Sustained fertility from first-wave follicle oocytes that pause their growth

Abstract

Ovulation results from the cyclical recruitment of non-renewing, quiescent oocytes for growth. Therefore, the primordial follicles that are established during development from an oocyte encapsulated by granulosa cells are thought to comprise the lifelong ovarian reserve ^1-4. However, using oocyte lineage tracing in mice, we observed that a subset of oocytes recruited for growth in the first juvenile wave remain paused for many months before continuing growth, ovulation, fertilization and development into healthy offspring. This small subset of genetically-labeled fetal oocytes, labeled with Sycp3-CreERT2, is distinguished by earlier entry and slower dynamics of meiotic prophase I. While labeled oocytes were initially found in both primordial follicles and growing follicles of the first wave, they disappeared from primordial follicles by puberty. Unexpectedly, these first-wave labeled growing oocytes persisted throughout reproductive lifespan and contributed to offspring at a steady rate beyond 12 months of age, suggesting that follicles can pause mid-growth for extended periods then successfully resume. These results challenge the conclusion from lineage tracing of granulosa cells that first-wave follicles make a limited contribution to fertility⁵ and furthermore suggest that growth-paused oocytes comprise a second and previously unrecognized ovarian reserve.

Extended Data Fig. 1.. Spatial dynamics of additional markers of meiotic progression and validation of the Sycp3-CreERT2 lineage tracing system.

(a) SYCE2 and (b) HORMAD1 expression exhibit no spatial bias in wholemount immunostained E13.5 wild-type ovaries. Scale bars, 50 μm. c, The targeting vector that was co-injected with CAS9 and sgRNA into FVB/N zygotes to generate the Sycp3-CreERT2 line. d, Genotyping of founder mice positive for targeting (#5, #7, #17 in bold). L; ladder, US; Upstream junction, DS; Downstream junction. e, Illustration of the tamoxifen dosing regimen designed to evaluate the specificity and efficiency of the Sycp3-CreERT2 line. Upon intraperitoneal tamoxifen injection at E11.5, the activated CreERT2 recombinase excises membrane-anchored tdTomato (mT) in Sycp3-expressing cells and initiates EGFP (mG) expression in cells, which leads to constitutive expression of membrane GFP. f, Whole-mount immunofluorescence detection of membrane GFP expression, observed in nearly all SYCP3 and VASA expressing oocytes at E14.5. Scale bars, 50 μm. g, Schematic for low dose oral administration of 4-OHT at E11.5, one day before meiotic initiation, which resulted in robust labeling in Sycp3^CreERT2/+; Rosa26^nT-nG/+ ovaries at E18.5, shown at (h) low magnification, scale bars; 10 μm (i) as GFP was detected in over 50% of TRA98+ oocytes. Scale bars, 50 μm. (i’) GFP+ cells also expressed endogenous SYCP3 protein, confirming their meiotic status. Scale bars, 5 μm. j, 4-OHT administration at E11.5 had no effect on the wave of radial meiotic initiation as evaluated by endogenous SYCP3 protein expression analysis at E13.5. SYCP3 positive cells were predominantly located in the anterior (bins 1 to 3 on x axis of A-P distribution graph; ***P=0.0001) and middle (bins 3 to 5 on x axis of M-L distribution graph; *P=0.0358) of E13.5 Sycp3^CreERT2/+; Rosa26^nT-nG/+ ovaries (n=4). Scale bars, 50 μm.

Extended Data Fig. 2.. Chromosomal analysis and spatiotemporal dynamics of apoptosis in Sycp3-CreERT2 labeled oocytes.

