The Notch2 gene is expressed at high levels in granulosa cells, the somatic-cell lineage of the ovarian follicle. To assess whether Notch2 gene function is required for correct development of the mouse ovary, we conditionally deleted the Notch2 gene in the somatic lineages of the mouse ovary using the anti-Müllerian hormone Type II receptor- Cre recombinase (Amhr2-Cre) driver line 19. Using Southern blotting analysis, we found efficient excision of the Notch2^flox allele by the Amhr2-Cre line in ovary DNA, but not in spleen DNA (Figure 1). Female Amhr2-Cre/+;Notch2^flox/- and control littermate mice were mated to wild-type C57BL/6J male mice over a 6-month period, then we assessed the fertility and reproductive performance of both groups of mice. The Amhr2-Cre/+;Notch2^flox/- mice exhibited reduced fertility compared with the littermate controls. Female Amhr2-Cre/+;Notch2^flox/- mice had only approximately one-third as many progeny as control littermates over the 6-month period, even though Amhr2-Cre/+;Notch2^flox/- mice had similar total litter numbers to the controls (Table 1).

Histological analysis of ovaries taken from Amhr2-Cre/+;Notch2^flox/- mice at 3 weeks of age identified the presence of multi-oocyte follicles (Figure 2). This phenotype was completely penetrant, as all Amhr2-Cre/+;Notch2^flox/- mice developed multi-oocyte follicles. As many as 10 oocytes were present in a single follicle in the Amhr2-Cre/+;Notch2^flox/- ovaries. By 7 weeks of age, the follicles in the Amhr2-Cre/+;Notch2^flox/- mutant ovaries became hemorrhagic (Figure 2D, F).

The presence of multi-oocyte follicles in Amhr2-Cre/+;Notch2^flox/- mice suggests that Notch2-mediated signaling may be required for breakdown of the primordial germ-cell nests. In wild-type mice, breakdown of germ-cell nests and formation of primordial follicles are normally complete within a few days of birth. Histological analyses on ovaries isolated at postnatal day (PND) 8 from Amhr2-Cre/+;Notch2^flox/- and control littermate mice showed that, whereas virtually all germ-cell nests in the control mice had broken down to form primordial follicles, there was persistence of germ-cell nests in the cortex of the Amhr2-Cre/+;Notch2^flox/- ovaries (Figure 3B). This was confirmed by immunofluorescence staining of ovaries isolated at PND3, using an antibody to the cytoplasmic germ-cell marker Vasa (Ddx4). In littermate control ovaries at PND3, most Vasa-positive cells had already been assembled into individual primordial follicles (Figure 3C). However, in the Amhr2-Cre/+;Notch2^flox/- ovaries, most Vasa-positive cells, particularly near the cortical surface of the ovary, remained in germ-cell nests (Figure 3D). By PND3, laminin immunostaining showed that in control littermate mice, the primordial follicles in the cortex of the ovary were surrounded by a basement membrane (Figure 3E), whereas in Amhr2-Cre/+;Notch2^flox/- ovaries, the oocytes remained in the germ-cell nests and were not individually surrounded by a basement membrane (Figure 3F). These data indicate that Notch2 gene function in pre-granulosa cells is required for breakdown of the primordial germ-cell nests and assembly of primordial follicles.

To quantify oocyte numbers in Amhr2-Cre/+;Notch2^flox/- and control littermate mice, we counted serial sections of ovaries isolated at PND18. This time point was chosen because it is a point shortly before the female mouse reaches sexual maturity. Amhr2-Cre/+;Notch2^flox/- mice had approximately 20% more oocytes than ovaries from littermate controls (Table 2). We then quantified the type of follicle present (that is, primordial, primary, secondary, pre-antral, and antral) and assessed whether these follicles contained a single oocyte, two or three oocytes, or four or more oocytes (Table 3, Figure 4). For all follicle types, only the Amhr2-Cre/+;Notch2^flox/- mice had four or more oocytes per follicle present in the ovaries (Figure 4). We divided the follicles into two categories: the resting pool (consisting only of primordial follicles) and the growing pool (primary, secondary, pre-antral, and antral follicles). The PND18 Amhr2-Cre/+;Notch2^flox/- ovaries contained significantly fewer total follicles (40.6% of the control mice), and also significantly fewer follicles in the resting pool (only 16.7% of the control) (Table 3). A significantly reduced number of normal primordial follicles was seen in the Amhr2-Cre/+;Notch2^flox/- ovaries. The mutant mice had approximately nine-fold fewer primordial follicles containing a single oocyte than did their littermate controls (11.4% of control), and had 13 times more primordial follicles containing four or more oocytes than controls (Figure 4A). Regarding the growing pool, the Amhr2-Cre/+;Notch2^flox/- ovaries had significantly more primary and secondary follicles containing four or more oocytes, and also had more follicles of the primary, secondary, and pre-antral types containing two to three oocytes (Figure 4B-D). However, at PND18 the total number of follicles in the growing pool (primary, secondary, pre-antral, and antral) was not significantly different in the Amhr2-Cre/+;Notch2^flox/- and their littermate controls (Table 3).

