overall homogeneity across a large area of sampled epithelium (Niklaus et al., 1999). Other than protocol variation or species differences, this variability may be explained by the variation in the content and acidity of the uterine glycocalyx. Perhaps the state of the plasma membrane and associated glycocalyx of each uterine epithelial cell reflects the transient metabolic and synthetic activity of the cell, especially related to glycolysis. Preference of CF staining for MV and inter-MV membranes has also been described previously in the rat (Hewitt et al., 1979) and gilt (Blair et al., 1991). How such localized arrangement of anionic sites influences blastocyst apposition is not known.

In preparation for implantation, the maternal surface of most mammals decreases its negativity (Nilsson et al., 1973; Murphy, 1993), but the degree to which the membrane charge is altered on the embryonic surface is not clear. As both have negatively charged surfaces, it seems unlikely that adhesion between these cells would be facilitated by surface charge alone. However, the decrease in luminal epithelial electronegativity may be important because cellular adhesion is enhanced when repulsive charges on opposing membranes are decreased (Vicker and Edwards, 1972; Weiss et al., 1975; Grinnell, 1976).

Estrogen and Progesterone Receptors

Differential ovarian hormone receptor staining of the luminal and glandular epithelium has been described in women (Press et al., 1986; Press and Greene, 1988) and monkeys (Okulicz et al., 1993). We also observed more intense receptor staining in the glandular than luminal epithelium. It was surprising that ERα and PR staining patterns of the luminal surface did not differ across the three marmoset stages examined. Okulicz and colleagues (1993) found ERα staining present in luminal epithelium of artificially stimulated rhesus monkeys with peak estrogen levels, but that this reactivity diminished by day 17 of the artificial cycle, when estrogen disappeared and progesterone levels rose. It is possible that implantation sites could have a different steroid receptor pattern and that species and methodology differences may account for the variability in receptor staining reported.

In summary, our studies have shown the complexity of the uterine luminal epithelium across the marmoset reproductive cycle. Ovarian steroids play a fundamental role in reorganizing the luminal epithelial surface and epithelial cell layer particularly at preimplantation stages. The major constitutive changes to the epithelial cells during early pregnancy, appear to be related to cell and protein turnover. Together with changes in cell surface charge and the contours of the apical plasma membrane, the events leading up to blastocyst implantation are seen to be highly dynamic. As evident from the different steroid receptor staining patterns in the luminal compared with glandular cells, the uterine luminal surface epithelium is a unique component of the endometrium that is remodelled morphologically and biochemically in preparation for blastocyst implantation.


We thank Judy Borg, Alice Franek, and Margot Hoise for their technical contribution; Angela Nelson for assistance with animal care; and Huon O'Sullivan for surgical procedures. Also, we thank James Bisley for comments on the manuscript.