a, Experimental design for labeling the first meiotic entrants with membrane GFP, followed by downstream chromosomal analysis and (b) numbers of cells sorted for each (GFP^high, GFP^low, and GFP^negative) group. Note that only GFP+ gates contain pure oocytes, while oocytes in GFP^negative gate were identified in meiotic spreads. c, Immunofluorescence labeling of SYCP3 and MLH1 and the frequency distribution of MLH1 foci across each synaptonemal complex (SC) within single nuclei. SCs were classified based on the number of MLH1 foci: those without an MLH1 focus were termed “E0” (0 exchanges), with one focus “E1” (1 exchange), with two foci “E2” (2 exchanges), and with three foci “E3” (3 exchanges). For E0, Chi square = 3.74, p = 0.15; for E1, Chi square = 0.44, p = 0.8; for E2, Chi square = 1.88, p = 0.39; and for E3, Chi square = 2.37, p = 0.31 as determined by Fisher’s Exact Test. GFP^negative oocytes exhibited a slightly lower frequency of E0s by comparison with GFP^low and GFP^high oocytes (1.3%, 2.7%, and 2.0%, respectively), though this difference was not statistically significant (Chi square = 3.74; p = 0.15). d, Pachytene oocytes were stained for SYCP3 and SYCP1 to assess the synapsis quality: complete synapsis (SYCP1 and SYCP3 perfectly co-localize without errors), partial asynapsis (white arrow pointing to a forked region of asynapsis), complete asynapsis (white arrow pointing to two axial elements without any SYCP1), and fragmented SC (SYCP1 and SYCP3 co-localize, but the SCs appear broken). e, Section view from whole-mount imaging of E18.5 Sycp3^CreERT2/+; Rosa26^nT-nG/+ ovaries, white arrows show GFP and cPARP co-expressing TRA98+ oocytes. Scale bars, 10 μm. f, cPARP+ oocytes were randomly distributed in E18.5 Sycp3^CreERT2/+; Rosa26^nT-nG/+ ovaries. g, Section view from whole-mount imaging of P5 Sycp3^CreERT2/+; Rosa26^nT-nG/+ ovaries. Dashed white circles delineate growing follicles. cPARP expression is concentrated within primordial follicles (arrows) at P5. Scale bars, 10 μm. h, A higher proportion of TRA98+ oocytes in the posterior (**p=0.0019) and lateral regions (cortex; *p=0.0261) of P5 ovaries exhibited cPARP expression.

Extended Data Fig. 3.. Localization studies of Sycp3-CreERT2 lineage traced oocytes in postnatal and adult ovaries.

a, Distribution of GFP+ oocytes in Sycp3^CreERT2/+; Rosa26^nT-nG/+ ovaries at P5. (b) and (c) High magnification images showing GFP+ growing follicles (arrows) and GFP+ primordial/primary follicles (arrowheads) in P16 and P21 whole-mount stained ovaries from Sycp3^CreERT2/+; Rosa26^nT-nG/+ mice. Oocytes in prepubertal ovaries were labeled with a NOBOX antibody. Scale bars are 5 μm for P16 and 20 μm for P21. d, High magnification images of NOBOX (top, magenta), VASA (middle, gray), and AMH (bottom, cyan) expressing GFP+ growing follicles (arrows) in Sycp3^CreERT2/+; Rosa26^nT-nG/+ ovaries at 2-month. Scale bars, 10 μm. e, Detection and quantification of GDF9 and GFP expression at 2-month (n=5). Scale bar, 50 μm. f, Whole-mount immunofluorescence and quantification of NOBOX, VASA, and GFP expressing growing follicles in Sycp3^CreERT2/+; Rosa26^nT-nG/+ ovaries (n=6) at 6-month. Scale bars, 50 μm. g, High magnification images showing GFP+ growing follicles (arrows) in 12-month-old ovaries expressing NOBOX (top, magenta) and AMH (bottom, cyan).

Extended Data Fig. 4.. Transcriptional signature of mature Sycp3-CreERT2 lineage traced oocytes collected at P21 is unremarkable.

a, Experimental approach to isolate GFP+ and RFP+ (GFP-) MII oocytes for Bulk RNA-seq analysis. Sycp3^CreERT2/+; Rosa26^nT-nG/+ females exposed to 4-OHT in utero were superovulated at P21 and MII oocytes were manually selected for transcriptome analysis. Two fluorescence intensities were detected in MII GFP+ oocytes: GFP high and GFP low. Oocytes with low GFP signal were kept as a separate group, and transcriptomic analysis was carried out on the three groups (GFP high, GFP low, and RFP+). Further analysis was primarily focused on the GFP high and RFP+ groups. b, The standardized average of readcount for each group. c, Fragments Per Kilobase per Million mapped fragments (FPKM) for each sample. d, A volcano plot showing a very small number of differentially expressed genes (8 upregulated and 4 downregulated) between GFP high and RFP+ oocytes (adjusted p-value ≤ 0.05). e, Differentially expressed genes between GFP high vs RFP, GFP low vs RFP, and GFP high vs GFP low. Immunofluorescence labeling of (f) GPX4 and (g) PRAMEF17 proteins in MII oocytes revealed similar expression in both GFP high and RFP+ oocytes. Scale bars, 10 μm.

