The Aging Ovary and the Tales Learned Since Fetal Development

While the term “aging” implies a process typically associated with later life, the consequences of ovarian aging are evident by the time a woman reaches her forties, and sometimes earlier. This is due to a gradual decline in the quantity and quality of oocytes which occurs over a woman’s reproductive lifespan. Indeed, the reproductive potential of the ovary is established even before birth, as the proper formation and assembly of the ovarian germ cell population during fetal life determines the lifetime endowment of oocytes and follicles. Menopause is the culmination of the ovarian aging process wherein follicles have been nearly exhausted. Menopause represents a major hormonal, psychological, and physiological event and confers an increased risk for comorbidities and overall mortality. The effects of ovarian aging on overall health and lifespan are due, then, to progressive, irreversible loss of follicles and oocytes. Furthermore, deterioration of gamete quality occurs with age, increasing the risk of birth defects and miscarriage. Importantly, human chronological lifespan has extended dramatically over the last century, while the timing of menopause has remained relatively constant [Nichols et al., 2006; Rocca et al., 2006]. This means that many women now spend a larger portion of their lives in the postmenopausal period, increasing the incidence of comorbidities. In women with primary ovarian insufficiency (POI), this process is catastrophically accelerated, causing loss of fertility and early menopause [Nelson, 2009].

The mechanisms thought to contribute to overall aging have been summarized under the term the “hallmarks of aging” and include such processes as mitochondrial dysfunction, telomere attrition, genomic instability, and stem cell exhaustion, among others. These processes are viewed as potential targets for the development of preventatives or therapeutics for age-related diseases [López-Otín et al., 2013]. Similarly, in the ovary, molecular processes have been identified that regulate the timing of ovarian aging and these are critical to ensuring follicular maintenance. As we will discuss, some of the general “hallmarks of aging” have been shown to regulate ovarian aging as well. In this review, we will outline critical processes involved in ovarian aging, highlight major achievements for treatment of ovarian aging, and discuss ongoing questions and areas of debate. The manuscripts included in this minireview encompass publications between January 2010 and August 2022. A combination of subject headings and keywords, including ovarian aging, DNA damage ovary, oocyte DNA damage repair (DDR), menopause, senescence, sirtuins, antioxidants, and oxidative stress, was used to identify relevant publications. Based on these criteria, 116 primary research articles were identified. We manually searched 22 subject-specific reviews for additional relevant publications.

As mentioned, ovarian aging includes both the loss of primordial follicles over time and also declining egg quality where the capability of eggs to be fertilized and give rise to healthy offspring declines. Although the median age of menopause is 51 years of age, egg quality is well-known to become significantly compromised much earlier, late in the third decade of life. We first discuss mechanisms that impact the rate of loss of primordial follicles and then move on to those that regulate DDR and chromosomal stability. Notably, there is some overlap in these mechanisms that regulate oocyte quantity and (egg) quality, particularly in genes that regulate DNA repair.

During folliculogenesis, primordial follicles develop to primary, preantral, antral and preovulatory stages, at which point they are able to release a mature oocyte for fertilization. Cohorts of growing follicles are continuously recruited from the large pool of primordial follicles that reflects the ovarian reserve. Most recruited follicles are destined to die via atresia, and this means that only about 450 follicles will mature and reach ovulation in the normal human reproductive lifespan (although the number of follicles lost over a lifespan is orders of magnitude higher than that) [Wallace and Kelsey, 2010].

Both inhibitory and activating pathways are known to function at the level of primordial follicles that regulate their timing of entry into the growing follicle pool. The rate of primordial follicle activation (PFA) determines the remaining ovarian reserve and reproductive lifespan. Consequently, much attention has been paid to the identification of gene expression pathways which regulate PFA and recruitment, with an eye toward strategies to activate dormant primordial follicles and restore ovarian function.

Central regulators of PFA include the PI3K/AKT/mechanistic target of rapamycin (mTOR) pathway, a prominent regulator of cell survival, growth, and migration. A key enzyme in this pathway is phosphatase and tensin homolog deleted on chromosome ten (PTEN), a dual protein/lipid phosphatase and master upstream that has been shown to function as a regulator of primordial follicle recruitment [Reddy et al., 2008; Jagarlamudi et al., 2009]. The main substrate of PTEN is the product of PI3K, phosphatidylinositol-3,4,5 triphosphate (PIP3); high levels of PIP3 recruit AKT to the cell membrane for activation [Maidarti et al., 2020]. Activation of AKT initiates localization of transcription factors which suppress follicle activation (e.g., FOXO3) [Kallen et al., 2018].

