| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Institute of Reproductive and Developmental Biology (S.A.S., S.F., K.H.), and Department of Metabolic Medicine (S.M.D.), Imperial College London, Hammersmith Hospital, London W12 0NN, United Kingdom; and Department of Mathematics (J.S.), Imperial College London, London SW7 2AZ, United Kingdom
Address all correspondence and requests for reprints to: Dr. Kate Hardy, Institute of Reproductive and Developmental Biology, Imperial College London, Hammersmith Hospital, London W12 0NN, United Kingdom. E-mail: k.hardy{at}imperial.ac.uk.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective: Our aim was to examine whether the accelerated transition of follicles from primordial to primary stages in polycystic ovaries (PCO) is due to increased granulosa cell (GC) division.
Design: This study is a comparison of expression of minichromosome maintenance protein 2 (MCM2) (present in the nuclei of cells that are licensed to divide) in archive tissue from normal and PCO.
Setting: This is a laboratory-based study.
Patients: There were 16 women with regular cycles (six with normal and 10 with PCO) and five anovulatory women with PCO, classified histologically, with reference to menstrual history and ultrasound.
Main Outcome Measures: The presence of MCM2 expression in the GCs of 1371 follicles was determined.
Results: GC proliferation was increased in anovulatory PCO compared with both normal and ovulatory PCO, with an increased proportion of preantral follicles with MCM2-positive GCs (P 
0.015). The number of GCs differed significantly among the three types of ovary at the transitional (P = 0.013) and primary (P = 0.0096) stages. This was accompanied by an altered relationship (P < 0.0001) between oocyte growth and GC division/cuboidalization.
Conclusions: These findings provide evidence for increased GC proliferation in early-growing follicles in PCOS. This offers an explanation for the increased proportion of primary follicles in PCOS.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Initiation of follicle growth is associated with oocyte growth, granulosa cell (GC) division, and transformation of GC morphology from flattened to cuboidal. The mechanism of the putative acceleration of follicle growth in PCO remains unclear. In a primate model for PCOS, Vendola et al. (3) reported more growing follicles associated with evidence of increased GC proliferation, as determined by Ki67 labeling. We used a similar approach to study archive sections of normal and polycystic human ovaries but found that Ki67-positive GCs were present in only 3% of primordial, transitional, and primary follicles (our unpublished data), in agreement with the results of a previous study (4). This is not surprising because in the human, primordial follicles can remain arrested for up to 50 yr, and the transition from primordial to primary stages probably takes several months (5). Furthermore, a long cell cycle time has been observed in GCs at early preantral stages in the rat (6).
Another common marker of proliferation that has been used in ovarian studies is proliferating cell nuclear antigen (PCNA) (7, 8, 9, 10, 11). However, it is now clear that PCNA has a number of roles in the cell cycle; it is not unique to proliferating cells, and is also required for DNA repair, so it can persist at low levels in nonproliferating cells (12). Therefore, we went on to investigate an alternative marker of cell proliferation, minichromosome maintenance protein 2 (MCM2), which labels a higher proportion of proliferating cells than Ki67 and is present throughout the cell cycle, but not in quiescent or differentiated cells. MCM2 is increasingly used to detect premalignant and malignant cells in various cancers (12).
MCM2 is part of a single protein complex, a heterohexamer (comprising proteins MCM2, -3, -4, -5, -6, and -7), which is essential for the initiation of DNA replication and the function of which is to "license" DNA replication so that it occurs only once during each cell cycle (12). It is becoming clear that replication licensing plays a central role in regulating proliferation (13). MCM2 is present in the nucleus, and is degraded rapidly after cell differentiation and more slowly after entry into G0 (14, 15, 16, 17). The increased proportion of MCM-positive cells compared with Ki67 (18, 19, 20) is thought to be due to the involvement of MCM proteins from early G1 in the cell cycle (21) and/or to the labeling of cells in prolonged G1 that are licensed to replicate (17). Therefore, MCM2 is ideal for detecting infrequently dividing cells such as GCs in early preantral follicles because of its nuclear localization during any part of the cell cycle.
The aims of this study were 2-fold. First, we wished to examine whether GC division was increased in early preantral follicles in PCO from anovulatory women, accounting for our earlier observation of an increased proportion and density of primary follicles in these ovaries (1). Second, in tandem with this first aim, we wanted to carry out a detailed morphometric analysis of oocyte size, together with GC number, morphology and division, to clarify the sequence of cellular events accompanying entry of follicles into the growing phase.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Whole ovary tissue samples (formalin fixed and paraffin embedded), from patients who had undergone oophorectomy for nonmalignant gynecological disorders between 1993 and 2000, were obtained from the histopathology tissue bank at St. Mary’s Hospital National Health Service Trust London, United Kingdom. Ethical approval for the use of this tissue was obtained from the Local Research Ethics Committee, and the tissue had been removed with informed surgical consent. The samples were anonymous, keeping only details of age, cycle history, and ovarian morphology on ultrasound for comparison.
