This study assessed 131 human germinal vesicle (GV) oocytes recovered from stimulated cycles to evaluate the relationship between morphometric, morphologic parameters and chromatin configuration. The oocytes were examined using contrast phase microscopy to measure size, nuclear characteristics and the nucleolus-like body (NLB). Chromatin configuration was determined using fluorescent DNA staining. The oocytes were retrospectively grouped into models based on chromatin condensation and distribution relative to the NLB. Results showed that nucleoplasm appearance, nucleus position, nuclear envelope continuity and oocyte size could predict chromatin condensation stage, and were used to develop a mathematical model for forecasting chromatin status based on noninvasive contrast phase microscopy.

1. Morphologic indicators predict the stage of chromatin
condensation of human germinal vesicle oocytes
recovered from stimulated cycles
Laura Escrich, Ph.D.,a Noelia Grau, Ph.D.,a Marcos Meseguer, Ph.D.,a Antonio Pellicer, M.D., Ph.D.,a,b,c
ıa-Jos Escrib, Ph.D.a
and Mar e a
a
Valencia University Institute of IVI; b Department of Obstetrics and Gynaecology, Dr. Peset Hospital; and c Department of
Paediatrics, Obstetrics, and Gynaecology, University School of Medicine, Valencia University, Valencia, Spain
Objective: To assess germinal vesicles (GV) recovered from stimulated cycles by means of morphometric and
morphologic examination (using contrast-phase and image analysis) and chromatin configuration (using fluores-
cent DNA imaging), and to evaluate the relevance of morphometric and morphologic parameters as forecasters
of chromatin status.
Design: Experimental study.
Setting: University-affiliated infertility clinic.
Patient(s): One hundred and thirty-one GV oocytes donated to patients for intracytoplasmic sperm injection.
Intervention(s): We evaluated 131 GVs by means of morphology and morphometry with the use of contrast phase
microscopy. They were subsequently fixed, DNA stained, and assessed by fluorescent microscopy. Compiled data
were retrospectively grouped according to three models.
Main Outcome Measure(s): Model A: ova were grouped according to chromatin condensation (noncondensed vs.
condensed). Model B: ova were grouped according to chromatin distribution in relation to the nucleolus-like body
(NLB) (not surrounding vs. surrounding and/or absent) but regardless of the condensation stage. Model C: GV
oocytes were grouped according to the combination of both of the previously mentioned parameters (chromatin
condensation and distribution in relation to the NLB).
Result(s): According to the GV classification of model A, nucleoplasm, nucleus position, nuclear envelope
continuity, and oocyte size were shown to be relevant and were included in a mathematical model for predicting
chromatin condensation stage.
Conclusion(s): Noninvasive analysis of GVoocytes using contrast-phase microscopy maintains oocytes in a viable
state and allows the chromatin condensation status to be predicted. (Fertil SterilÒ 2010;93:2557–64. Ó2010 by
American Society for Reproductive Medicine.)
Key Words: Chromatin configuration, germinal vesicle, morphologic predictors, nuclear envelope continuity,
nucleolus-like body, oocyte size
At birth, mammalian oocytes are arrested at the diplotene stage of surge of each cycle, the oocyte reinitiates its first reductive meiosis
the first meiotic prophase. After delivery, one or more follicles aban- division, which culminates in the extrusion of the first polar body
don the ovarian reserves, and the first phase of concerted follicle and (PB). The mature oocyte is once again meiotically arrested at the
oocyte growth commences. During the second phase, the follicle metaphase II (MII) stage and remains so until fertilization.
enlarges, principally through formation of an internal cavity (the In humans, after the application of controlled ovarian hyperstim-
antrum), while the oocyte, though under meiotic arrest, possesses ulation protocols, oocytes isolated from follicles of different sizes
a large nucleus named the germinal vesicle (GV), which is charac- display certain heterogeneity within the same follicle size category
terized by a completely compact nucleolus known as the nucleo- (12 to 20mm). Of the dozen recovered oocytes, most are mature
lus-like-body (NLB). At this point, having achieved sexual (85% at the MII stage), while other immature examples are at either
maturity and receiving gonadotropic support, follicle growth the metaphase I (MI: 4%; oocytes without both GV and first PB) or
continues, while oocyte size increases slightly and acquires compe- the GV stage (11%) (1).
tence (namely, the ability to resume meiosis and sustain the first Early attempts at in vitro maturation (IVM) in human oocytes
cleavage divisions after fertilization). Thus, after the ovulatory LH date back to 1965 (2), but the first successful birth was reported in
1991 (3). However, clinical outcomes are variable and continue to
Received November 5, 2008; revised May 5, 2009; accepted May 25, be poor after transfer of embryos derived from in vitro matured
2009; published online July 10, 2009. oocytes with respect to those derived from in vivo matured oocytes
L.E. has nothing to disclose. N.G. has nothing to disclose. M.M. has (4). These variations have been attributed to intrinsic differences
nothing to disclose. A.P. has nothing to disclose. M-J.E. has nothing among recovered oocytes.
to disclose. Detailed analysis by fixation, specific Hoechst staining, and ultra-
Funded by IMPIVA (grant IMIDTF/2007/163; Generalitat Valenciana) and
violet exposure of fully grown mammalian oocytes recovered from
IVI Valencia.