Consequently, PTEN-specific inhibitors have been used as pharmacologic strategies to activate AKT, by blocking PTEN-mediated AKT suppression [Adhikari et al., 2012]. While in vitro activation of primordial follicles using pharmacologic PTEN inhibition in human ovarian tissue has been shown to increase growth activation of primordial follicles, isolated and cultured follicles treated with PTEN inhibitor exhibit limited growth and reduced survival compared to untreated follicles, suggesting that PTEN inhibition severely compromises follicle survival and the need for caution before applying this and other PFA strategies to a wider patient pool [Li et al., 2010; McLaughlin et al., 2014; Novella-Maestre et al., 2015].

AKT signaling activates downstream mTOR action. mTOR forms two multimolecular complexes, mTORC1 and mTORC2 [Johnson et al., 2013; Cheng et al., 2015] that regulate downstream protein translation, and cytoskeletal interactions, respectively. While mTOR signaling in oocytes contributes to PFA, it is not a requirement for the initial transition from primordial to primary follicles. This is evidenced by the fact that conditional mTOR knockout in primordial follicle oocytes does not inhibit follicular growth, although compensatory PI3K signaling may underlie this outcome [Gorre et al., 2014], or, action within somatic granulosa cells may be responsible for the effect. However, mTOR inhibition with small molecule inhibitors in mice has been shown to protect ovarian reserve, primordial follicles, anti-Müllerian hormone (AMH) levels, and fertility in cyclophosphamide-treated mouse model [Goldman et al., 2017; Guo and Yu, 2019].

Next, we consider the Hippo signaling pathway, an evolutionarily conserved pathway that regulates cell proliferation, apoptosis, and stem cell self-renewal [Csibi and Blenis, 2012]. Hippo signaling restrains follicle growth, while mechanical Hippo disruption promotes follicle growth [Grosbois and Demeestere, 2018]. Thus, mechanical ovarian fragmentation has been used as a strategy to activate resting follicles in vitro. For example, ex vivo fragmentation of mouse ovaries followed by reimplantation of fragmented tissue increases the percentage of late secondary and antral follicles, although an overall loss of follicles occurs after grafting. Twenty-seven patients with POI underwent ovarian tissue harvesting and fragmentation of harvested tissues, followed by in vitro AKT treatment and autotransplantation of the fragmented tissue. While follicle growth was subsequently observed in 8 patients and mature oocytes were retrieved from 5 patients, only one woman achieved a live birth; this success rate is similar to the spontaneous pregnancy rate in women with POI [Kawamura et al., 2013].

Finally, AMH represents a critical “gatekeeper” hormone which regulates follicular quantity by inhibiting recruitment and growth. AMH is involved in two key steps of follicle development: inhibition of primordial follicle recruitment, and suppression of small follicle growth by reducing the follicle’s sensitivity to follicle-stimulating hormone (FSH) [Durlinger et al., 1999, 2001, 2002]. The use of exogenous AMH represents an innovative potential preventative strategy and therapeutic option for gonadotoxic-induced follicle loss. In a recent study, co-treatment with exogenous delivery of AMH in chemotherapy-exposed mice resulted in inhibition of PFA, though the protective effect varied between agents used [Kano et al., 2017]. Because the PI3K signaling pathway is not activated by AMH and therefore phosphorylation of FOXO3 (which is conserved in human granulosa cells [Zhang et al., 2020]) has been postulated as a causative mechanism for AMH’s protective effects [Sonigo et al., 2019].

It is important to note that in our summary of pathways that control the rate of PFA, in some cases, the oocyte has been implicated (as in oocyte-specific knockout models) as the key cell type that controls whether a primordial follicle begins to grow or stays dormant. In other cases, it appears that it is the somatic (pre)granulosa cells of primordial follicles that control the decision. We next consider ovarian aging at the level of oocytes, and signals and mechanisms that control oocyte development such that high-quality eggs can be produced. While the focus of oocyte aging will be upon events that take place within oocytes, somatic follicle cells including granulosa cells utilize many of the same mechanisms to manage their own DNA damage (see Fig. 1).