We used the PND18 oocyte quantification data to estimate the fraction of primordial germ-cell nests that failed to break down properly in Amhr2-Cre/+;Notch2^flox/- mice. We combined the oocyte numbers from all follicle types (primordial to antral), and assessed in both control littermate and Amhr2-Cre/+;Notch2^flox/- mice (n = 4 for each group) whether the follicles contained a single oocyte, or more than one oocyte (multi-oocyte follicle). In the control mice, 96.9% of the follicles contained a single oocyte, and the remaining 3.1% (399/12865) contained more than one oocyte. In the Amhr2-Cre/+;Notch2^flox/- mutants, 60.3% of the follicles contained a single oocyte, and the remaining 39.7% (2074/5223) contained more than one oocyte. Assuming that all multi-oocyte follicles resulted from a failure of primordial germ-cell-nest breakdown, these data suggest that approximately 40% of the primordial germ-cell nests in the Amhr2-Cre/+;Notch2^flox/- mice failed to break down and assemble into follicles with a single oocyte each.

To address the mechanisms responsible for the 20% increase in oocyte numbers in Amhr2-Cre/+;Notch2^flox/- mice, we assessed cell proliferation (by 5-bromo-2'-deoxyuridine (BrdU) incorporation) and apoptotic cell death (by terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate (TUNEL) assay) in ovaries of Amhr2-Cre/+;Notch2^flox/- and control littermate mice. In wild-type mouse embryos, oogonia cease mitosis and enter meiosis at approximately embryonic day 13.5. We assessed whether oocytes were still proliferating in Amhr2-Cre/+;Notch2^flox/- embryos during the period E16.5 to E18.5, using BrdU injection into timed pregnant females. Ovaries from the embryos of the BrdU-injected females were isolated and sectioned, and oocytes were identified by immunofluorescence staining with an antibody to the cytoplasmic germ-cell marker Vasa. We identified no oocytes throughout the period E16.5 to E18.5, in either Amhr2-Cre/+;Notch2^flox/- embryos or their littermate controls, which were double-stained with the anti-Vasa and anti-BrdU antibodies (Figure 5, and data not shown). We next examined cell proliferation in the ovaries of neonatal mice at PND1 and PND2 by BrdU injection at these time points. Quantifying either all BrdU-positive cells (Figure 6A) or BrdU-positive pre-granulosa/granulosa cells (which we defined as a BrdU-positive cell directly adjacent to a Vasa-positive oocyte) (Figure 6B), we detected no significant differences in the numbers of proliferating cells between the control and the Amhr2-Cre/+;Notch2^flox/- ovaries at either PND1 or PND2. Similarly to the embryonic BrdU exposure (Figure 5), we did not find any oocytes that were double-stained with the anti-Vasa and anti-BrdU antibodies, in either Amhr2-Cre/+;Notch2^flox/- embryos or their littermate controls, at PND1 or PND2 (Figure 6, and data not shown). Therefore, it did not appear that excess or extended proliferation was the cause of the increase in oocyte numbers in Amhr2-Cre/+;Notch2^flox/- mice.

We next assessed apoptotic programmed cell death using TUNEL (Figure 7). Counting either TUNEL-positive oocytes (defined both by double staining with the TUNEL reaction and anti-Vasa antibody, and by the distinct morphology of the oocyte nucleus) (Figure 7A) or TUNEL-positive pre-granulosa/granulosa cells (defined as a TUNEL-positive cell directly adjacent to a Vasa-positive oocyte) (Figure 7B), we detected significantly fewer apoptotic cells of either type in Amhr2-Cre/+;Notch2^flox/- mutant ovaries at PND1. By PND2, the numbers of apoptotic cells (both oocytes and pre-granulosa/granulosa cells) were not significantly different in Amhr2-Cre/+;Notch2^flox/- and control ovaries. These data support a model in which the increase in oocyte number in Amhr2-Cre/+;Notch2^flox/- mutant ovaries is caused by reduced oocyte apoptotic cell death during the early postnatal stages.

All animal experiments were carried out with strict adherence to the National Institutes of Health (NIH) Guidelines for animal care and safety, and were approved by the Animal Care and Use Committee of the Jackson Laboratory (where the experiments were performed).