Reprint requests: Mar
ıa-Jose Escriba, Ph.D., Instituto Universitario IVI, large antral follicles have revealed that immature GVoocytes consti-
Instituto Valenciano de Infertilidad, Plaza Polic Local, 3, Valencia
ıa tute a heterogeneous population in terms of chromatin configuration
46015, Spain (FAX: 34-963050999; E-mail: mjescriba@ivi.es). (5, 6), which is related to the subsequent meiotic progression
0015-0282/$36.00 Fertility and Sterilityâ Vol. 93, No. 8, May 15, 2010 2557
doi:10.1016/j.fertnstert.2009.05.077 Copyright ª2010 American Society for Reproductive Medicine, Published by Elsevier Inc.

2. (5, 7, 8), cytoplasmic maturation, and developmental competence of (Conseller de Sanitat, Generalitat Valenciana). Eggs were acquired after
ıa
the oocyte (4, 5, 9, 10). obtaining the written informed consent of donors.
The most detailed works on human GVoocytes are those by Com-
belles et al. (4) and Miyara et al. (11). These investigators established Oocyte Collection
four GVoocyte categories according to both chromatin configuration The oocytes employed in this study were obtained from healthy donors aged
and NLB dynamics after fixation and Hoechst staining. Both groups between 18 and 34 years (average: 27.7 years, standard deviation [SD] Æ3.9
agreed that modifications in chromatin organization consist of a shift years) and with no family history of chromosomal diseases. They underwent
from a decondensed, dispersed configuration around several NLB that a complete gynecologic examination, karyotype, and screening for infectious
are small or heterogeneous in size to a condensed chromatin, diseases such as human immunodeficiency virus, hepatitis B and C, gono-
organized around a single well-defined nucleolus. In mammals, these cocci, and syphilis. Oocytes were recovered after controlled ovarian hyper-
chromatin modifications correspond to a transition from a transcrip- stimulation treatment and follicle punction, and were incubated in vitro for
tionally active state to an inactive one as the end of the growth phase 4hours in 50mL of human tubal fluid medium (hTF; IVI Barcelona, Barce-
lona, Spain), under standard culture conditions (at 37 C and 5% CO2 in
of oogenesis approaches (4, 6, 10, 12, 13). Thus, the definition of
a humidified atmosphere). Cumulus cells were then gently removed by
morphologic markers of chromatin stage in fresh (nonfixed and
hand pipetting after a brief period of incubation in 40IU/mL of hyaluronidase
stained) GV oocytes before IVM represents an initial step toward (Sage In Vitro Fertilization Inc, Tumbull, CT).
future studies on the developmental competence of GV oocytes. Cumulus-free oocytes were classified as mature (MII) or immature. The
We first aimed to characterize a population of immature GV latter included oocytes with no PB or nuclear structure (MI) (Fig. 1D) and
oocytes recovered from stimulated cycles through morphometric those with a GV structure with no PB in the perivitelline space (see Fig.
and morphologic examination of different parameters using phase- 1A–C). Mature oocytes were used for reproductive purposes, whereas GV
contrast and image analysis, and to determine the configuration of oocytes were donated to be used in this research.
the GV chromatin by means of fluorescent DNA imaging. Second, The GVoocytes were collected from 66 donors and cultured in vitro in hTF
we defined three models with which to assess the value of selected medium for 1 to 3hours. The nuclear maturation state of eggs was confirmed
so that only oocytes at the germinal vesicle stage were included in the study.
morphometric and morphologic parameters as predictors of chroma-
tin status.
Morphologic GV Observations
Individualized immature GV oocytes were morphologically examined by
contrast-phase microscopy at Â400 magnification (Olympus, Barcelona,
MATERIALS AND METHODS Spain).
This work was approved by both the ethics committee of the Instituto For every oocyte (see Fig. 1), at least three images were recorded for
Universitario IVI (Valencia, Spain) and the Valencian regional government subsequent image analysis. The image capture process lasted no more than
FIGURE 1
(A–D) Contrast-phase microscopy images at Â40 magnification and (E, F) detailed nuclear observation of immature oocytes at the germinal
vesicle stage (GV: A–C) or metaphase I stage (MI: D). In images A, B, and C, arrows represent diameters of the (A) oocyte, (B) nucleus, and (C)
nucleolus. (A) GV oocyte with a well-defined nuclear envelope (grade 1) and one large eccentric NLB. (B) GV oocyte with an irregular nuclear
envelope (grade 2). Nucleus is centrally located in the ooplasm. (C) GV oocyte with a discontinuous nuclear envelope in some areas (grade 3)
that defines a nuclear area containing an clearly visible NLB. (D) Immature oocyte at the MI stage. Note the absence of both the nucleus and
the first polar body in the perivitelline space. (E) Detailed image of oocyte from A. Note smooth appearance of nucleoplasm. (F) Note rough
appearance of nucleoplasm of GV oocyte from C.
Escrich. Morphologic prediction of GV condensation. Fertil Steril 2010.