There were 21 archive ovary samples, classified as normal (n = 6) or polycystic (n = 15) on the basis of cycle history, ultrasound examination, and histology, used in this study (Table 1
). Their classification has been described in detail previously (22); 10 of the subjects with PCO were ovulatory (ovPCO), and five were anovulatory (anovPCO). One section from each sample was used for immunohistochemical staining for MCM2, and all sections were stained in a single session. There were no significant differences among the three groups in terms of age or body mass index (BMI) (Kruskal-Wallis). Overall, there was a significant difference in ovarian volume (Table 1).
|
Immunohistochemistry was performed as described recently (22). Sections were incubated with mouse antihuman MCM2 (clone D1 12A3, 
15 µg/ml undiluted) overnight at 4 C (23). A secondary antibody, goat antimouse (1 in 200 for 60 min; DakoCytomation Ltd., Ely, UK), was used, and labeling was visualized with a peroxidase-conjugated avidin biotin complex and 3,3'-diaminobenzidine tetrahydrochloride (22). As a negative control, the primary antibody was omitted.
Analysis
Coded anonymized slides were examined on an E600 microscope (Nikon UK Ltd., Kingston-upon-Thames, UK). Using a x60 objective, every follicle within the section was located and assessed at the microscope. Only healthy follicles with a visible oocyte were analyzed; atretic follicles with pyknotic GCs or oocytes were excluded. First, each follicle was classified as: 1) primordial, with one layer of flattened pre-GCs; 2) transitional, in which at least one but not all GCs were becoming cuboidal (intermediate) or were already cuboidal (Fig. 1A); 3) primary, with one complete layer of intermediate or cuboidal GCs; 4) secondary, with two or more layers of GCs and beginnings of a thecal layer; and 5) antral, with the appearance of an antral cavity. Second, GCs were examined carefully, while focusing through the section, for cell shape (flattened, intermediate, or cuboidal GCs; Fig. 1A) and absence (Fig. 1B) or presence (Fig. 1, C–H) of moderate and strong staining for MCM2. Finally, the presence or absence of a nucleus in the oocyte was noted and used as a marker of the largest cross-section of the follicle (22). In such follicles, an estimate of oocyte diameter was obtained by measuring, in two perpendicular dimensions, the inner follicle diameter. Direct measurement of oocyte diameter was not possible because formalin fixation does not preserve oocyte shape well.
|
High-resolution images of every follicle in the entire section were captured on a DXM 1200 digital camera (Nikon), using the Lucia image analysis program (Nikon), and saved in tagged image file format. The presence or absence of staining was independently confirmed by a second investigator (K.H.).
Statistical analysis
Statistical analysis of MCM2 staining of GC nuclei was carried out using Stata 8 for the Macintosh (StataCorp, College Station, TX). Mean proportions of both MCM2-positive follicles (with one or more positive GCs) and MCM2-positive GCs at each stage of development and in each type of ovary were computed and compared using logistic regression ("blogit" command). This was also used to calculate and compare the coefficients for the trends in such proportions with stage. In computing confidence intervals and P values, standard regression estimates of the variance of the estimates of proportions are sensitive to within sample or within follicle correlations. Instead, we have, therefore, used robust clustered SE values, which take account of such correlations and are much less sensitive to model misspecification (24). For comparisons between proportions of follicles, robust SE values were calculated with clustering by sample. This takes account of possible within-ovary correlation in MCM2 staining. For comparisons between proportions of GCs, robust SE values were calculated with clustering by follicles, taking into account possible within-follicle correlation of dividing cells in those follicles that have initiated growth. P values were corrected for multiple comparisons at each stage of follicle development (i.e. multiplied by three for comparisons among the three ovary groups).
Oocyte diameters and the numbers of GCs in the largest cross-section of follicles were analyzed using ordinary regression ("regress" command), again using robust SE values with clustering by sample. In effect, this gives a more flexible and robust ANOVA type analysis. A few groups were not normally distributed, and significant differences were confirmed using nonparametric analysis (Kruskal-Wallis for multiple groups; if P < 0.05, pairwise comparisons were performed using the Mann-Whitney U test, correcting for multiple comparisons by multiplying the P value by the number of comparisons made, Stata 8).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A total of 1371 follicles was examined, of which 89 were antral, and 1282 were preantral (Table 1
). Positive staining for the cell cycle marker MCM2 was observed from the primordial stage onwards (Fig. 1). Follicles with one or more GC nuclei labeled with MCM2 antibody are termed MCM2-positive follicles. In total, 13,187 preantral and 5,495 antral GCs were scored for shape (unilaminar follicles only) and immunolabeling for MCM2 (all follicles).