Fig. 1.. The earliest differentiating oocytes are distinct in their spatial distribution and meiotic progression.

a, Schematic of earliest SYCP3 expressing cells (magenta) in the anterior-medial region of the E12.5 ovary, radiating outward by E14.75. For analysis, the ovary was divided into seven segments. Y-axes depict the number of germ cells in each bin, dashed lines represent individual ovaries, and color-coded bold lines indicate mean values for each antibody. Secondary y-axes (gray) show the percentage of cells positive for antibody of interest (i.e. SYCP1 in yellow) normalized to total germ cells in each bin. n=4 wild-type ovaries for each time point. Scale bars, 70 μm. b, Illustration of induced in vivo lineage tracing of the first meiotic entrants. Following 4-OHT feeding at E9.5, CreERT2 recombinase excises nT and initiates nG expression, causing constitutive expression of nuclear GFP in SYCP3-expressing cells. c, Whole-mount immunolabeling of TRA98 and GFP (white arrows) in Sycp3^CreERT2/+; Rosa26^nT-nG/+ ovaries at E16.5 and E18.5. GFP labeled first meiotic entrants were predominantly located in the anterior (bins 1 to 3 on x axis of A-P distribution graph; ***P=0.0005 for E16.5 and ****P<0.0001 for E18.5) and middle (bins 3 to 5 on x axis of M-L distribution graph; ****P<0.0001 for E16.5 and ***P=0.0005 for E18.5) of E16.5 (n=10) and E18.5 (n=5) Sycp3^CreERT2/+; Rosa26^nT-nG/+ ovaries. In graphs, distributions of GFP are shown in green. Scale bars are 50 μm for the whole ovary view, 8 μm for higher magnification at E16.5, and 10 μm for higher magnification at E18.5. d, Quantification of TRA98-positive total germ cells and GFP expressing first meiotic entrants at E16.5 (n=10) and E18.5 (n=7). e, FACS isolation of GFP^high GFP^low, and GFP^negative cells from Sycp3^CreERT2/+; Rosa26^mT-mG/+ ovaries at E17.5. f, MPI staging of chromosome spreads from nuclei of sorted cells in (e) following SYCP1/3 immunostaining (100 nuclei examined in each population, a total of n=900 cells were analyzed).

Fig. 2.. Sycp3-CreERT2 labeled early meiotic entrants are more resistant to FOA.

Quantification at E17.5 of (a) persistent double-strand breaks with RAD51 (n=40 nuclei for GFP^high, n=74 for GFP^low, and n=81 for GFP^negative), (b) class I crossovers with MLH1 (n=10 nuclei for GFP^high, n=43 for GFP^low, and n=38 for GFP^negative), and (c) synaptic defects with SYCP3/1 (n=49 nuclei for GFP^high, n=98 for GFP^low, and n=81 for GFP^negative). d, Whole-mount immunolabeling of cells expressing TRA98, cPARP, and GFP in E18.5 Sycp3^CreERT2/+; Rosa26^nT-nG/+ ovaries following E9.5 4-OHT administration. White lines delineate ovary borders and white arrows indicate a cPARP+ GFP+ TRA98+ oocyte. Scale bars are 50 μm for whole ovary view (top) and 10 μm for higher magnification (bottom). e, Timeline for 4-OHT administration and evaluation. f, Quantification of TRA98+ (magenta) and GFP+ cells in E18.5 Sycp3^CreERT2/+; Rosa26^nT-nG/+ ovaries following 4-OHT at E9.5 (labeling first meiotic entrants, n=5 ovaries) and E14.5 (random oocyte labeling, n=5 ovaries). The number of TRA98+ oocytes in each labeling condition was similar by T-test. g, Frequency of cPARP colocalization with TRA98+ oocytes, GFP+ and GFP-negative oocyte subsets at E18.5 following 4-OHT dosing at E9.5 or E14.5. Frequencies by labeling time were compared by Chi-square test whereas the p-value for the interaction term of 4.94 computed between 4-OHT timing, cPARP and GFP is shown above with Z-statistic. h, Whole-mount immunostaining and (i) quantification of GFP, LINE-1 ORF1p and TRA98 in E18.5 Sycp3^CreERT2/+; Rosa26^nT-nG/+ ovaries following (h’) 4-OHT administration at E9.5. Scale bars, 5 μm. LINE-1 ORF1p expression was predominantly observed in GFP-TRA98+ oocytes (white dashed lines) compared to GFP+ TRA98+ oocytes. j, Whole-mount immunostaining of TRA98, cPARP, and GFP in P5 Sycp3^CreERT2/+; Rosa26^nT-nG/+ ovaries following (j’) 4-OHT at E9.5. Scale bars, 50 μm. k and l, Quantification of TRA98 and GFP+ cells revealed lower frequency of cPARP positivity in GFP+ compared to GFP-negative oocytes.