2558 Escrich et al. Morphologic prediction of GV condensation Vol. 93, No. 8, May 15, 2010

3. 1 minute per oocyte. We used OCTAX software (OCTAX Microscience After defining predictive variables, a statistical model was generated. By
GmbH, Herborn, Germany) for image analysis to measure the size of the introducing the values of these variables, a prediction was provided. The
oocyte (whole cytoplasmic diameter, excluding the zona pellucida; see predictive value of the model was assessed by calculating the area under
Fig. 1A), nuclear diameter (see Fig. 1B), and diameter of the largest NLB the curve (AUC 95% CI) according to receiver operating characteristic
(see Fig. 1C). Each parameter was measured three times by two of the three (ROC) analysis.
observers (LE, NG, and ME). Statistical analysis was performed using the Statistical Package for Social
Morphologically identified GV oocytes were also assessed to confirm Sciences 15 (SPSS Inc., Chicago, IL) and MedCalc Software (Ghent, Belgium).
location of the nucleus (central or peripheral), appearance of nucleoplasm
(smooth or rough; see Fig. 1E and 1F, respectively), and presence of NLB, RESULTS
whose position (central or peripheral; see Fig. 1F and 1E, respectively) and
number were recorded. In addition, the continuity of the nuclear envelope
A total of 1027 oocytes were recovered from 66 donors (17.8 Æ 7.8
of the GV oocytes was studied, and a progressive variation was established ova per donor). Donor age varied from 19 to 34 years (27.8 Æ 3.9
(see Fig. 1): well-defined (grade 1, see Fig. 1A and 1E), irregular features years). Seven hundred and seventy-five of the eggs were mature
(grade 2; see Fig. 1B), or discontinuous envelope in some areas (grade 3; (MII stage: 75.5%), 91 oocytes were at MI (8.7%), and 146 were
see Fig. 1C and 1F). Nuclear envelope was eventually broken down and at GV (14.2%). The remaining 15 oocytes (1.5%) were degenerated
became unidentifiable in MI oocytes (see Fig. 1D). or showed evident signs of atresia.
Of the 146 GVoocytes recovered, five had degenerated by the time
Observation of GV Chromatin Configuration of the second microscopic assessment, 10 eggs (6.8%) had matured to
Immediately after GV image capture, each immature oocyte was fixed and metaphase I within 1 to 3hours of recovery, and none progressed to
stained in 70% (v/v) ethanol with 25 mg/mL of Hoechst 33342 (Sigma, MII. Thus, 131 GV oocytes were finally included in the study.
Madrid, Spain) for 10minutes at room temperature. Oocytes were then equil- Each GV oocyte was morphologically studied using phase-contrast
ibrated in a mounted medium, placed on glass slides, and compressed with microscopy, and its chromatin conformation was determined by
coverslips to visualize GVs. Chromatin observations were performed with fixation, Hoechst DNA staining, and fluorescent microscopy, as previ-
an Olympus AX-70 epifluorescence microscope (Olympus Optical Co., ously described.
Hamburg, Germany) equipped with a triple-band pass filter for DAPI/Texas
Red/FITC and single-band pass filters for FITC, Texas Red, and Aqua Blue.
Images were recorded with an Olympus DP-70 video camera at both Â400
Model A: GV Classification According to Chromatin
and Â1000 magnifications. Condensation
It was critical to minimize the elapsed time between oocyte image capture After fluorescent chromatin observation, GV oocytes were retro-
and oocyte fixation because we were interested in registering the intrinsic spectively assigned to one of two stages of chromatin condensation
variability of recovered GV oocytes. Thus, rather than using drugs to block (decondensed vs. condensed) (Table 1). In the decondensed group
the meiotic cell cycle, we chose to record instantaneous morphologic and (GV-DC), the chromatin had a dispersed and fibrillar appearance
chromatin images. (Fig. 2A) or was observed to be a more condensed, continuous,
ring-like mass (see Fig. 2B). In the condensed chromatin group
GV Classification According to Chromatin Status (GV-CD), large clumps or strands of very condensed chromatin
After studying chromatin condensation and organization in relation to the were detected throughout the nucleoplasm (see Fig. 2C, D). These
NLB, we retrospectively grouped GV oocytes according to three models. classifications were established regardless of the chromatin distribu-
Model A was based on chromatin condensation stage, so GV oocytes were tion in relation to the NLB.
separated into two categories—noncondensed versus condensed—regardless The frequency distribution of GV oocytes with respect to oocyte
of their NLB dynamics. For model B, we grouped eggs into two categories classification, which was based on chromatin condensation, is pre-
according to chromatin distribution in relation to the NLB—surrounding
sented in Table 1. Data analysis indicated a decondensed (GV-DC)
versus not surrounding and/or absent—regardless of the chromatin conden-
or condensed (GV-CD) chromatin condensation stage (P¼.08) in
sation stage. Finally, model C, which was inspired by reports by Combelles
et al. (4) and Miyara et al. (11), consisted of grouping eggs into four GV comparable percentages of GV oocytes.
categories according to both chromatin condensation and distribution in Morphometry revealed statistically significant differences
relation to the NLB (combination of models A and B). between the mean diameters of GV-DC and GV-CD oocytes
(106.7 Æ 4.2 mm and 104.8 Æ 4.4 mm, respectively; see Table 1).
On the other hand, mean diameters of both the nucleus (mean:
Data Analysis
30.3 Æ 2.5 mm) and the largest NLB (mean 8.4 Æ 1.0 mm) were
Morphometric measures consisted of oocyte, nucleus, and NLB diameters,
comparable in the GV chromatin groups (P¼.1 and P¼.9,
and are presented in mm as either meanÆSD or 95% confidence interval
values (95% CI).
respectively; see Table 1).