DDR pathways have emerged from large-scale human population analyses [Stolk et al., 2012; Day et al., 2015; Huhtaniemi et al., 2018] and from individual cellular and molecular investigations as key regulators of follicle survival and of oocyte quality control. During female mammalian meiosis, which begins during fetal life, gametes undergo a single round of DNA replication followed by two rounds of chromosome segregation to produce haploid oocytes. This process arrests at meiotic prophase I during gestation until oocytes recommence meiosis decades later or undergo atresia [Gebel et al., 2020]. Indeed, prophase-arrested oocytes are among the longest-living cells in the human body. Before entering prophase arrest, oocytes are subject to a series of programmed DNA double-strand breaks (DSBs) generated by the topoisomerase-like enzyme SPO11 [Rinaldi et al., 2017]. Oocytes undergoing meiotic recombination are highly tolerant of DSBs; repair of these DSBs through homologous recombination generates inter-homolog crossovers and maintains genomic integrity [Winship et al., 2018; Stringer et al., 2020]. To identify and address DSBs, cells utilize the ataxia-telangiectasia mutated (ATM) DDR pathway [Titus et al., 2013b]. Breaks in the genome are first marked by phosphorylation of γH2AX, an indicator of DNA damage. Following this, RAD51 (RAD51 recombinase, a DNA strand exchange protein that promotes homologous single-stranded DNA intrusion into DSBs) and BRCA1 (Breast cancer gene 1, a tumor suppressing protein with multiple active binding domains implicated in cell cycle and transcriptional regulation and DNA repair) [Prakash et al., 2015] localize to these sites to initiate repair [Titus et al., 2013a]. If repairs cannot be made, oocytes are targeted for apoptosis and ultimately follicular atresia [Stringer et al., 2020].

TAp63, a member of the p53 family of proteins, is essential for DNA-damage-induced apoptosis in primordial follicle oocytes [Stringer et al., 2020] and highly expressed in primordial follicles [Hu, 2009]. Integrating different DNA damage signals in the oocyte, TAp63 serves as a transcription factor upregulating the expression of the proapoptotic factors PUMA and NOXA, potent proteins that initiate the apoptotic cascade [Kerr et al., 2012]. Postnatal expression of TAp63 remains high in the oocytes of primordial, primary, and early secondary follicles [Nakamuta and Kobayashi, 2004].

Interestingly, while apoptosis is the preferential mechanism for elimination of prophase-arrested oocytes after DNA damage, and as few as four DNA DSBs are sufficient to induce apoptosis in postnatal immature oocytes [Kim and Suh, 2014], studies in apoptosis-deficient (TAp63−/−) mice have shown that prophase-arrested oocytes are also capable of highly efficient DNA repair through homologous recombination [Stringer et al., 2020]. These findings implicate DNA repair as a critical quality control mechanism both in pre-arrested and prophase-arrested oocytes and suggest that apoptosis is not the sole quality control mechanism oocytes deployed in response to DNA damage. The implications for future therapeutic strategies designed to rescue aging oocytes from apoptosis are consequential: given that depletion of oocytes in the aging ovary results in infertility and menopause, therapeutic strategies to block oocyte loss can be expected to be successful only if rescued oocytes are viable and healthy.

Intact DDR pathways are associated with a robust ovarian reserve and prolonged fertility. For example, mouse oocytes engineered to overexpress BRCA1 have higher survival rates and fewer DSBs after exposure to genotoxins [Titus et al., 2013a]. Conversely, in rat follicles, BRCA1,RAD51,andH2AX messenger RNA (mRNA) transcripts decrease with aging, indicating decreased efficiency in recruiting the necessary machinery for DDR [Govindaraj et al., 2015]. Furthermore, γBRCA1 (the phosphorylated, active form) also decreases with aging [Govindaraj et al., 2015] and aged mice exhibit an increase in the number of γH2AX marks at DSBs [Titus et al., 2013a], indicative of increased oocyte DNA damage as compared to young mice.

Another critical aspect of ovarian aging and oocyte development is the regulation and maintenance of the meiotic spindle; indeed, successful reproduction relies on the well-orchestrated development of oocytes through meiosis. During meiotic prophase, homologous chromosomes must pair, synapse, and undergo meiotic recombination. Stable, tightly bound chromosomal centromeres are essential for the prevention of oocyte aneuploidy. Centromere tightness maintains proper tension at the chromosome to prevent premature or delayed separation by the meiotic spindle [Gorbsky, 2015] and spatial control is exerted by sister kinetochore pairing [Gorbsky, 2015]. Without a tight centromere, spindle attachments can occur randomly in germ cells and promote aneuploid events in daughter cells [Mihajlović and FitzHarris, 2018]. Indeed, greater interkinetochore distance has been observed in older mice (16–19 months of age) compared to younger mice (6–14 weeks of age) in both metaphases I and II, and this distance correlated with 90% of the aneuploidies seen in older mice [Chiang et al., 2010].

