Ultrastructural identification of CD9 positive extracellular vesicles released from human embryos and transported through the zona pellucida

ABSTRACT

Extracellular vesicles (EVs) are highly specific and multi-purpose vesicular structures that are released by various cell and tissue types in the body. However, the secretion of EVs from mammalian embryos, especially human, has not been well characterized. Thus, the aim of this study was to 1) identify EVs in human preimplantation embryos at different stages of their development using scanning and electron microscopy, and 2) investigate whether EVs can cross the zona pellucida (ZP) and be released from human embryos cultured in vitro. Human oocytes, zygotes, cleavage embryos and blastocysts donated for research were labeled with the tetraspanin EV marker CD9 and analyzed by scanning and transmission electron microscopy. Embryo culture conditioned media collected 3- and 5-days post fertilization were examined for the presence of EVs using electron microscopy. We detected numerous CD9 positive vesicles released from all embryos examined. They were observed on the surface of the plasma membrane, within the perivitelline space as well as throughout the zona pellucida. Interestingly, EVs were not seen in the ZP of all mature metaphase II oocytes, however, were detected just after fertilization in the ZP of zygotes and embryos. Electron microscopy using negative staining, and nanoparticle tracking analysis (NTA) of embryo conditioned culture media also showed the presence of vesicles of various sizes, which were round shaped, and had a lipid bilayer. Their size ranged from 30 to 500 nm, consistent with the sizes of exosomes and microvesicles. In conclusion, the results of the study provide evidence that human preimplantation embryos at all developmental stages secrete EVs into the perivitelline space, which then traverse through the ZP, and are then released into the surrounding culture medium.

Abbreviations: EVs: extracellular vesicles; ZP: zona pellucida; CD9, CD63, and CD81: tetraspanin EV markers; NTA: nanoparticle tracking analysis; ESCRT: endosomal sorting complexes required for transport; SEM: scanning electron microscopy; TEM: transmission electron microscopy; TE: trophectoderm; ICM: inner cell mass; PVS: perivitelline space; MI: metaphase I; MII: metaphase II; GV: germinal vesicle; MVs/EXs: microvesicles/exosomes; hCG: human chorionic gonadotrophin; GnRH: gonadogrophin releasing hormone; ICSI: intracytoplasmic sperm injection; SPS: serum protein substitute; 1PN: one pronuclear zygote; 3PN: tri-pronuclear zygote; IgG: immunoglobulin G; PBS: phosphate buffer saline; ETHO: ethanol; ESED: Environmental Secondary Electron Detector; BSA: bovine serum albumin

KEYWORDS:

• Extracellular vesicles
• CD9 tetraspanin
• human embryos
• conditioned culture media
• electron microscopy

Introduction

Extracellular vesicles (EVs) are highly specific and multi-purpose vesicular structures that are secreted by various cell types of the body (Lee et al. 2012; Andaloussi et al. 2013; Pitt et al. 2016; Kalra et al. 2016; Cruz et al. 2018). They have been identified in many biological fluids in-vivo, including blood, milk, saliva and urine, as well as in-vitro in cell culture conditioned media (Andaloussi et al. 2013; Tkach and Théry 2016). The three main subtypes of EVs are exosomes, microvesicles and apoptotic bodies, and can be distinguished by their biogenesis, presence of a lipid bilayer, specific surface markers and size (Andaloussi et al. 2013). Both, exosomes (30–120 nm) and microvesicles (100–1000 nm) mediate cell-to-cell communication and have been shown to selectively deliver a broad spectrum of proteins, lipids, and nucleic acids to target cells, thereby modifying the cellular activities of recipient cells. (Valadi et al. 2007; Andaloussi et al. 2013; Desrochers et al. 2016; Kalra et al. 2016; Cruz et al. 2018). EVs can be characterized by protein markers, such as tetraspanins (CD9, CD63, and CD81), adhesion integrins and endosomal sorting complexes required for transport (ESCRT). In contrast, apoptotic bodies are secreted by cells undergoing apoptosis and their contents are typically limited to nuclear fragments and cell debris. They do not exhibit surface markers characteristic for exosomes or microvesicles and are not involved in cellular communication (Tannetta et al. 2014; Kalra et al. 2016). EVs have been shown to play an important role in the regulation of normal physiological processes such as maintenance of stem cells, tissue repair, and control of the immune system (Desrochers et al. 2016; Pitt et al. 2016; Tkach and Théry 2016). Disruption of the physiologically normal composition and release of EVs has been shown to lead to, or be indicative of, pathological disorders, such as neoplasia or autoimmune diseases (Lee et al. 2012; Pitt et al. 2016). Therefore, the diverse nature of EVs renders them of great interest for clinical applications, including the development of diagnostic biomarkers as well as for potential therapies (Andaloussi et al. 2013).

