Human Embryonic Behavior Observed with Time-Lapse Cinematography

Assisted reproductive technologies (ART) have enabled the in vitro observation of a human embryo as it develops; however, evaluating the morphokinetics of embryonic development in still images remains difficult. In 1997, Payne et al. [1] developed time-lapse video cinematography (TLC) to overcome the limitations of intermittent observation by providing high-resolution, continuous imaging, in which various cellular events could be distinguished in real time and monitored during the observation period. However, this system was restricted to an observation period of 17-20 hours. Based on Payne’s work, we developed a new TLC system that maintains optimal and stable culture conditions on the microscope stage for periods of up to a week [2,3]. Using this system, we have observed human embryonic development from the fertilization process to the hatched blastocyst stage. Herein, we present several novel aspects and events identified using our system during that developmental period.


Introduction
Assisted reproductive technologies (ART) have enabled the in vitro observation of a human embryo as it develops; however, evaluating the morphokinetics of embryonic development in still images remains difficult. In 1997, Payne et al. [1] developed time-lapse video cinematography (TLC) to overcome the limitations of intermittent observation by providing high-resolution, continuous imaging, in which various cellular events could be distinguished in real time and monitored during the observation period. However, this system was restricted to an observation period of 17-20 hours. Based on Payne's work, we developed a new TLC system that maintains optimal and stable culture conditions on the microscope stage for periods of up to a week [2,3]. Using this system, we have observed human embryonic development from the fertilization process to the hatched blastocyst stage. Herein, we present several novel aspects and events identified using our system during that developmental period.

Materials and Methods
We developed a new system for TLC, based on the concept of Payne et al. [1][2][3], wherein the optimal embryo culture conditions (temperature: 37 ± 0.3°C, pH: 7.45 ± 0.03) are maintained by continuously monitoring and adjusting the flow of CO 2 (40 mL/ min), the air temperature in the large chamber (38.0°C), and the thermoplate temperature on the microscope stage (41.8°C). The inverted microscope is equipped with a CCD digital camera (Roper Scientific Photometrics, Tuscon, AZ) connected to the computer and display by Meta Morph (Universal Imaging Co, Downingtown, PA). In this study, digital images of the cultured embryos were acquired for 2-5 days using an exposure time of 50 ms. In total, 2000-8000 images were taken during the observation period. The time interval of image storage was 10 sec during fertilization and 2 min thereafter until the end of the observation.
In this study, we used our TLC system to observe 1) oocytes for clinical use and 2) embryos for research work. In oocytes for clinical use, after receiving informed consent from the couples, one oocyte out of 10 or more oocytes retrieved in each patient was randomly selected for TLC observation.
In oocytes derived by in vitro fertilization (IVF), the cumulus cells were mechanically removed so as not to damage the tail of the sperm that had penetrated the zona pellucida (ZP) at one hour after insemination with 50,000 of the husband's motile sperm. The oocyte was transferred into a microdrop of culture medium (5 μl, Fertilization medium ® , COOK, Australia) covered by mineral oil (SAGE, USA), and the site of sperm penetration into the ZP was carefully identified by micromanipulation (Narishige, Japan). The microscope was focussed on the sperm within the ZP that was penetrated most deeply and TLC observations were carried out at 10-second intervals until the leading sperm penetrated ZP and attached to the oocyte membrane. Thereafter, TLC was continued for approximately 40 hours at 2-minute intervals. Oocytes for intracytoplasmic sperm injection (ICSI) were transferred immediately into a microdrop of culture medium covered by mineral oil and TLC observation was commenced for up to 40 hours at 2-minute time intervals. At the end of the observation period, goodquality embryos developed in both cIVF and ICSI were cryopreserved for future clinical use.
On the other hand, the donated frozen/thawed day-2, -3 embryos for research work were thawed and cultured for 3 to 5 hours to assess viability and embryo quality. After the assessment of embryos, the earlystage, good-quality embryos were transferred into 5 μl of the culture medium covered by mineral oil and TLC observation was commenced for up to 5 days at 2-to 5-minute time intervals. These TLC observations were approved by the ethical committee of JISART (Japanese Institution for Standardizing Assisted Reproductive Technology).