Spatial control of chromosomes is achieved in part via meiotic cohesion complexes. Cohesin proteins are important leading up to separation in anaphase because they act to establish the metaphase plate [Mihajlović and FitzHarris, 2018]. Cleavage should occur only when all chromosomes are properly aligned to ensure proper genetic distribution. Consequently, cohesion loss has been proposed as one mechanism contributing to oocyte aneuploidy in older women. Cohesin loading is primarily a prenatal event [Burkhardt et al., 2016], and age-related meiotic segregation errors in mammalian oocytes are preceded by cohesin depletion [Lister et al., 2010]. In mice, levels of the meiosis-specific cohesin protein REC8 decrease in aging oocytes [Chiang et al., 2010]. However, only levels of chromosome-associated REC8 decrease, rather than total levels of soluble REC8, suggesting that aging impairs the ability of REC8 to associate with chromatin but does not change overall protein levels [Chiang et al., 2010]. Further supporting a role for cohesion dysfunction in oocyte aging, multiple cohesion variants have been associated with familial POI [Beverley et al., 2021].

Loss of sister chromatid cohesion is prevented in part by a safeguard known as the spindle assembly checkpoint (SAC). One necessary protein of the SAC is Chromosome Transmission Fidelity Factor 18 (CHTF18), a component of the Replication Factor C-like complex, which aids in sister chromatid cohesion and loads the clamp subunit of the eukaryotic replisome [Holton et al., 2020]. Although mice with a global knockout of CHTF18 are still able to undergo recombination events, they exhibit significantly more DNA damage and greater outer kinetochore distances resulting in early homolog disjunction in oocytes, indicating a critical function for CHTF18 in facilitating correct chromosomal synapsis and preventing aneuploidy [Holton et al., 2020]. Activation of the SAC also requires Aurora kinase B (AURKB), which monitors kinetochore-microtubule interactions and recruits proteins to the SAC complex. Oocyte-specific depletion of AURKB reduces expression of SAC response pathway proteins, compromising SAC integrity and resulting in increased numbers of aneuploid oocytes and infertility [Blengini et al., 2021] and further implicating spindle dysfunction in ovarian aging.

Telomeres, or repeats of noncoding DNA located at the ends of chromosomes, function as physical “caps” to protect the genome from shortening during DNA replication. The process of replication of linear eukaryotic DNA results in a loss of genetic material after each round of new strand synthesis. Telomeres represent noncoding sequences that can be sacrificed in this process, protecting other functional parts of the genome from being deleted [Rocca et al., 2019]. In somatic cells, telomere lengths shorten through cell division until they are exhausted, triggering cellular senescence [Rocca et al., 2019]. Thus, there has been an increasing interest in the potential role of telomeres in human reproduction. In support of a link between telomeres and female reproductive aging, studies have shown reduced activity of telomerase (a reverse transcriptase that adds TTAGGG sequences to telomeres, thus preserving them from progressive loss with successive cell replications) in granulosa cells from women with occult ovarian insufficiency, as well as reduced leukocyte granulosa cell telomere length (TL) [Butts et al., 2009; Xu et al., 2017].

However, while shortened somatic cell TL correlates with epigenetic age acceleration, ovarian cumulus cells do not demonstrate changes in telomere length with increasing female age [Morin et al., 2018]. Furthermore, the TL of leukocytes does not correlate with the TL of cumulus cells or granulosa cells [Lara-Molina et al., 2020] suggesting that telomere measurement in leukocytes is not an accurate measurement of follicular TL and precluding the use of this measure as an estimate of gamete TL length. Indeed, cumulus cell telomeres are significantly longer than leukocyte telomeres [Lara-Molina et al., 2020] which may be a result of a unique “coping strategy” in follicles to protect against telomere shortening. Finally, embryo polar body and blastomere biopsies have shown no correlation between TL and female age, or between embryo aneuploid status and age [Turner et al., 2019]. As such, the ovarian germ cell microenvironment may employ other means of protecting and preserving telomeres than what is established in somatic tissues.