Proportion of MCM2-positive follicles
Twice as many preantral follicles from anovPCO contained one or more MCM2-positive GCs compared with follicles from ovPCO and normal ovaries. The proportion of positive preantral follicles in ovPCO and normal ovaries was similar (Fig. 2A).
|
). Although only a few secondary follicles were observed (three to seven per group), they all, regardless of ovary type, had MCM2-positive GCs that were scattered throughout the layers with no evident regionalization.
In all three types of ovary, the proportion of MCM2-positive follicles increased as follicle development progressed, with a significant positive trend in each (Fig. 2B; P < 0.0001). Between the primordial and secondary stages, the profile of the positive trend (combining slope and intercept with the y-axis) for anovPCO was significantly different from ovPCO (Fig. 2B; P = 0.0012).
Proportion of MCM2-positive GCs in preantral follicles
Overall, twice as many GCs from anovPCO were MCM2 positive compared with those from normal ovaries (P = 0.027; Fig. 2C).
There was a significant difference among the three types of ovary in terms of the proportion of MCM2-positive GCs at both the primordial (P = 0.0014) and transitional (P = 0.034) stages (Fig. 2D). At the primordial stage, corrected pairwise comparisons showed that the proportion of MCM2-positive GCs in anovPCO was significantly higher than in normal (P = 0.024) or ovPCO (P = 0.005). At the transitional stage, there were no significant pairwise differences among types of ovary. The proportion of MCM2-positive GCs at the primary or secondary stages was not significantly different among the three types of ovaries (Fig. 2D).
The overall proportion of MCM2-positive GCs increased as follicle development progressed, with a significant positive trend in all three types of ovary (P < 0.0001; Fig. 2D). As the follicles progressed from the primordial through to the secondary stages, the rate of increase in MCM2-positive GCs in anovPCO was not as steep as in normal (P = 0.006) and ovPCO (P = 0.0012). During preantral development, the profile of the positive trend (combining slope and intercept with the y-axis) for anovPCO was significantly different from normal (P = 0.009) and ovPCO (P = 0.0009).
Number of GCs in preantral follicles
The observation of significantly more MCM2-positive follicles and cells in anovPCO led us to ask whether this resulted in more GCs in anovPCO preantral follicles, as a consequence of cell division.
The number of GCs surrounding the oocyte in the largest cross-section of primordial follicles varied widely from two to 12 in normal ovary. At the primordial stage, there was no significant difference among the three types of ovary in the mean number of GCs (Fig. 2E). However, by the transitional stage, oocytes in anovPCO were enclosed by significantly more GCs than oocytes in ovPCO or normal ovaries (P < 0.05). Similarly, at the primary stage, oocytes in anovPCO had significantly more GCs than normal ovaries (P < 0.05; Fig. 2E). In anovPCO, the increased number of GCs at the transitional and primary stages, coupled with an increased incidence of MCM2-positive cells, suggested that GC division was increased.
We then examined whether the presence of cells in the cell cycle was related to increased GC numbers. In anovPCO, MCM2-positive primordial follicles had significantly more GCs in the largest cross-section than negative follicles. (Fig. 3A). At the transitional and primary stages in all three types of ovary, MCM2-positive follicles had significantly more GCs than MCM2-negative follicles (Fig. 3, B and C).
|
GC number increased significantly between the primordial and primary stages in all three types of ovary (Fig. 2E), suggesting that cell division was occurring at the same time as the GCs became cuboidal. We wished to explore the association between GC morphology and cell division in more detail. GCs were categorized as flattened, intermediate, and cuboidal (Fig. 1A). Overall, the proportion of MCM2-positive cells significantly increased as cells transformed from flattened to intermediate (P < 0.0001) and from intermediate to cuboidal (P < 0.0001; Fig. 4A). In both flattened and intermediate cells, the proportion of MCM2-positive cells was significantly higher in anovPCO compared with normal and ovPCO, but in cuboidal cells, levels of staining were similar in the three types of ovary (Fig. 4B).
|
). At the transitional stage, as cells transformed from flattened to cuboidal morphology, the proportion of dividing cells increased in all types of ovary. Oocyte growth
At the transitional stage, anovPCO oocytes were significantly larger than normal and ovPCO oocytes. At the primary stage, oocytes from anovPCO were significantly larger than those from ovPCO (Fig. 2F).