After grouping GV oocytes according to the three models, statistically Table 1 shows the incidence of the morphologic parameters
significant differences among continuous variables (diameters) were studied in immature GV oocytes according to each chromatin
analyzed by analysis of variance (ANOVA), using Bonferroni’s correction condensation group. Nearly all GV oocytes (121 of 131) exhibited
for multiple comparisons. a swollen, subcortical nucleus (92.4%; P¼.9). Study of the nucleo-
Statistical analysis of discrete variables (i.e., nucleus location, nucleo- plasm revealed that a rough appearance correlated with a compact
plasm appearance, nucleolus presence, and number and continuity of nuclear chromatin (72.6%: GV-CD; 37.9%: GV-DC; P¼.0001; see Table 1).
envelopes) was performed using the chi-square test. P.05 was considered The NLB was present in 125 of the 131 GV oocytes and was
statistically significant. usually (76%) peripherally located (P¼.1 and P¼.1, respectively;
Multinomial and binomial logistic regression analyses were employed to
see Table 1). Nuclear envelope continuity altered as chromatin
determine the predictive value of the variables analyzed in this study. A pre-
dictive model was obtained for each of the three classifications proposed
condensed (P.05) (Fig. 3A), with statistically significantly more
herein (model A, B, or C) by the forward step method. In short, variables GV-CD oocytes exhibiting a more dispersed (grade 3) nuclear
for whom P.2 (determined by the Wald method) were introduced progres- envelope than GV-DC oocytes.
sively for computer analysis. Values were ordered according to the P value To determine the predictive value of all the variables with re-
from the most to least statistically significant. spect to model A, we performed a binomial logistic regression
Fertility and Sterilityâ 2559

4. TABLE 1
Frequency distribution of germinal vesicle oocytes according to chromatin condensation stage (model A): morphometric and
morphologic parameters.
Appearance of oocytes
Mean diameter in mm (95% CI) (n, % of corresponding chromatin GV group)
Peripherally Nucleoplasm Largest NLB
Chromatin Number of Largest located with rough At least located
stage oocytes, n (%) Oocyte GV NLB nucleus appearance one NLB peripherally
GV-DC 58 (44.3) 106.7 (105.6–107.8)a 30.7 (29.9–31.6) 8.4 (8.1–8.7) 53 (91.4) 22 (37.9)a 57 (98.3) 39 (68.4)
GV-CD 73 (55.7) 104.8 (103.7–105.8)b 30.0 (29.7–30.4) 8.3 (8.1–8.6) 68 (93.1) 53 (72.6)b 68 (93.1) 56 (82.3)
Total 131 30.3 (29.9–30.8) 8.4 (8.1–8.5) 121 (92.4) 75 125 (95.4) 95 (76.0)
a,b
Different superscripts within a column indicate statistically significant differences between germinal vesicle (GV) classification groups (P .05). DC: decon-
densed group; CD: condensed chromatin group; NLB: nucleolus-like body.
Escrich. Morphologic prediction of GV condensation. Fertil Steril 2010.
analysis using a forward step method. According to this analysis, oocyte is R.5, there is a high probability that the chromatin of that
the most relevant variable was the appearance of nucleoplasm, oocyte is in a condensed stage.
followed by nucleolus position, continuity of nuclear envelope,
and, finally, oocyte size. The rest of the continuous morphometric
(nuclear and nucleolus sizes) and discrete morphologic (position Model B: GV Classification According to Chromatin
of the nucleus and number of nucleolus) variables did not prove Distribution with Respect to the Nucleolus-like Body
to be statistically relevant and, therefore, were not included in To test model B, the GV oocyte population was divided into two
the model. groups according to chromatin in relation to the NLB: surrounding
As a result, the predictive model was as follows: versus not surrounding (Table 2).
Model A ¼ e ½9:586À0:114 ðoocyte sizeÞ þ 1:013 ðnuclear envelope grade 1Þ þ 2:099 ðnuclear envelope grade 2Þ þ 1:500 ðnucleolus locationÞ þ 1:648 ðnucleoplasm appearanceÞŠ
The value that estimated the fitting of the model (R2 Nagelkerke) In the first group (GV-SN), chromatin was organized throughout or
was 0.33. around a large NLB (see Fig. 2A–C). This group included various
Predictive value was obtained by introducing all the variables of chromatin condensation stages, namely, fibrillar-dispersed chromatin
each of the oocytes analyzed and calculating the exponential func- (see Fig. 2A), chromatin compacted into a continuous ring (see Fig.
tion. The predictive value of this model was defined according to 2B), or more condensed chromatin in clumps or strands (see Fig. 2C).
the AUC (and 95% CI) provided by ROC analysis. In this study, In the GV-NSN group, chromatin was not related to the presence
the AUC was 0.786 (95% CI, 0.709–0.864), which was considered of a small, barely detectable NLB. These oocytes frequently
statistically significant (P.0001). Thus, if this value for a particular exhibited highly compacted chromatin strands (see Fig. 2D).