Mitochondria are double-membrane-bound intracellular organelles essential for anaerobic metabolism and energy production. Mitochondria produce ATP during energy metabolism and during this process constantly generate reactive oxygen species (ROS); these are eliminated by antioxidant enzymes in the mitochondria. While most cellular DNA is located in the nucleus, DNA is also found in mitochondria; unlike the large, linear structure of nuclear DNA, mitochondrial DNA (mtDNA) are small structures (spanning about 16,500 nucleotides) which are packaged in a double-stranded, closed, circular conformation. Human mtDNA is inherited solely from the mother, except in rare cases of inheritance of both maternal and paternal mtDNA [Parr and Martin, 2012]. mtDNA encodes 37 genes which are essential for mitochondrial function, including genes encoding tRNA, rRNA, and enzymes involved in oxidative phosphorylation and synthesis of ATP. Mitochondria contain cellular machinery to maintain, replicate, and transcribe their own mtDNA, as well as mechanisms to limit the effects of cellular damage on optimal mitochondrial function [Kasapoǧlu and Seli, 2020].

The “Free Radical Theory of Aging” postulates that oxidative stress contributes to cellular aging due to the accumulation of reactive oxidative species [Yang et al., 2021]. Published results have not been entirely concordant, but most studies report that mtDNA levels either remain unchanged or decrease with advancing age [Pasquariello et al., 2018]. Mitochondrial dysfunction contributes to excessive generation of ROS and reduced antioxidant ability. Increased susceptibility to reactive oxidative species in aging oocytes, in turn, disrupts mitochondrial function as well as oocyte metabolism and DNA repair [Sasaki et al., 2019], further limiting the ability of oocytes to recover from oxidative stress [Choi et al., 2007; Goud et al., 2008; Miyamoto et al., 2010; Mihalas et al., 2017; Tamura et al., 2020]. Mitochondrial dysfunction can also result from mitochondrial mutations or from aberrant DNA repair [Chiang et al., 2020]. mtDNA mutation rates are on the order of 100-fold higher than that of nuclear DNA, and individuals may harbor a mixture of wild type and mutant mtDNA (heteroplasmy). However, the burden of mutant mtDNA must reach a particular threshold for clinical manifestations of a mitochondrial disorder to occur (the thresholdeffect) [Parr and Martin, 2012]. Human studies as well as mouse models have linked mitochondrial mutations with a decrease in the quality and quantity of oocytes [Van Blerkom et al., 1995; Bentov et al., 2011; May-Panloup et al., 2016; Cimadomo et al., 2018; Kasapoǧlu and Seli, 2020]. Thus, mitochondrial replacement has been explored as a method for improving oocyte mitochondrial function in older women. However, in a study of autologous mitochondrial transfer as a treatment for ovarian aging in women with IVF failure, embryos created from mitochondrial transfer had poorer development and the study was discontinued [Labarta et al., 2019].

Antioxidants, which target damaging ROS found in the aging cellular environment, have been explored as a tool to combat aging. For example, melatonin, a free radical scavenger, is present in follicular fluid and oocytes [Nakamura et al., 2003; Cruz et al., 2014]. Melatonin treatment in vitro increases expression of hyaluronan synthase-2 (HSA2) and progesterone receptor (PGR) genes, which are critical for cumulus cell expansion, and improves fertilization rates in mouse cumulus cell-oocyte complexes treated with melatonin [Ezzati et al., 2018]. A randomized trial of melatonin supplementation in 40 women with unexplained infertility concluded that daily melatonin supplementation (3 mg or 6 mg) improved oxidative balance and oocyte quality [Espino et al., 2019]. Follicular fluid melatonin levels have been shown to correlate with progesterone levels in women undergoing in vitro fertilization [Nakamura et al., 2003]; however, treatment of granulosa cells with melatonin in vitro did not increase progesterone levels, suggesting that melatonin does not directly stimulate progesterone production in human granulosa cells. Consequently, more work is needed to determine the potential therapeutic role of antioxidants, including melatonin, in treatment of female reproductive aging.