There is growing evidence for the possible role of EVs, particularly exosomes, in key reproductive processes including follicular growth, oocyte and sperm maturation, fertilization, embryo implantation, as well as in pre-term labour and pre-eclampsia (Di Pietro 2016; Machtinger et al. 2016; Marin and Scott 2018; Martinez et al. 2018). The fusion of oocyte and sperm during fertilization in mice has been shown to be dependent on the release of CD9 positive vesicles by oocytes into the perivitelline space (PVS) where they bind to sperm promoting fertilization (Miyado et al. 2008). Additionally, CD81 associated vesicles have been reported to be expressed on the surface of mouse oocytes and granulosa cells, and possibly participate in gamete fusion by mediating the sperm acrosome reaction (Tanigawa et al. 2008). In addition, recently it has been shown that EVs containing the Juno protein appear to be shed by oocytes into the perivitelline space and may play a role in preventing polyspermy (Bianchi and Wright 2014; Machtinger et al. 2016).

Moreover, recent data suggest the potential roles of EVs in mediating the embryo-endometrial crosstalk essential for successful implantation (Tannetta et al. 2014; Saadeldin et al. 2015; Machtinger et al. 2016; Homer et al. 2017; Parks et al. 2018). Importantly, there are emerging research studies related to EVs cargo and their roles in facilitating intercellular communication during human embryonic development and embryo-endometrial cross-talk (Marin and Scott 2018). During pregnancy, exosomes and microvesicles are present in the uterine environment and are thought to be secreted by the apical surface of the endometrium (Ng et al. 2013; Homer et al. 2017). As these EVs are found to contain miRNAs it has been proposed that they enhance the adhesive capacity of the blastocyst and enable the invasive potential of trophoblast cells, thereby enabling implantation. (Ng et al. 2013; Greening et al. 2016).

Although EVs have been shown to be released from many different cell and tissue types in the body, the secretion of EVs from mammalian, especially human embryos has not been well characterized. Thus, the aim of the study was to 1) identify EVs in human preimplantation embryos at different stages of their development using scanning and electron microscopy, and 2) investigate whether EVs can cross the ZP and be released from human embryos cultured in vitro.

Results and discussion

Scanning electron microscopy (SEM) analysis of early cleavage embryos (2- to 9-cell) showed the presence of CD9 positive vesicular structures on the surface of the embryonic membrane (Figure 1). Numerous small or large groups of spherical CD9 vesicles of different sizes were seen dispersed on the embryonic plasma membrane, suggesting that these vesicles are released by human embryos. To localize the vesicles, view vesicular morphology in greater detail, and determine whether EVs can cross the ZP, transmission electron microscopy (TEM) images were taken of embryos from different stages of preimplantation embryo development, including 1-cell zygotes, cleavage embryos (2-cell, 4-cell, and 8–10-cell), morulae, and blastocysts, including a compromised blastocyst which only contained the trophectoderm layer (TE) and no inner cell mass (ICM).

Figure 1. SEM images of day 3 cleavage embryo. A and B show CD9-positive 20 nm immuno-gold labeling on two different areas of the same embryo. The embryonic membrane is shown in green while the red shows CD9 labeling.

Strong CD9 signals were observed in the EVs in all examined embryos (Figure 2). Positively labeled, bilayer vesicles were seen budding off the embryonic plasma membrane into the PVS as single or multiple vesicle clusters (Figure 3). EVs were also detected throughout the ZP from the inner to its external surface (Figure 2). Although the microscopic images of blastomeres’ cytoplasm appeared to be very dense and filled with different organelles, some large groups of CD9-immuno-gold positive vesicles were also observed within embryonic blastomeres just beneath the plasma membrane (Figure 3). Moreover, numerous CD9 positive EVs were also observed within the ZP of ‘abortive’ blastocysts which did not develop an ICM, suggesting that trophectoderm cells actively contribute to the EV population produced by human embryos. In fact, the trophoblast cells obtained from early human placenta secrete membranous EVs, of the nanometer range, into culture, suggesting their potential role in modulation of the maternal immune system (Kshirsagar et al. 2012). Additional evidence in the porcine model shows that EVs are released by both the endometrial and trophectoderm cell lines and are involved in bidirectional cell-to-cell communication. Such EV transportation between both cell types strongly supports the idea that EVs play an important biological role in conceptus-endometrial cross-talk crucial for a successful pregnancy (Bidarimath et al. 2016).