FIGURE 2
Photographs of human oocytes at the immature stage of either GV (A–D) or MI (E), showing different patterns of chromatin configuration
observed with fluorescence microscopy (Â100) after Hoestch 33342 staining. (A) Chromatin is dispersed and has a fibrillar appearance in the
nucleolar area. (B) Chromatin is more condensed, forming a continuous mass close to the surface of a large NLB with a ring or horseshoe
appearance. (C) Similar to B, except that a few chromatin clumps or strands are detected in the nucleoplasm. (D) Chromatin is further
condensed into larger clumps and tend to be organized around a single small or absent nucleolus. (E) Chromatin is completely condensed and
organized on a meiotic spindle.
Escrich. Morphologic prediction of GV condensation. Fertil Steril 2010.
2560 Escrich et al. Morphologic prediction of GV condensation Vol. 93, No. 8, May 15, 2010

5. FIGURE 3 FIGURE 3 Continued
Frequency distribution of GV oocytes according to the continuity
Model A of the nuclear envelope grade (grade 1 to grade 3) for the three
1 proposed classification models. (A) Percentage of oocytes of each
GV-DC nuclear envelope grade (grade 1 to grade 3) in GV-DC and GV-CD
0,8
GV-CD
Percentage
groups (i.e., chromatin condensation stage: model A). Once again,
0,6 as the nuclear envelope advanced from grade 1 to grade 3,
becoming progressively more disorganized, more eggs of the GV-
0,4 DC group turned into GV-CD oocytes. At analysis of oocyte
frequency within the grade 3 nuclear envelop group, an increased
0,2 frequency of GV-CD oocytes was evident (P .05). (B) Percentage
of oocytes of each grade of nuclear envelope (grade 1 to grade 3)
0 in GV-SN and GV-NSN classification groups (i.e., chromatin
1 2 3 assembly in relation to the nucleolar-like body [NLB]: model B).
A Nuclear envelope grade Within all three nuclear envelope grades, there were high and
comparable percentages of GV-SN oocytes; the frequency of GV-
NSN oocytes progressively increased as the degree of nuclear
Model B envelope increased. However, these differences were not
1 GV-SN statistically significant. (C) Percentage of oocytes of each grade of
GV-NSN nuclear envelope (grade 1 to grade 3). According to GV1, GV2,
0,8
Percentage
GV3, or GV4 chromatin classifications (i.e., chromatin
0,6 condensation and assembly to NLB: model C), grade 1 and grade
2 nuclear envelopes were more frequent among GV1 and GV2
0,4 oocytes. As degree of nuclear envelope progressed, the
frequency of GV1 and GV2 oocytes diminished and that of GV3
0,2 and GV4 oocytes augmented. However, these variations in oocyte
frequency distribution did not reach statistical significance.
0
1 2 3
B Nuclear envelope grade Table 2 also shows the incidence of morphologic parameters in
GV-SN and GV-NSN oocytes. Almost all GV oocytes (121 of 131)
possessed a swollen, peripheral nucleus. However, there was a statis-
Model C
1 tically significant effect of chromatin–nucleolar dynamics on
GV1 GV3 frequency distribution (P¼.02). Study of the nucleoplasm revealed
0,8 GV2 GV4 a rough nucleoplasm appearance in almost 60% of GV oocyte, but
Percentage
this distribution was regardless of chromatin–nucleolar dynamics
0,6
(P¼.2; see Table 2).
0,4 A NLB was observed in 125 of the 131 GV oocytes, and its inci-
dence varied according to classification model. Thus, although all
0,2
GV-SN oocytes had one large nucleolus, 25% of GV-NSN oocytes
0 had no nucleolus (P.05; see Table 2). When a nucleolus was iden-
1 2 3 tified, it was peripherically located in 76.0% of the oocytes, regard-
C Nuclear envelope grade less of their classification (P¼.09; see Table 2).
With respect to the nuclear envelope continuity of GV oocytes
Escrich. Morphologic prediction of GV condensation. Fertil Steril 2010. and its relationship with the different chromatin–nucleolar dynam-
ics, no statistically significant differences were detected in the
frequency distribution of the different grades of continuity (see
Fig. 3B).
To determine the predictive value of all the variables analyzed
here with respect to model B, a binomial logistic regression analysis
The frequency distribution of GV oocytes according to chroma- was performed using a forward step method. None of the variables
tin assembly is presented in Table 2. Data analysis indicated that were found to be relevant; consequently, this model was confirmed
the majority of oocytes possessed chromatin organized throughout to have no practical application.
or around a single NLB (GV-SN: 84.7% vs. GV-NSN: 15.3%;
P.01).
Morphometry revealed no statistically significant differences Model C: GV Classification According to Chromatin
between the mean diameters of GV-SN and GV-NSN oocytes Configuration
(P¼.05; see Table 2), probably because of the variation of The GVoocytes were grouped into four categories according to their
sizes within the sample. Thus, GV oocytes measured on average chromatin configuration, which considered condensation and distri-
(105.6 Æ 4.5 mm). Similarly, comparable mean diameters of the bution in relation to the NLB. In the first group (GV1, see Fig. 2A),
nucleus (30.3 Æ 2.5 mm; P¼.5) and the largest NLB (8.4 Æ 1.0 chromatin was distributed throughout a large nucleolar area, and had
mm; P¼.6) were observed regardless of the oocyte classification a dispersed and fibrillar appearance. In the second group (GV2; see
(see Table 2). Fig. 2B), chromatin was more condensed. It occupied a smaller
Fertility and Sterilityâ 2561

6. TABLE 2
Frequency distribution of germinal vesicle oocytes according to chromatin assembly in relation to the nucleolus-like body
(model B): morphometric and morphologic parameters.