Fig. 3.. GFP expression from Sycp3-CreERT2 disappears from primordial/primary follicle oocytes by puberty but persists in growing follicle oocytes for up to 12 months.

a, Whole-mount immunostaining and (b) quantification of growing follicles using AMH (granulosa cell marker) expression and low TRA98 expression in Sycp3^CreERT2/+; Rosa26^nT-nG/+ ovaries (n=6) revealed a comparable percentage of GFP+ oocytes (c) in both growing and primordial/primary follicles at P5. Ovarian medulla is outlined by the yellow dashed line, with white arrows indicating AMH+ growing follicles containing GFP+ oocytes. The scale bars are 50 μm for the whole ovary view (left), 30 μm for the higher magnification (middle), and 15 μm for the primordial/primary and growing follicles (right). d, At P16 and (e) P21, whole-mount immunostaining against oocyte marker NOBOX and GFP showed a dramatic reduction in primordial/primary follicles containing GFP+ oocytes in the cortex (n=6 for each timepoint). Scale bars, 50 μm. Contribution of GFP+ oocytes to the dynamic pool of growing follicles (f) in 2-month (n=13) and (g) 12-month-old (n=9) Sycp3^CreERT2/+; Rosa26^nT-nG/+ ovaries, assessed using NOBOX and VASA (oocyte markers) and AMH (granulosa cell marker). Scale bars, 50 μm. Each data point in b, d, e, f, and g corresponds to an individual ovary and the percentage of GFP+ oocytes are displayed at the top of the graphs. Quantifications of GFP+ oocytes (h) in primordial/primary (P5 to P21) and (i) growing (P5 to 12-month) follicles highlighted the disappearance of GFP+ primordial/primary follicles, while GFP+ oocytes in growing follicles were observed until 12-months. The primary y-axes display the total number of GFP+ oocytes, while the secondary y-axes represent the percentage of GFP+ oocytes within the total oocyte population. The x-axes indicate various time points.

Fig. 4.. Sycp3-CreERT2 oocytes labeled at the onset of meiosis contribute to offspring at a constant frequency throughout reproductive lifespan.

a, Sycp3^CreERT2/+; Rosa26^nT-nG/+ females exposed to 4-OHT at E9.5 (n=14) were crossed with wild-type males and pups were screened to detect embryos derived from GFP+ oocytes (left), RFP+ oocytes (middle) or GFP-;RFP- unlabeled oocytes (right). b, Percentage of all pups expressing fluorescent protein (FP). c, Percentage of GFP+ pups among FP-expressing pups remained similar across litters despite increasing maternal age. d, Half of females ceased reproducing after the ninth litter. e, Ages of the females and (f) the number of pups produced per female in each litter. The number of pups per litter decreased in the 9th (**P=0.0038) and 10th (*P=0.0381) litters compared to the first litter. g, Cumulative number of GFP+ and RFP+ pups produced by each female (circles, with mean indicated as gray line) upon reaching 12 months of age. h, Model of paused follicle growth in mouse ovaries; (top) the paradigm of continuous follicle growth in which growing follicles are cyclically recruited from primordial follicles and either undergo atresia or reach ovulation stage which predicts that GFP-labeled first-wave growing oocytes should be rapidly depleted, with ovulation in adulthood and later reproductive life resulting from continuous recruitment from unlabeled primordial follicles. However, detection of GFP+ pups up to 12 months of age, alongside persistence of GFP labeled growing oocytes/the loss at puberty of GFP labeled primordials, leads to the model in which growing follicles in the medulla derived from the first meiotic entrants can pause, and later resume growth and maturation to contribute to offspring until 12-months (bottom).