Sirtuins are a family of nicotinamide adenine dinucleotide-dependent deacetylases which can catalyze either deacetylation or ADP ribosylation and are important modulators of cellular metabolic processes. Sirtuins can be present in the nucleus (SIRT1, SIRT6, and SIRT7), in mitochondria (SIRT3, SIRT4, and SIRT5), or in the cytoplasm (SIRT2), and can relocate during cell differentiation. Sirtuins are implicated in diseases of aging including cancers, metabolic disorders, and neurodegenerative diseases, and have emerged as candidate predictive markers and therapeutic targets for ovarian aging [Zhang et al., 2016]. Several sirtuins, including SIRT1, SIRT3, and SIRT6, are positively correlated with primordial follicle numbers in mice [Zhang et al., 2014, 2016]. SIRT1, one of the most well-studied enzymes of the sirtuin family, is involved in the regulation of DNA repair and apoptosis, mitochondrial production, and the cell stress response [Chandrasekaran et al., 2017]. Caloric restriction, which preserves ovarian follicle numbers in rodent models, results in elevated ovarian SIRT1 and SIRT6 ovarian protein levels [Luo et al., 2012; Long et al., 2019] while SIRT1 knock-in mimics caloric restriction in the ovary and preserves ovarian reserve. Treatment of mice with a SIRT1 activator also results in a significantly increased number of primordial follicles and suppresses mTOR signaling; mTOR inhibition has been shown to prolong reproductive longevity in a mouse model of physiologic ovarian aging [Goldman et al., 2017]. Conversely, depletion of SIRT1 in oocyte-specific SIRT1-knockout mice results in premature sterility due to defective oocyte development and quality [Iljas et al., 2020]. These data suggest that loss of sirtuin expression may result in activation of mTOR pathways and depletion of the ovarian reserve.

As another potential mechanistic explanation for these findings, loss of SIRT1 has been shown to reduce expression of NRF2, a protein that regulates antioxidant proteins and modulates oxidative stress [Ma et al., 2018]. Disrupted NRF2 expression through NRF2 mRNA and protein knockdown in mouse oocytes result in aberrant oocyte maturation and chromosome and spindle organization, suggesting relevance for a NRF2-SIRT1 signaling pathway in oocyte aging [Ma et al., 2018]. Loss of SIRT2 and SIRT3 in mouse oocytes also impairs spindle organization and chromosome alignment in mice, while SIRT3 overexpression in mice oocytes reduces spindle defects and chromosome misalignment [Zhang et al., 2014], demonstrating a role for sirtuins in oocyte meiotic spindle assembly and chromosome segregation during meiosis.

Inflammation, another hallmark of aging, has been linked to female reproductive aging. Aged mouse ovaries demonstrated an enhanced inflammatory environment, with increased expression of TNF-α, IL6,IL1a, IL1b, and IL10 mRNA [Lliberos et al., 2021] and presence of multinucleated giant cells [Briley et al., 2016]. Corresponding to these results, female TNF-α knockout mice (TNF-α−/−) have increased frequency of estrous cycles, increased granulosa cell proliferation and reduced oocyte apoptosis, and produce 21% more pups than control mice during a 12-month breeding period [Cui et al., 2011]. Similarly, pharmacologic inhibition of the neutrophil-to-lymphocyte ratio (NLR) family pyrin domain containing 3 (NLRP3) inflammasome, a multiprotein complex which regulates activation of caspase-1 and inflammatory signals in response to a range of stressors, attenuates ovarian aging and prolongs fertility in female C57/Bl6 mice [Navarro-Pando et al., 2021]. In humans, a prospective study found that early menopause risk was lower in women with moderate tumor necrosis factor receptor 2 (TNFR2, an antagonist of TNF-α activity) levels compared to women with low TNFR2 levels [Bertone-Johnson et al., 2019]. Women with POI also exhibit an increased serum inflammatory environment, as exhibited by a lower NLR ratio than normal fertile women; in multivariate logistic regression analysis, a NLR ≤1.5 was an independent risk factor for POI [Ilhan et al., 2016]. It has been postulated that age-related inflammation can impair meiosis and oocyte quality [Snider and Wood, 2019]. Inflammation also plays a fundamental role in the development of tissue fibrosis, and fibrotic changes have been observed in the aging mouse ovary [Briley et al., 2016]. However, whether suppression of inflammatory processes can delay human ovarian aging remains to be determined.