At the primordial stage, there was wide variation in oocyte diameter from 24–48 µm. This led us to ask whether GC division only occurred in follicles in which the oocytes may have started growing. At the primordial stage, there was no evidence that this was the case. At the transitional and primary stages, MCM2-positive follicles generally contained larger oocytes (Fig. 3, E and F).
Relationship between oocyte growth and GC division
Finally, in light of increasing evidence that folliculogenesis relies on bidirectional communication between GCs and oocyte, we compared the relationship between GC division and oocyte growth in the three types of ovary (Fig. 5). Both parameters were measured in the largest cross-section of the follicle, so only follicles with a visible oocyte nucleus were included in this analysis. Categorization by oocyte diameter suggested a nonlinear relationship between the number of GCs and oocyte diameter (data not shown). To explore this, we fitted a regression of log GC number against log oocyte diameter, as described previously (10). This gives a relationship of the form N = aDb, where N = number of GCs, D = oocyte diameter, and a and b are coefficients obtained by the fit. For normal and ovPCO, b was very close to 1, giving a linear relationship, whereas for anovPCO b was more than 2, giving an approximately quadratic relationship (Fig. 5). A test of the overall relationship between oocyte growth and GC division shows that this is significantly different in the three categories (P < 0.0001). A test of the coefficient b that describes the nature (e.g. linear or quadratic) of the relationship shows that this is the same in the normal and ovPCO categories (P = 1), whereas it is significantly different between anovPCO and normal (P = 0.0066), and between anovPCO and ovPCO (P = 0.0033). Combining the normal and ovPCO categories and testing against the anovPCO category yields P = 0.0004. This means that as oocyte diameter increases, GC division in anovPCO is significantly faster than in the other two types of ovary.
|
The majority of antral follicles showed widespread MCM2 labeling of GC nuclei, indicating extensive cell division in this population. On average, a significantly higher percentage of GCs was MCM2 positive in ovPCO (56%) compared with anovPCO (39%; P = 0.033). In normal tissue, 37% of GCs was positive (not significant). The positive GCs were scattered throughout the mural GC layer, and there was no regionalization of staining in relation to the basal lamina or antrum. Although we did not perform a quantitative analysis of the presence of MCM2 in theca cells, it was clear that in some of the larger follicles, a high proportion of theca cells stained positively with the MCM2 antibody (Fig. 1H).
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
It has recently been suggested that the DNA replication licensing system, in which the MCM protein complex plays a key part, performs a universal role in determining the proliferative capacity of cells in many tissues. Absence of replication licensing (in the form of lack of MCM) has been proposed as a mechanism for down-regulating cell division (17) and, hence, may be involved in the maintenance of a stock of relatively quiescent primordial follicles. In anovPCO we have demonstrated that an increased proportion of follicles has cells with proliferative potential, suggesting that the regulation of proliferation at these early stages of follicle development has been disturbed. Whether this is due to alterations in the expression of transcription factors, local paracrine factors, or their receptors remains to be elucidated.
Even though primary follicles are considered to be growing (5), between 20% (anovPCO) and 60% (ovPCO and normal) were completely MCM-negative (Fig. 2D). This is not surprising because early preantral follicle growth is thought to be slow, taking several months to complete (5). Immunohistochemistry of normal human ovary for the cell proliferation marker Ki67 did not detect any labeling of primordial and primary follicles (4), and our own studies using this approach showed that the majority of primary follicles were unlabeled (90%; data not shown). Thus, it appears that GC proliferation is a sporadic event at the earliest stages of follicle growth, with the majority of cells in G0, and becomes more widespread as multiple layers of GCs develop (Fig. 2D). At later preantral stages, when follicles are considered to be committed to an irreversible, if slow, growth trajectory, the maximum proportion of MCM2 positive GCs never exceeded 50%, and averaged around 30% at the secondary stage. MCM proteins are lost from the cell after entry into quiescence (15, 16, 17), therefore, it is clear that a substantial proportion of GCs in preantral follicles was in G0. Even in antral follicles, at a time of extensive GC proliferation, there were varying proportions of quiescent cells. In preantral follicles, the distribution of MCM-positive GCs appeared random, with no apparent regionalization in relation to either oocyte or basal lamina surrounding the granulosa layer.