Figure 2. TEM images of embryos of different embryonic stages labeled with anti-CD9 5nm immuno-gold particles. 1 cell zygote (A-C); 3-cell embryo (D-F); 5-cell embryo (G); 8–10 cell embryo (H); blastocyst with no ICM (I). The images show double membrane EVs in the PVS and in the ZP.

Figure 3. TEM images of a day 3 embryo labeled with 5nm immuno-gold antibodies against CD9. In A clusters of EVs are seen in the PVS, in the embryo under the PVS, and throughout the ZP (scattered black dots). Higher magnification of the ZP in B shows CD9 positive vesicles from A. Immuno-gold labeling is highlighted with arrows.

To date, there is no specific marker for different types of EVs; however, some tetraspanins (CD9, CD63, and CD81) and members of the ESCRT machinery have been reported to be characteristic for EVs and particularly present in exosomes (HardiHarding et al. 2013; Kowal et al. 2016). Therefore, the strong CD9 positive staining of human embryos, as observed in the present study, suggests that these spherical structures are either exosomes or microvesicles, but not apoptotic bodies (Tannetta et al. 2014; Kalra et al. 2016). Abundance of CD9-containing ‘exosome’ like vesicles, bound to gold particles was also identified in the PVS of mouse oocytes by electron microscopy analysis as well as by immunostaining in the oocyte membrane (Miyado et al. 2008). Accumulation of such vesicles was seen during oocyte meiotic maturation, except during the germinal vesicle stage, and it was proposed that their release from oocytes before fertilization facilitates sperm–oocyte interactions. In other animal studies, strong immunofluorescent signal of the CD9 tetraspanin molecule was also observed during meiotic stages on the membrane of mouse and porcine oocytes; and has been found to be essential for sperm-oocyte fusion (Chen et al. 1999; Le Naour et al. 2000; Miller et al. 2000; Li et al. 2004). Similarly, using immunostaining, CD9 positive vesicles have been found uniformly distributed on the oolemma of human zona-intact or zona-free GV oocytes, as well as on metaphase I and metaphase II oocytes (MI and MII; Coskun et al. 2003; Ziyyat et al. 2006).

Interestingly, in the present study, although CD9 positive vesicles were observed in the PVS of MII oocytes, they were not observed within the ZP (Figure 4). Similar findings have been reported in mice, where extracellular vesicles seem to be absent from the ZP of mature MII oocytes (Miller et al. 2000). The synthesis of CD9 protein on the oocyte membrane has been well characterized in mouse and porcine models, where the abundance of the CD9 protein gradually increases as the oocyte matures from germinal vesicle (GV) to MII (Chen et al. 1999; Li et al. 2004). Our results, which show that although EVs were absent from the ZP of MII oocytes, there was already an abundance of EVs present in the ZP of a 1-cell stage zygote, suggesting that intensive secretion of EVs begins shortly after fertilization and continues throughout cleavage to the blastocyst stage (Figures 3,4).

Figure 4. Representative images of MII oocytes labeled with anti-CD9 5nm immuno-gold particles. EVs are present in the oocyte microvilli (black dots) but are not seen in the ZP.

Indeed, we confirmed, by electron microscopy analysis using negative staining, that embryo-derived EVs are present in embryo conditioned culture media. The images from pooled samples of day 3 and day 5 conditioned media showed the presence of CD9 immuno-gold, vesicles of various sizes, ranging from 50 to 500 nm (Figure 5). Additional NTA analysis indicated that the majority of both, day 3 and day 5/6 embryo-derived EVs were of similar diameter ranging from 100 to 200 nm, consistent with the size of exosomes and microvesicles (Figure 6). The results also indicated a considerably higher concentration of EVs in conditioned culture media collected from blastocysts (7.89 ± 0.3x10⁸ particles/ml) compared to those from cleavage stage embryos (7.0 ± 3.3x10⁸ particles/ml). The EVs were detected as well in unconditioned culture media. The concentration of EVs in these unconditioned media appeared to be slightly higher (7.33 ± 2.4x10⁸ particles/ml) than in day 3 conditioned samples, which may suggest uptake of EVs by the embryos. Interestingly, other reports, in animal models and human, have shown that the transport of exosomes and microvesicles through the ZP may indeed be bidirectional. EVs present in the surrounding media have been shown to traverse through the ZP and be internalized by blastomeres of porcine preimplantation embryos. (Saadeldin et al. 2014). Additionally, a human study reported a high abundance of miR-645 in unconditioned control media, compared to cultured media. It is proposed that this miRNA may be contained within EVs which traverse through the ZP and are internalized and used by healthy human embryos (Rosenbluth et al. 2014). Several other reports have also observed the presence of EVs in unconditioned media (Tannetta et al. 2014, Giacomini et al. 2017) and the origins of these vesicles is proposed to be from protein supplements routinely used in embryo culture media such as albumin and α, β, and γ-globulins. Nonetheless, techniques used to characterize embryo-derived EVs remain still largely empirical and the isolation of embryo-derived EVs from those of non-embryonic origin, from a small volume of fluid (20–30 ul) is the most common technical challenge awaiting improvements (Nguyen et al. 2016).