Mean diameter Appearance of oocytes
in mm (95% CI) (n, % of corresponding chromatin GV group)
Peripherally Rough At Largest NLB
Chromatin Number of Largest located nucleoplasm least located
stage oocytes (n, %) Oocyte GV NLB nucleus appearance one NLB peripherally
GV-SN 111 (84.7)a 106.0 (105.1–106.8) 30.4 (29.9–30.9) 8.4 (8.2–8.6) 105 (94.6)a 61 (54.9) 110 (99.1)a 81 (72.9)
GV-NSN 20 (15.3)b 103.9 (102.1–105.6) 30.0 (29.4–30.6) 8.2 (7.7–8.7) 16 (80.0)b 14 (70.0) 15 (75.0)b 14 (93.3)
Total 131 105.6 (104.9–106.4) 30.3 (29.9–30.8) 8.4 (8.1–8.5) 121 75 (57.5) 125 95 (76.0)
a,b
Different superscripts within a same column indicate statistically significant differences (P .05) between germinal vesicle (GV) classification groups. SN:
surrounding group; NSN: not surrounding group; NLB: nucleolus-like body.
Escrich. Morphologic prediction of GV condensation. Fertil Steril 2010.
portion of the nucleolar area and was organized around a large NLB Thus, GV oocytes recovered from ovarian stimulated cycles had
in a continuous mass with a ring or horseshoe appearance. In the an average size of 105.6 Æ 4.5 mm (see Table 3).
third group (GV3; see Fig. 2C), clumps or strands of condensed To determine whether the size of GV oocytes was comparable to
chromatin were detected throughout the nucleoplasm and around that of MI oocytes, we compiled images of 10 MI oocytes that
the NLB. The GV4 group (see Fig. 2D) displayed a more condensed matured within 1 to 3hours of recovery. The cytoplasmic diameter
chromatin, arranged in large clumps or strands, particularly around of MI oocytes measured 108.4 Æ 3.6 mm on average (95% CI,
a very small or practically undetectable nucleolus. 105.9–111.0mm, data not shown in tables), which made them larger
The frequency distribution of GV oocytes according to the afore- than GV oocytes. However, such differences were not statistically
mentioned chromatin classification is presented in Table 3. Data significant (P¼.055), probably because of the size disparity of
analysis indicated that statistically significantly more GV oocytes the sample. Though this aspect had not initially formed part
possessed a GV3 chromatin configuration (40.5%; P.05), whereas of the objectives of the study, the data obtained encouraged us to
comparable percentages of immature oocytes presented GV1, GV2, analyze it.
or GV4 patterns (19.8%, 24.4%, and 15.3%, respectively; P.05; In relation to nuclear and NLB diameters (see Table 3), no
see Table 3). satistically significant differences were observed among the four
Concerning morphometric parameters (see Table 3), statistically chromatin configurations (P¼.291 and P¼.635, respectively).
significant differences were observed in the mean size of oocytes Thus, regardless of chromatin status, GV oocytes had an average
(P¼.048); however, after applying Bonferroni’s multiple compari- nucleus size of 30.3 Æ 2.5 mm and one large NLB of 8.4 Æ 1.0 mm.
son tests (by pairs), differences were shown not to be statistically Table 3 shows the frequency of morphologic parameters observed
significant. This was probably due to the low number of eggs. in immature GV oocytes (see Fig. 1). Nearly all the GV oocytes
TABLE 3
Frequency distribution of germinal vesicle oocytes according to chromatin condensation and assembly to the nucleolus-like
body (model C): morphometric and morphologic parameters.
Mean diameter Appearance of oocytes
in mm (95% CI) (n, % of corresponding chromatin GV group)
Peripherally Rough At Largest NLB
Chromatin Number of Largest located nucleoplasm least located
stage oocytes, (%) Oocyte GV NLB nucleus appearance one NLB peripherally
GV1 26 (19.8)b 106.4 (104.7–108.0) 30.3 (29.6–31.0) 8.5 (8.1–8.9) 23 (88.5) 7 (26.9)c 25 (96.1)a,b 17 (68.0)
GV2 32 (24.4)b 107.0 (105.5–108.6) 31.0 (29.5–32.5) 8.2 (7.8–8.6) 30 (93.7) 15 (46.9)b,c 32 (100)a 22 (68.7)
GV3 53 (40.5)a 105.1 (103.8–106.4) 30.0 (29.6–30.5) 8.4 (8.1–8.6) 52 (99.9) 39 (73.6)a 53 (100)a 42 (789.2)
GV4 20 (15.3)b 103.9 (102.1–105.6) 30.0 (29.4–30.6) 8.2 (7.7–8.7) 16 (80.0) 14 (70.0)a,b 15 (75.0)b 14 (93.3)
Total 131 105.6 (104.9–106.4) 30.3 (29.9–30.8) 8.4 (8.1–8.5) 121 (92.4) 75 125 95 (76.0)
a,b,c
Different superscripts within a column indicate statistically significant differences between germinal vesicle (GV) classification groups (P .05).