In the last half-century, the central dogma of molecular biology – that genes are transcribed into mRNA, and mRNA is translated into protein – has been revised. It’s now known that mRNA makes up only about 1.5% of the genome. It has become abundantly clear that noncoding RNAs (ncRNAs), RNA transcripts with no protein product, are integral to the function of cells, particularly in the control of gene activity and translational regulation. ncRNAs can be further divided into subtypes of ncRNAs based on their transcript size and include small ncRNAs (less than 200 bases) and long ncRNAs (lncRNAs, more than 200 bases). ncRNAs have multiple and diverse cellular functions, with one critical function being the silencing of target mRNA translation [Brosnan and Voinnet, 2009; Kallen et al., 2013; Cech and Steitz, 2014; Grimson, 2015]. It is therefore unsurprising that ncRNAs are intriguing candidate regulators of female reproductive aging. Ovaries from young and old mice [Schneider et al., 2017], as well as oocytes from aged women [Barragán et al., 2017] display global, age-related changes in ncRNA expression, suggesting that reproductive aging is associated with significant alterations in ncRNA pathways [Barragán et al., 2017; Schneider et al., 2017]. In studies using Ames dwarf mice, which have extended reproductive longevity as compared to normal mice, differential ovarian micro RNA (miRNA) profiles were observed in young and aged mice as well as in ovaries from dwarf mice as compared to normal mice [Schneider et al., 2017]. This divergent miRNA profile suggests that the aging process has a differential impact on the ovarian miRNA profile and implicates miRNAs can be central players in the maintenance of a younger ovarian phenotype.

Also highlighting the scope of reproductive processes that are simultaneously altered by lncRNAs and that affect fertility, in vitro activation of neonatal mouse primordial follicles is associated with widespread changes in lncRNA expression, with 6,541 lncRNAs upregulated and 2,135 lncRNAs downregulated, suggesting a role for lncRNA involvement in PFA [Zheng et al., 2019]. The lncRNA H19 represents one candidate regulator of ovarian aging [Kallen et al., 2013]. H19 was discovered as the first lncRNA over 3 decades ago and is a highly conserved imprinted gene, suggesting a physiologic pressure to maintain function. H19 is abundantly expressed in the early stages of embryogenesis; in the adult, expression is observed in the skeletal muscle, heart, and reproductive organs including the ovary and uterus. Mice witha targeted deletion of H19, characterized by deletion of the 3-kb transcription unit upstream of the H19 gene itself, exhibit accelerated primordial follicle loss in part due to decreased expression of AMH, a downstream target of H19 [Men et al., 2017; Qin et al., 2018; Chen et al., 2019]. Corresponding to these results, women with diminished ovarian reserve (indicative of ovarian aging), have decreased circulating and ovarian H19 levels [Xia et al., 2020]. Antagonist and agonists which up- or downregulate ncRNAs are already in use in other fields; as such, harnessing this technology to target candidate ncRNAs such as H19 may be key to the development of therapeutic interventions aimed at slowing ovarian aging.

We have reviewed several pathways that must combine in their actions to regulate the rate of PFA and oocyte quality. While there may be pathways unique to each cell type, there is also significant overlap and integration of signals to control overall ovarian aging. Stated another way, internal communications between the cell types of the ovarian follicle can be regulated in response to external signal in order to control the rate of PFA and the quality of oocytes. A recent report of how the many signals that regulate PFA investigated the aptly-named Integrated Stress Response (ISR) pathway [Pakos-Zebrucka et al., 2016]. In the ISR, different signals and stressors are received by a set of four kinases, which converge upon common machinery to regulate protein translation, and ultimately, the cell cycle. Intriguingly, the ISR regulatory machinery is expressed in granulosa cells and oocytes [Llerena Cari et al., 2021] and the different stress and damage signals that activate the ISR are present in the follicle itself, suggesting that the ISR and pathways like it might serve to integrate complex signaling events in order to keep the cell cycle of granulosa cells in primordial follicles arrested or allow them to begin to grow. We now move on to highlight areas in the field where clinical progress is being made to target accelerated ovarian aging.

Currently, the most successful examples of intervening in cases when patients are at risk for early ovarian failure involves fertility preservation (FP) strategies such as the collection of oocytes, embryos, or ovarian cortical tissue. Cryopreservation of oocytes and/or embryos has become an established method of fertility preservation in women at risk for accelerated ovarian aging such as those receiving gonadotoxic therapy for cancer treatment. Women with a male partner or who wish to use donor sperm may choose to pursue controlled ovarian hyperstimulation (COH) with in vitro fertilization (IVF) and subsequent embryo cryopreservation [Carson and Kallen, 2021]. For those who cannot or wish not to cryopreserve embryos, methods for ultrarapid oocyte freezing (vitrification) now allow for banking of mature oocytes after COH, an excellent fertility preservation strategy for eligible patients. Ovarian tissue cryopreservation (OTC), in which a portion or the entire ovary is removed and frozen for future use, is also an option for women who will receive gonadotoxic therapies, especially for prepubertal individuals or those who otherwise cannot undergo COH [Dolmans et al., 2013; Anderson et al., 2015; Donnez and Dolmans, 2015; Oktay et al., 2021]. A major challenge, however, is the identification of “at risk” individuals who would benefit from FP. While some of these individuals are easily identified, such as those undergoing planned gonadotoxic chemotherapy or oophorectomy, others (such as women with impending POI) may not be readily apparent; for these women, a reliable noninvasive diagnostic test for early prediction of ovarian insufficiency could be transformative.