The similar numbers of GCs in primordial follicles from normal and PCO strongly suggest that follicles are endowed with comparable numbers of pre-GCs at follicle formation. MCM2 immunolabeling was seen in occasional flattened GCs in primordial follicles, suggesting that these cells were either in the process of, or were licensed to undergo, cell division. Positive immunostaining for PCNA has been observed occasionally in primordial follicles of cow, primate, human, and sheep (8, 9, 10, 11), but not rat (7, 25). However, positive labeling of flattened GCs after long-term infusion of rats with tritiated thymidine (6) or 5-bromo-2'-deoxyuridine (26) suggested that very low levels of cell division were occurring. The labeling of a few flattened GCs in the human either marks initiation of follicle growth or represents a process of slow turnover, which contributes to the variable number of GCs in primordial follicles. Not all follicles showing evidence of GC division are committed to the growth phase; the persistence of 5-bromo-2'-deoxyuridine labeled GCs in some rat transitional follicles for more than 20 wk shows that such follicles can be considered to be still resting (26). Our data suggest that human primordial follicles are not entirely quiescent and that entry into the cell cycle can precede cuboidalization. Most importantly, our data further show that cell division, or the potential to divide, is increased in primordial follicles in anovPCO.
Cuboidal GCs expressed MCM2 more frequently than flattened cells (Fig. 4A), and transitional follicles with one or more cuboidal cells had more GCs (Fig. 2E). Together, these observations suggest that change in GC shape results in, or at least is accompanied by, an increased proliferative capacity. Interestingly, in flattened GCs and intermediate cells that were undergoing transformation, MCM2 labeling was more frequent in anovPCO. However, the incidence of MCM2 labeling in fully cuboidal cells was similar in all three types of ovary (Fig. 4B), and this was reflected by the similarity of labeling levels in primary follicles (Fig. 2D). Thus flattened pre-GCs and those undergoing the earliest changes in morphology had higher levels of MCM2 labeling in anovPCO, but equivalence was restored among the three types of ovary after the cells became cuboidal. This suggests an accelerated progress to the primary stage in anovPCO, followed by a decelerated progress afterwards (Fig. 2D). This model is in keeping with the recent suggestion that primary follicles are "stockpiled" in PCO (2).
It has previously been proposed that division and cuboidalization of GCs precede the onset of oocyte growth in cow (8, 27) and human (28). In this study we also observed increased GC numbers and cuboidalization before the onset of oocyte growth in normal ovary. However, in anovPCO, oocytes in transitional follicles (with one or more cuboidal cells) were significantly larger than those in normal or ovPCO (Fig. 2F). This strongly suggests that, in anovPCO, oocyte growth and GC cuboidalization are concurrent processes, rather than sequential as we see in the normal ovary and as previously described (28). Furthermore, our data show that the relationship between oocyte growth and GC division is different in anovPCO, with GCs dividing at a faster rate relative to oocyte growth than in ovPCO and normal follicles (Fig. 5). From these observations, it can be postulated that there is a disturbance in the dialogue between oocyte and GC in anovPCO, perhaps resulting from alterations in the onset or amount of expression of a particular growth factor that is either involved in oocyte growth, transformation of GCs, or regulation of GC division.
In summary, this is the first study to compare GC proliferation in early preantral follicles in human normal and PCO. We have found that GC proliferation in anovulatory women with PCO differs from that in ovulatory women with polycystic or normal ovaries, with an increase in the proportion of follicles showing evidence of GC division, increased GC numbers in transitional and primary follicles, and an altered relationship between oocyte growth and GC division or cuboidalization (Fig. 5). These findings provide a possible explanation for the increase in primary follicles seen previously in anovPCO, but the mechanisms contributing to these disturbances remain to be elucidated.
|
Footnotes |
|---|
Disclosure Statement: The authors have nothing to disclose.
First Published Online August 14, 2007
Abbreviations: anovPCO, Anovulatory polycystic ovary; BMI, body mass index; GC, granulosa cell; MCM2, minichromosome maintenance protein 2; ovPCO, ovulatory polycystic ovary; PCNA, proliferating cell nuclear antigen; PCO, polycystic ovaries; PCOS, polycystic ovary syndrome.
Received March 29, 2007.
Accepted August 3, 2007.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  P. Da Silva-Buttkus, G. S. Jayasooriya, J. M. Mora, M. Mobberley, T. A. Ryder, M. Baithun, J. Stark, S. Franks, and K. Hardy Effect of cell shape and packing density on granulosa cell proliferation and formation of multiple layers during early follicle development in the ovary J. Cell Sci., December 1, 2008; 121(23): 3890 - 3900. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Franks, J. Stark, and K. Hardy Follicle dynamics and anovulation in polycystic ovary syndrome Hum. Reprod. Update, May 22, 2008; (2008) dmn015v1. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