NLB: nucleolus-like body.
Escrich. Morphologic prediction of GV condensation. Fertil Steril 2010.
2562 Escrich et al. Morphologic prediction of GV condensation Vol. 93, No. 8, May 15, 2010

7. possessed a swollen, peripherical nucleus (92.4%; P¼.058), regard- have initiated GVBD, and represents 15.3% of the GV population
less of their chromatin status. However, nucleoplasm appearance (GV4, model C). Accordingly, it should be highlighted that, in
changed according to chromatin configuration (P¼.00038). Data this study, 6.8% of initially GV oocytes progressed to MI within
indicated that a nucleoplasm with a rough appearance was more 1 to 3hours of in vitro culture. This immediately spontaneous GV
common in oocytes pertaining to the groups with compact chroma- maturation was also reported by Combelles et al. (4), who observed
tin (GV3 and GV4; see Table 3). that almost 24% of oocytes spontaneously resumed meiosis and
With regard to NLB, no variable was affected by chromatin matured to prometaphase I/MI after 6hours of in vitro culture.
configuration. We found NLBs were present in 125 of the 131 GV Reports on human GV oocytes (4, 11) are in agreement concern-
oocytes (95.4%; see Table 3), and this incidence did not vary with ing the relevance of oocyte size in GV arrangement, but they
chromatin configuration stage (P¼.2). Moreover, when a large approach the question from different perspectives. In our study,
NLB was present, it was peripherically located in 95 of the 125 oocyte size contributed to the predictive model; all GVoocytes mea-
nuclei, regardless of chromatin configuration (P¼.2; see Table 3). sured between 91.1mm and 108.9mm, with the smallest size being
With respect to the study of the nuclear envelope continuity of associated with GVBD. Miyara et al. (11) suggested that changes
GV oocytes and its relationship to chromatin configuration groups, in chromatin configuration in human oocytes vary with size, and
no statistically significant differences were detected in frequency they established three oocyte size categories (small, intermediate,
distribution (P¼.05; see Fig. 3C). and large), with the largest oocyte size being associated with
To determine the predictive value of all the variables with respect GVBD. Combelles et al. (4) also reported the relevance of oocyte
to model C, a multinomial logistic regression analysis was size, which in their study varied between 106.6mm and 115.5mm,
performed using a forward step method. Analysis of all the variables and they also established two oocyte size categories related to
revealed that only nucleoplasm appearance was relevant with oocyte competence. It can be said that the larger the oocyte is, the
respect to this classification model. Therefore, it was concluded earlier the ovum matures, but small oocytes are associated with atre-
that model C has a weak practical application. sia and are thought to be incompetent (4, 10). In general terms, we
have observed that our oocyte size values (105.6 Æ 4.4 mm) are
considerably lower than those reported by other investigators
DISCUSSION (4, 11). Differences in oocyte size may be due not only to chromatin
We evaluated a GV oocyte population from both morphologic and stage but also to variations in the ovarian hyperstimulation protocols
chromatin organization perspectives using phase-contrast micros- administered before oocyte recovery, the fertility status of our
copy and image analysis and fluorescent DNA imaging, respec- population (all oocytes were obtained from fertile young women),
tively. This compilation of data allowed us to group GV oocytes culture conditions, the inclusion of zona pellucida in the measure-
retrospectively, according to three hypothetical models based on ment of total oocyte diameter, or the use of drugs to synchronize
chromatin condensation (model A), chromatin distribution in rela- the GV oocyte cell cycle (i.e., dbdAMP or IBMX).
tion to the NLB (model B), or chromatin condensation with respect To our knowledge, ours is the first report in which a mathematical
to NLB dynamics (model C). Finally, and most relevant from a prac- model has been shown to predict chromatin condensation stage
tical point of view, we propose a predictive model of GV chromatin using morphometric and morphologic variables. Though an AUC
condensation stage based on quick image analysis of four relevant under 0.8 is questionable from a purely statistical point of view,
morphologic variables. we feel that the AUC obtained in this study (0.786; see the results
The GV oocyte chromatin classification showed our GV popula- section) is sufficiently high considering the intrinsic variability of
tion to be decondensed (DC) and condensed (CD) in similar propor- the biological samples.
tions (44.3% and 55.7%, respectively). We have found that this As represented in the results section, our predictive model is as
parameter can be predicted by four of the variables studied herein: follows:
Model A ¼ e ½9:586À0:114 ðoocyte sizeÞ þ 1:013 ðnuclear envelope grade 1Þ þ 2:099 ðnuclear envelope grade 2Þ þ 1:500 ðnucleolus locationÞ þ 1:648 ðnucleoplasm appearanceÞŠ
appearance of nucleoplasm, nucleolus position, continuity of the To take advantage of the practical benefits of this model, it is nec-
nuclear envelope, and oocyte size. essary to assess the morphometrically and morphologically relevant
The biological and physiologic significance of a nucleoplasm variables in the captured image of a GV oocyte and to apply the
with a rough appearance and the position of the peripherical nucle- values to the formula. Whereas oocyte size is a continuous variable,
olus is a question of great relevance. The relation of these aspects to the rest of the variables studied herein have different binary values
chromatin condensation stage is clear, as can be seen in the contri- (namely, 0 or 1) or variable categories (grade 1 to 3). In this way,
bution of such variables to our proposed predictor model. However, an oocyte with nuclear envelope grade 1 will have a variable value
as far as we know, no direct relationship has previously been estab- of 1, and that with a nuclear envelope grade 2 will have a variable
lished between these morphologic features and subsequent oocyte value of 0. On the other hand, nucleolus location is valued at
competence. 0 when it is central to the GV oocyte, and at 1 when it is peripheral.
