Here, we present a protocol for non-invasive assessment of oocyte developmental competence performed during their in vitro maturation from the germinal vesicle to the metaphase II stage. This method combines time-lapse imaging with particle image velocimetry (PIV) and neural network analyses.
The overall goal of this procedure is the identification of occytes with a backside developmental capacity using a non-invasive tool. The main advantage of this method is that by simply observing the oocyte movements occurring during in vitro maturation, one may evaluate the female gamete developmental potential. This technique is base on time-lapse imaging coupled with particle image velocimetry, a neural network based analysis.
Here, we will describe the steps taken to demonstrate that mouse oocytes during the transition from the germinal vesical to metaphase two stage, display a different profile of said plasmic movements significantly correlated to their developmental competence or incompetence. According with the World Health Organization, infertility is a pathology that affects about 20 percent of couples and in addition, one third of the women undergoing cancer treatment are at risk of a premature ovarian failure. This corresponds to about 300, 000 or 30, 000 women in countries like U.S.A or Italy respectively.
The design of non-invasive procedures for the identification of the oocyte developmental competence would help to improve the overall success of pregnancy outcomes. Here, we give details of the most critical steps of the procedure as developed for the mouse, in order to make it readily available to be tested and used for other mammalian species such as bovine or humans. Ovaries are transferred in a Petrie dish containing M2 medium pre-warmed and equilibrated and 37 degrees centigrade and five percent CO2.
If the oviduct and some fat are still present around the ovary, they ought to be removed. Then, holding the ovary with a pair of tweezers, it's surface is punctured using a sterile 1G needle. The punctured ovaries are removed from the M2 medium and, using a hand-pulled micropipette, cox our first move to a new clean M2 medium, and then pipetted in and out for a few seconds until the cumulus cells surrounding the oocyte are completely removed.
Next, oocytes are transferred in 20 microliter drops of M2 medium and if some cumulus cells are persisting, they are thoroughly removed through several passages from drop to drop until cumulus cells are completely eliminated. In order to prepare Hoechst 33342 at a final concentration of 0.05 microgram per mil, we begin with a mild solution of Hoechst at concentration of five miligram per mil, prepared in one-time pbs and this is diluted further until the required concentration is obtained. 3.5 microliter microdrops of Hoechst solution are placed at the bottom of a Petrie dish lid and each single diluted oocyte is then transferred in a single Hoechst drop, being careful to reduce light exposition during the Hoechst staining process, by covering the dish with a dark lid.
After ten minutes staining, oocytes are observed under a fluorescent microscope with UV light. Based on the chromatin conformation, they are classified either as Surrounded Nucleolus, SN, if they present a ring of x-positive chromatin around the nucleolus or as Not-Surrounded Nucleolus, NSN, if the nucleolus lacks a ring of x-positive chromatin and the overall chromatin appears dispersed in the nucleolus. Four two-microliter drops of alpha mem medium are laid onto a glass-bottom Petrie dish being careful to keep the drops within a specific distance, drops are overlaid with pre-warmed and equlibrated mineral oil and a total of three to four oocytes are transferred to each drop.
The Petrie dish is transferred into the incubation chamber of a time-lapsed microscope, which allows prolonged cylature maintenance of temperature at 37 degrees centigrade and five percent CO2 and illumination with red LED light. Prior to time-lapse analysis, using the BioStation software, the coordinates of each drop are set in order to automate the passage from one drop to the next during the time-lapse routine. Time-lapse frames are taken at eight-minute intervals for a total of 15 hours.
The recorded movies are then analyzed using the CellPave software, which allows the detection of cytoplasmic movements expressed in term of velocity vectors. First, with the CellPave tool, the perimeter of each oocyte is manually delimited. Then, the software calculates the cytoplasmic movement's velocity occurring in the comparison between two consecutives time-lapse frames and expresses it by placing a colored arrow whom color intensity and length defines the speed and direction of movements.
The histogram underneath the oocyte seen here describes the overall cytoplasmic movement's velocity throughout the 15 hour recording period of an SN oocyte. The same type of analysis is made on an NSN oocyte. Notice the delayed polar body one extrusion peak shown by NSN oocyte compared to SN oocyte.
The bottom of the screen shows the temporal pattern of the oocyte's cytoplasmic movement velocity. The velocity is calculated by comparing two consecutive frames. The raw data of cytoplasmic movement velocity are then exported to an Excel sheet reporting, on the rows, the Frens sequence, and on the columns, the single oocyte analyzed.
On the left-hand side, the data of an SN oocyte, whereas on the right-hand side, those of NSN oocytes. The data of either SN or NSN oocytes are then split into 10 sub-groups. Nine are used for training, green color.
And one for the blind testing, red color. This procedure is repeated for a total of 10 times. Each time changing the blind testing sample.
The data are then examined through a feed-forward artificial neural network, or FANN. The FANN is built to analyze the whole time-lapse frames, here named Input Neurons"The latter are connected to three hidden neurons, themselves linked to two output neurons. One output gives the probability that the input is a competent oocyte.
The other, on the contrary, that the input is an incompetent oocyte. The comparison between the two outputs predicts the oocyte dula mental competence with an accuracy of 91.03 percent. The technique has been set-up for mouse oocytes as shown here.
However, the future recall is to test it on human oocytes. In fact, that non-invasive procedures from the analysis of oocyte quality are particularly required in the field of artificial reproductive technologies.