At present, there are no effective treatment modalities for women with POI who wish to have children with their own eggs. Efforts to delay ovarian aging include the use of stem cell therapies or platelet-rich plasma (PRP) to support the viability of existing oocytes and follicles. It has been hypothesized that the administration of stem cells (e.g., via bone marrow transplant) can replenish intraovarian factors necessary for follicular growth [Sheikhansari et al., 2018; Yoon, 2019]. A study of the effects of bone marrow derived stem cells (BMDSC) on fertility in mice with chemotherapy-induced POI demonstrated higher ovarian weight, serum estradiol levels, and ovarian vascularization, as well as more ovulatory follicles, metaphase II oocytes, two-cell embryos, and live births [Herraiz et al., 2018a]. Another recent study of endometrial mesenchymal stem cells (MSC) injected into mice with chemotherapy-induced POI reported higher serum AMH levels, and ovulation rates, more developing follicles, and higher live birth rates [Liu et al., 2021]. A trial of BMDSC use in 15 women with poor ovarian response to ovarian stimulation reported improved ovarian function in 81% of patients, and five live births were reported, but no statistically significant differences were observed when total antral follicular count (AFC) results were analyzed [Herraiz et al., 2018b]. Larger clinical trials are needed to confirm these findings.

PRP is a blood-derived product which is growth factor and chemokine rich, including factors such as platelet derived growth factor (PDGF), epidermal growth factor (EGF), insulin-like growth factor (IGF), transforming growth factor beta-1 (TGFβ-1), vascular endothelial growth factor (VEGF), hepatocyte growth factor (HGF), and basic fibroblast growth factor (bFGF) [Bos-Mikich et al., 2018; Scott Sills and Wood, 2019]. PRP is used for therapeutic purposes in various tissue types (e.g., tendons, muscles, nerves) with the goal of improved neoangiogenesis, wound healing, and regeneration. Consequently, intraovarian injection of PRP has been proposed as a strategy to activate dormant ovarian follicles before in vitro fertilization cycles in patients with diminished ovarian reserve or POI. While studies have reported increased angiogenesis and antral follicle growth after culture with PRP [Shahidi et al., 2018; Hsu et al., 2020], and live births have been reported after PRP injection [Cakiroglu et al., 2020; Hsu et al., 2020], to date there is no randomized placebo-controlled trial available that has evaluated intraovarian PRP injection in terms of efficacy and safety for POI or diminished ovarian reserve.

A crucial conundrum in ovarian biology is why some primordial follicles are maintained in dormancy for many years, while others are activated for growth. What is the mechanism by which adjacent follicles can be earmarked for such different fates? Whether an individual follicle stays in the resting pool or initiates growth likely rests on the balance of pro-aging and antiaging processes at a particular point in time. At the very least, ovarian aging is recognized as what may be a complex process in which age, genetics, environment, and many other factors contribute to the size and depletion of the follicle pool. The putative hallmarks of reproductive aging outlined above include a diversity of plausible processes contributing to the depletion of the ovarian reserve. More research is needed to clarify if and to what extent these putative regulators do in fact govern follicle and oocyte behavior, and how these signals might be integrated in order to control the overall pattern of ovarian aging. When these regulatory mechanisms are better understood, we may be able to intervene and slow the loss of primordial follicles in women at risk for POI and to noninvasively improve the quality of eggs in women that currently produce only eggs incapable of giving rise to healthy children.

The authors have no conflicts to declare.

Dr. Kallen gratefully acknowledges funding and research support provided by the NIH-NICHD (R01HD101475), the Global Consortium for Reproductive Longevity and Equality Junior Scholar Award, the Reproductive Scientist Development Program (NIH-NICHD Project #2K12HD000849-26), the American Society for Reproductive Medicine, and the NIH Loan Repayment Program.

Amanda N. Kallen, Jesus Lopez, and Gabe Hohensee outlined the manuscript and performed relevant literature searches. Amanda N. Kallen, Jesus Lopez, and Gabe Hohensee took the lead in writing the manuscript. Jing Liang and Meriav Sela provided critical feedback and revisions for important intellectual content. All authors reviewed and approved the final draft of the manuscript.