We have previously discussed the possible relation between Finally, nucleoplasm appearance is also a binary variable. It is
nuclear envelope continuity imminent germinal vesicle breakdown scored 0 when smooth in appearance and 1 when rough. Values
(GVBD). At compacted stages, all clumps and stands were detected are obtained by means of arithmetic. If the value is R.5, the GV
in the nucleoplasm and were organized either around a small, prac- oocyte is likely to have a compact chromatin.
tically undetectable nucleolus or independently of any nucleolar We have proposed a predictor method based on four morphologic
structure. This GV category corresponds to oocytes that would and morphometric variables (nucleoplasm appearance, nucleolus
Fertility and Sterilityâ 2563

8. position, continuity of nuclear envelope, and oocyte size) that pro- Acknowledgments: The authors thank all the IVI staff, in particular the
vide information about the chromatin condensation stage of the oo- embryologists who collected the GV immature oocytes day after day; Empar
cyte. Because the GV chromatin condensation stage previously has Garcıa-Rosell, Ph.D., for her assistance with the Hoeschst stainings, and the
o
been related to oocyte quality in IVM, transcriptional activity, and staff of the PGD department for their help with fluorescence; Mrs. Celia
Urios, Mrs. Mar Jos Ferri, Mrs. Eva Prez, Mrs. Mar Diana, and Mrs.
ıa e e
nuclear, cytoplasmic, and genomic competence (4, 11, 14), our
Amparo Bald for their help with donor recruitment and informed consent
o
next objective is to determine whether this new predictive GVoocyte management; and Mr. Brian Normanly for editing this manuscript.
classification is related to meiotic and cytoplasm competence.
REFERENCES
1. De Vos A, Van de Velde H, Joris H, Van isolated from antral follicles exhibit different chro- 10. Albertini DF. Origins and manifestations of oocyte
Steirteghem A. In-vitro matured metaphase-I oocytes matin organization and follow different maturation maturation competencies. Reprod Biomed Online
have a lower fertilization rate but similar embryo qual- dynamics. Mol Reprod Dev 1993;36:59–74. 2003;6:410–5.
ity as mature metaphase-II oocytes after intracytoplas- 6. Zuccotti M, Piccinelli A, Giorgi Rossi P, Garagna S, 11. Miyara F, Migne C, Dumont-Hassan M, Le Meur A,
mic sperm injection. Hum Reprod 1999;14:1859–63. Redi CA. Chromatin organization during mouse oo- Cohen-Bacrie P, Aubriot FX, et al. Chromatin config-
2. Edwards RG. Maturation in vitro of mouse, sheep, cyte growth. Mol Reprod Dev 1995;41:479–85. uration and transcriptional control in human and
cow, pig, rhesus monkey and human ovarian oocytes. 7. Schramm RD, Tennier MT, Boatman DE, mouse oocytes. Mol Reprod Dev 2003;64:458–70.
Nature 1965;208:349–51. Bavister BD. Chromatin configurations and meiotic 12. Mattson BA, Albertini DF. Oogenesis: chromatin and
3. Kim BK, Lee SC, Kim KJ, Han CH, Kim JH. In vitro competence of oocytes are related to follicular diam- microtubule dynamics during meiotic prophase. Mol
maturation, fertilization, and development of human eter in nonstimulated rhesus monkeys. Biol Reprod Reprod Dev 1990;25:374–83.
germinal vesicle oocytes collected from stimulated 1993;48:349–56. 13. Bouniol-Baly C, Hamraoui L, Guibert J, Beaujean N,
cycles. Fertil Steril 2000;74:1153–8. 8. Wickramasinghe D, Ebert KM, Albertini DF. Szollosi MS, Debey P. Differential transcriptional ac-
4. Combelles CMH, Cekleniak NA, Racowsky C, Meiotic competence acquisition is associated tivity associated with chromatin configuration in fully
Albertini DF. Assessment of nuclear and cytoplasmic with the appearance of M-phase characteristics grown mouse germinal vesicle oocytes. Biol Reprod
maturation in in-vitro matured human oocytes. Hum in growing mouse oocytes. Dev Biol 1991;143: 1999;60:580–7.
Reprod 2002;17:1006–16. 162–72. 14. Liu Y, Sui HS, Wang HL, Yuan JH, Luo MJ, Xia P,
5. Debey P, Szollosi MS, Szollosi D, Vautier D, 9. Abeydeera LR. In vitro production of embryos in et al. Germinal vesicle chromatin configurations of
Girousse A, Besombes D. Competent mouse oocytes swine. Theriogenology 2002;57:256–73. bovine oocytes. Microsc Res Tech 2006;69:799–807.
2564 Escrich et al. Morphologic prediction of GV condensation Vol. 93, No. 8, May 15, 2010