The Amhr2-Cre knock-in allele 19 was used to drive Cre recombinase expression in granulosa cells of the ovary. The Amhr2-Cre allele is expressed by E11.5 in both male and female urogenital ridges, and by E12.5, expression is localized to somatic cells of the gonads and mesenchyme cells of the Müllerian ducts 19. At E17.5, Amhr2-Cre expression is present in almost all somatic cells of the ovary 40. Construction and utilization of the Notch2 conditional allele (Notch2^flox) and Notch2 null allele (Notch2^del3) have been described previously 1741. To generate mice with targeted disruption of the Notch2 gene in ovarian granulosa cells, mice heterozygous for the Notch2 null allele and the Amhr2-Cre knock-in allele were crossed with mice homozygous for the Notch2^flox allele. Mice were genotyped by PCR. The Notch2 null allele was detected using the primers. Notch2 forward and Notch2 reverse (Table 4), and the Notch2^flox allele was detected using the primers Flox forward and Notch2 reverse (Table 4).

To assess fertility, 7-week-old Amhr2-Cre/+;Notch2^flox/- and littermate control females (either Notch2^flox/-, Notch2^flox/+, or Amhr2-Cre/+;Notch2^flox/+) were paired with a single male C57BL6/J mouse 7 to 8 weeks old. Cages were monitored daily, and the numbers of litters and litter sizes were recorded for a 6-month period.

Ovaries were removed, fixed overnight in Bouin's fixative for hematoxylin and eosin or Periodic-Acid-Schiff (PAS) staining, or at 4°C in 4% paraformaldehyde for immunohistochemistry. Fixed ovaries were embedded in paraffin wax, and cut into sections 6 μm thick. Sections used for immunohistochemistry were washed with phosphate-buffered saline and boiled for 10 minutes in 10 μmol/l sodium citrate (pH 6.0) for antigen retrieval. Sections were then incubated overnight at 4°C with primary antibodies (1:100 dilution) directed against: BrdU (BD Pharmingen, San Jose, CA, USA), Vasa (Ddx4) (Abcam, Cambridge, MA, USA), AMH (MIS, Santa Cruz, Dallas, TX, USA), and laminin (Sigma, St. Louis, MO, USA). After washing, sections were incubated with secondary antibodies, either Alexa fluor 488-conjugated or 546-conjugated donkey anti-mouse, anti-goat, or anti-rabbit (1:200; Invitrogen, Carlsbad, CA, USA), for 2 hours at room temperature, washed, and developed. Negative controls were prepared by omitting the primary antibody, and no staining above the background was detected. Sections from at least three mice of each genotype were processed for comparison of immunostaining, and immunohistochemical studies were repeated three to five times to ensure reproducibility of results.

To examine germ-line cell proliferation, timed pregnant female mice were injected with a BrdU solution (Sigma, 10 mg/ml in 50:50 PBS/DMSO; final dose, 3 mg per mouse), twice daily (at 8-hour intervals) between E16.5 and E18.5. Following euthanasia, the ovaries were fixed and embedded in paraffin wax for sectioning. To study cell proliferation in mice on PND1 and PND2, female mice were injected intraperitoneally with BrdU solution at a final dose of 100 mg/kg (normal body weight approximately 2 g, hence 20 μl each). After 2 hours, the ovaries were harvested. Tissues were embedded in paraffin wax, and serial sections were cut at 6 μm. Sections were dewaxed and rehydrated, and stained with anti-BrdU and anti-Vasa antibodies. The total number of BrdU-positive cells in every 15th section was determined. Granulosa cells were defined as cells adjacent to the Vasa-positive germline cells.

Ovaries of PND18 mice were sectioned at 6 μm and stained with PAS. The follicles in each ovary were counted serially in every third section through the entire ovary. Only healthy, non-atretic follicles with visible oocyte nuclei were scored. Follicles were classified as primordial, primary, secondary, pre-antral, or antral. Primordial follicles had a compact oocyte surrounded by a single layer of flattened granulosa cells; primary follicles had an enlarged oocyte surrounded by a single layer of cuboidal granulosa cells; secondary follicles had an enlarged oocyte surrounded by two layers of granulosa cells; and pre-antral follicles had an enlarged oocyte surrounded by three or four layers of granulosa cells. Any follicles with antral spaces were considered antral follicles. Total oocyte numbers at PND18 were counted serially in every 15th section through the entire ovary by counting every oocyte in the section. Total oocyte numbers at PND1 and 2 were counted serially in every 15th section through the entire ovary by counting every oocyte with positive Vasa and DAPI staining in the section.

For terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end-labeling (TUNEL) assay, ovaries were collected on PND1 and PND2, and fixed in cold 4% paraformaldehyde for 1 hour. Tissues were embedded in paraffin wax, and serially sectioned at 6 μm. Detection of apoptotic cells was carried out using a TUNEL apoptosis detection kit (Roche Applied Science, Indianapolis, IN, US), with co-staining using an anti-Vasa antibody. The total number of apoptotic oocytes and granulosa cells in every 15th section was counted.

Data are presented as mean ± SD. A two-tailed t-test was performed to compare means between two groups, and one-way analysis of variance (ANOVA) was performed to compare means between multiple groups. P≤0.05 was considered significant.