Fertilization process (representative oocyte) in IVF (Supplementary Movie 1)
human oocyte following IVF and up to second cleavage by TLC [2,3]. This initial study also revealed several novel phenomena occurring during fertilization and cleavage [4]. In this study we further analyzed these novel phenomena by TLC.

Fertilization cone (FC) (Supplementary Movie 2):
In other species including echinoderms [5,6] and rodents [7,8], the FC first appears during the fertilization process wherein it plays an important role. In the sea urchin oocytes, the FC was closely associated with the polyspermic block through depolarization of the oocyte membrane [5,6], while in rodent oocytes, the FC was detectable from 2 to 3 hours postinsemination prior to the decondensation of sperm chromatin, and it disappeared after the decondensation was complete.
By TLC analyses of the human fertilization process, we confirmed the appearance of FC at the sperm entry point (SEP), 30 minutes after extrusion of the second polar body (PB) [2,3], and then evaluated the fertilization time course from insemination to disappearance of the FC ( Figure 1). Considering that oocytes showing an FC were more likely to develop into a good quality embryo (GQE) than those without FC, the appearance of FC may be associated with embryo quality (Table 1). Such physiological effects should be studied further in human fertilization.

Cytoplasmic flare (Flare) (Supplementary Movie 3):
After introduction of the sperm centrosome following sperm incorporation, the radial sperm aster was formed by the polymerization of microtubules adjacent to the decondensing male pronucleus (mPN). The female pronucleus (fPN) then migrated toward the mPN and both PN abutted [9]. Payne et al. [1] first showed by time-lapse video recording that the Flare radiated out from the center of the oocyte with ICSI. In our TLC, the Flare was clearly observed emanating from the site where the FC disappeared and migrating toward the ooplasm cortex. Then, the fPN moved toward the mPN until the two PN abutted ( Figure 2). We then examined the relationship between appearance of the Flare and movement of the fPN toward the mPN in oocytes with IVF (n=29) and ICSI (n=66) (Figure 3). In the 25 IVF oocytes showing normal fertilization (2PN/2PB) and the 4 fertilized abnormally (>3PN/2PB), the Flare appeared and the fPN moved toward the mPN. In the ICSI oocytes, 63 of the 66 showed the Flare and the remaining did not, and in those with the Flare, 59 showed normal fertilization (2PN/2PB) and 4 showed abnormal fertilization. Thus, all IVF and ICSI oocytes with the Flare showed migration of the fPN toward the mPN, while three ICSI oocytes without the Flare showed neither both PNs nor normal fertilization. The Flare therefore seemed to represent a visual manifestation of the sperm aster.

Nucleolar precursor body (NPB) (Supplementary Movie 4):
The nucleolar precursor bodies (NPBs) are condensed components responsible for ribosomal assembly [10] and the production of some   growth factors and developmental regulatory proteins [11]. NPBs are therefore likely to play important biological roles in early embryonic development. Several studies have correlated NPBs with embryo quality [12][13][14]. However, our TLC analysis revealed the distributions of NPBs to be quite dynamic with polarity changes from time to time, which made it extremly difficult to predict embryo quality based on NPBs in a single observation. Therefore, we evaluated the number of NPBs just after the PN appeared and before it disappeared. Just after the PN appeared, the average NPB number in GQEs and PQEs was 7.5 and 6.6 in male PN (mPN) and 6.1 and 4.9 in female PN (fPN), respectively, whereas immediately before the PN disappeared, NPBs in GQE and PQE were 7.1 and 6.8 in mPN, and 5.6 and 4.9 in fPN, respectively. Thus, we found no correlation between NPB number and embryo quality. However, the average number of NPB in mPN was higher than that in fPN [1]. These results suggest that evaluation of NPBs is not a useful biomarker of embryonic quality.