Figure 5. Negative staining of EVs present in embryo conditioned culture media after 3 (A) and 5 days (B, C) of culture. The size of EVs (57-190 nm) is shown for each vesicle.

Figure 6. Results of nanoparticle tracking analysis from day 3 and day 5–6 conditioned culture media and unconditioned controls.

To date, only three recent human studies were performed using TEM and/or flow cytometry to describe the size, number and markers of EVs in order to explore their potential as reproductive biomarkers (Abu-Halima et al. 2017; Pallinger et al. 2017; Giacomini et al. 2017). It has been demonstrated that human preimplantation embryos at different developmental stages can release CD9 and CD63 positive EVs in size range from 50 to 200 nm with exosomes as the predominant vesicle type, exhibiting much greater number and variability by day 5 than day 3 embryos (Giacomini et al. 2017). These experiments also showed the uptake of embryo-derived EVs by human endometrial epithelial and stromal cells suggesting possible way of communication at the maternal–fetal interface. Other studies also indicated a possible link between the decreased miRNA variety and decreased EV quantity in the spent culture media from embryos leading to successful pregnancy compared to those which failed (Abu-Halima et al. 2017; Pallinger et al. 2017). Similarly, the latest TEM experiments in animal models have also demonstrated that bovine blastocysts secret a heterogeneous population of EVs (Exs/MVs) into the culture media which vary depending on embryo quality and competency (Mellisho et al. 2017).

Although the presence of extracellular vesicles in spent culture media has already been demonstrated in animal models and in human; evidence of ultrastructural imaging has not previously been reported. Here we showed for the first time (to the best of our knowledge), ultrastructural analysis using SEM/TEM, that human preimplantation embryos at different developmental stages produce and secret EVs, that transport across the ZP into the surrounding culture medium during growth in vitro.

Materials and methods

Material collection, ethical approval, and laboratory procedures

Ethics approval for this project was obtained from the Health Sciences Research Ethics Board of the University of Toronto (Ref 30251). Informed consent for oocyte and embryo donation and use of conditional media from cultured embryos was obtained from all individuals participating in the study. Controlled ovarian hyperstimulation was carried out with the use of a standard antagonist protocol (Engmann et al. 2016), and laboratory procedures including oocyte retrieval, sperm preparation, ICSI as well as embryo culture were performed as previously described (Balakier et al. 2016).

Study materials

All donated human oocytes and embryos for this study were unsuitable for use in infertility treatment. The study material consisted of 1) metaphase II oocytes obtained from germinal vesicles and metaphase I oocytes after 24 h of in vitro culture; 2) abnormal zygotes bearing one or three pronuclei (1PN or 3PN); 3) cleavage stage embryos derived from those abnormally fertilized oocytes; 4) arrested day 5–6 morulae; and 5) abortive blastocysts containing low cell number in the inner cell mass and trophectoderm, and blastocysts without the ICM.

Antibodies

To detect extracellular vesicles in human oocytes (n = 6) and embryos (n = 45), scanning and transmission electron microscopy was performed using specific EVs marker: CD9 monoclonal anti-human primary antibodies (Cedarlane). The Goat Anti-Mouse IgG antibody conjugated with 5 nm or 20 nm immune-gold particles were used as the secondary antibodies (Cedarlane).