Translucent zone in peripheral ooplasm (cytoplasmic halo) (Supplementary Movie 5):
The mPN and fPN was formed at 6.6 hours and 6.8 hours, respectively, after fertilization [3]. As they migrated toward the center of ooplasm and abutted (8.8 hours after the insemination), a peripheral cytoplasmic translucent zone called the cytoplasmic halo was formed. Prior to syngamy, the halo disappeared and the first cleavage division commenced. This cytoplasmic halo is thought to be the consequence of microtubule-mediated withdrawal of mitochondria and other cytoplasmic components to the perinuclear regions [15], and this hypothesis matches a sequence of cytoplasmic dynamics captured by TLC. Movie 6): Our TLC observation of ICSI oocytes revealed for the first time a novel phenomenon in which a third PB-like material was extruded in the perivitteline space shortly after extrusion of the second PB ( Figure  4). These oocytes developed into single-PN zygotes and poor-quality, cleaved embryos. Until now, several mechanisms have been proposed for the formation of the 1PN zygote after IVF or ICSI, as follows: 1) parthenogenetic activation [16][17][18]20], 2) asynchronous pronucleus formation [18][19][20], and 3) early fusion prior to syngamy [21]. However, our TLC study demonstrated a novel candidate mechanism in which ICSI oocytes extruded all chromosomal material including the fPN, leaving only mPN injected. Indeed, this is quite an aberrant phenomenon, and there is no background information to support such an occurrence. However, we observed this eventuality in two separate movies by TLC.

Uneven 2PN zygote (Supplementary Movie 7):
Our TLC analyses also revealed other aberrant phenomena in the formation of uneven 2PN zygotes after ICSI. As seen in the 1PN zygote formed after ICSI, a third PB-like matter was extruded in addition to the first and second PB, and it appeared to contain a small PN-like body, formed after the mPN formed in the ooplasm. The PN-like body was then absorbed into the ooplasm and migrated toward the mPN, leading to formation of an uneven 2PN embryo ( Figure 5). Although it has been reported that zygotes with uneven 2PN may correlate with ooplasmic immaturity and chromosomal abnormalities [22,23], the etiology and physiological significance of those zygotes are still unknown. The TLC analyses in this study demonstrated for the first time a possible mechanism for the formation of uneven 2PN zygotes.

Fragmentation (Supplementary Movie 8)
It is well known that embryo fragmentation or the generation of anucleate cell fragments is an important biomarker of embryo quality during early embryonic development after ART [24,25]. Several hypotheses regarding embryo fragmentation have been reported such as programmed cell death (apoptosis) [26], telomere length [27], increased maternal age [28], abnormalities in cytoskeleton and microtubule organization [29,30]. However, the precise underlying mechanism remains controversial and we considered our TLC observation to be a powerful tool to elucidate the dynamic morphology of this phenomenon.
We obtained one of the best TLC movies showing embryo fragmentation after ICSI. After the cytoplasmic flare appeared from the center of the ooplasm toward the cortex, mPN was formed in the center of the ooplasm, followed by the appearance of fPN just beneath the first and second PB. The fPN migrated quickly toward the mPN and abutted. In 17 hours and 20 minutes after the commencement of TLC observation, this zygote reached syngamy. Two hours later, first cleavage started and was completed within 30 minutes. During the division, a few fragments appeared along the cleavage furrow and   then decreased in number as the time passed. At 31 hours of TLC observation, the second cleavage commenced and the embryonic morphology dramatically changed into poor quality by 31 hours and 48 minutes. However, at 33 hours, the blastomere shape changed into good quality. More fragments were observed during the second cleavage process than during the first cleavage ( Figure 6).
In addition, we also confirmed that fragmentation occurred during cytokinesis and was not observed at any other time during the cell cycle. Some of the fragments were resorbed into the blastomere in the first 2 hours following both first and second cleavage (Figures 7 and 8).

Blastocoel collapse (Supplementary Movie 9)
Using donated frozen/thawed 4-cell embryos (IVF; n=21, ICSI; Figure 5: Novel mechanism of uneven 2PN zygote after ICSI. After the second PB was extruded, a third PB-like material was also extruded. At the same time as mPN formation in the ooplasm, a small fPN-like body was also formed within the third PB-like material and then resorbed into the ooplasm to migrate toward the mPN.   n=13), TLC observation was carried out for up to 5 days. One of these TLC images showed a novel mechanism for blastocoel collapse. During the expanding blastocyst stage, a localized part of trophectoderm (TE) became suddenly disturbed followed by blastocoel collapse. After that event, while the TE was repaired, the blasotocoel cavity was quickly re-expanded ( Figure 9). Furthermore, we compared the time course of two typical embryos; one was a hatched embryo and the other was an unhatched embryo showing degeneration ( Figure 10). The hatched embryo developed slowly and steadily without the collapse of TE until hatching was complete. Contrarily, the unhatched embryo developed quickly, showed a number of blastocoel collapses, and finally degenerated. We compared the relationship between the number of collapses and culture outcome in two types of blastocysts [hatched (hBL) vs. unhatched (uBL)] (Figure 11), and found a significantly higher number in uBL than in hBL (P<0.01). Therefore, although the nature of normality and/or differences in blastocoel collapse between in vivo and in vitro conditions remain controversial [25,26], we concluded that the blastocoel collapse is a super-physiological phenomenon and may be indeed be an artifact of extended culture in vitro.

Hatching process (Supplementary Movie 10 and 11)
Escape from the ZP of the expanded blastocyst (hatching) is an epilogue in embryonic development before implantation. Although several hypotheses have been proposed regarding the mechanism of hatching in mouse, bovine, equine, and human [33][34][35], only limited information exists about the entire process. Using donated day-2 frozen/thawed human embryos, we looked at embryonic development up to hatching by TLC [2,3]. In this analyses, we found another novel phenomenon regarding the hatching pattern of human expanded blastocysts in that two patterns exist, which we named the "inward" and "outward" patterns. In the "inward" pattern, after the blastocyst reached the fully expanded stage, the blastocoel cavity suddenly collapsed and, at the same time, the ZP retracted inwardly followed by a break in the ZP, through which the blastocyst then escaped (Figure 12).
In the "outward" pattern, after the blastocoel cavity expanded steadily without any blastocoel collapse, the ZP was broken and the blastocyst escaped quickly from the break site ( Figure 13). Of the expanded blastocysts that escaped successfully from the ZP (n=33), only 9 (27.3%) showed no blastocoel collapse when the blastocysts went through the ZP, and this occurrance was irrespective of the grade and frequency of blastocoel collapse. In addition, the time required for escape from the ZP in blastocysts without blastocoel collapse was significantly less than that in blastocysts showing collapse (P<0.05) ( Table 2). Blastocysts with fewer blastocoel collapses could escape from the ZP quicker than those with an increased number of collapses.
Although most of the expanded blastocysts escaped from the ZP showed blastocoel collapse under our in vitro culture conditions, this result suggested that blastocysts in vivo are more likely to follow the "outward" pattern, because in vitro culture conditions are still more consistent than the in vivo situation.
In conclusion, herein we presented novel in vitro data and information regarding human embryonic development observed by using TLC. This microscopy system enables precise visual analyses of human embryonic behavior, and the knowledge derived from using the system provides novel and important information about both the physiology of embryonic development and the quality of in vitro culture conditions to establish safer and more effective ART programs.   6.1 ± 2.8* a 7.5 ± 2.8 9.9 ± 3.1* b p<0.05 Major collapse indicated ≧ 30% of blastocoel cavity