Scanning electron microscopy (SEM)

Embryos were washed 3X with PBS and fixed overnight at room temperature with 0.15% gluteraldehyde and 4% paraformaldehyde in PBS. The zona pellucida (ZP) was removed using Acidic Tyrode’s Solution, according to manufacturer’s instructions (Sigma-Aldrich). Then, embryos were labeled with CD9 primary antibody for 30 min at room temperature (1:100), followed by labeling with 20 nm immune-gold secondary antibody for 1 hour at room temperature (1:200). After a series of alcohol dehydration steps (ETHO: 50%, 70%, 80%, 90%, 95%, 100%), the samples were processed in a critical point dryer and then carbon coated. Imaging was performed with a Hitachi S-570 SEM, equipped with the Environmental Secondary Electron Detector (ESED).

Transmission electron microscopy (TEM)

Oocytes and embryos were washed 3X with PBS and fixed overnight with 0.15% gluteraldehyde and 4% paraformaldehyde in PBS. Zona permeabilization was achieved using 0.1% Tween-20 incubation for 10 min at room temperature. Embryos were then embedded in LR White Embedding Media (Electron Microscopy Sciences) and incubated overnight at 60°C. Sections were cut using a microtome and placed on nickel grids. The grids were labeled with CD9 primary antibodies by placing them on top of a series of 100 μL droplets consisting of the following: 1) Blocking using 0.05% PBS-BSA for 3 × 5 min, followed by 3 × 30-s washes with ultra-pure water; 2) primary CD9 antibody labeling at a concentration of 1:100 in 0.05% PBS-BSA for 1.5 h, followed by 3 × 30-s washes with ultra-pure water; 3) Blocking using 0.05% PBS-BSA for 5 min, followed by 3 × 30-s washes with ultra-pure water; 4) secondary antibody labeling with 5 nm immune-gold particles for 45 min followed by 3 × 30-s washes with ultra-pure water, 5) 2% Uranyl acetate stain for 30 s followed by 3 × 30-s washes with ultra-pure water; and 6) stained with 4% osmium tetroxide (OsO₄). The stained grids were then imaged using the Hitachi H-700 TEM.

Embryo culture conditioned media

Embryo culture conditioned media were individually collected after embryo removal (approximately 20 μL) and centrifuged at 300 x g for 10 min, 2000 x g for 10 min, 12,000 x g for 30 min and stored in −80°C until needed for experimentation. Simultaneously, unconditioned 20 uL media droplets, that were cultured in vitro under the same gas and temperature environments as conditioned media but not exposed to an embryo, were also processed as a negative control. A total of 118 samples were collected (40 spent media from day 3 and 78 from day 5/6).

Negative staining

For negative staining, samples were thawed for 30 min at room temperature and then pooled. Following preliminary centrifugation (300 x g for 10 min, 2000 x g for 10 min, 12,000 x g for 30 min), the supernatants were placed in polycarbonate thick-walled ultracentrifugation tubes (Beckman) and centrifuged at 49 000 rpm for 3 h at 4°C. After removing the supernatant, 10 μL of fixative was added to the pellet. In cases where there was no visible pellet, 5 μL of media collected from the bottom of the tube was fixed with 10uL of fixative comprised of 0.15% gluteraldehyde and 4% paraformaldehyde in PBS, after the rest of the supernatant was removed. After incubation overnight at 4°C, 1 μL of the solution was placed on a copper grid, along with 1 μL of negative stain. The grids were visualized with a Hitachi-H 700 TEM.

Nanoparticle tracking analysis (NTA)

Nanoparticle tracking analysis was used to characterize the population of EVs secreted to the culture media by day 3 and day5/6 embryos. Analysis was performed on pools of 10 droplets, of 25ul microdroplets, in each group. Pooled droplets were centrifuged to remove dead cells and debris (200 x g for 5 min, 2000 x g for 10 min, 14,000 x g for 30 min). 750 uL of PBS was added to the supernatant to reach a final volume of 1000 uL. Samples were analyzed using the NanoSight LM10 (Malvern Panalytical). Each sample was measured in triplicate with the same camera settings, acquisition time of 30 s and detection threshold of 8.

Authors contributions

Designed the study, performed experiments, analyzed the data, and assisted with manuscript writing: PV. Collected and analyzed the data, wrote the manuscript: HB. Medical Director of Create Fertility Centre, and critical revision of the manuscript: CL. All authors have reviewed and approved the final version of the manuscript.

Acknowledgments

The authors wish to thank Dr. Battista Calvieri, Dr. Steven Doyle, and Yan Chen for their assistance with electron microscopy (Microscopy Imaging Laboratory, Faculty of Medicine at the University of Toronto). We are also indebted to the patients who generously donated for this study their oocytes and embryos which were unsuitable for infertility treatment.

Disclosure statement

No potential conflict of interest was reported by the authors.