Studying ovarian aging and its health impacts: modern tools and approaches

In this review, Benayoun et al. discuss the importance and systemic consequences of ovarian aging and their implications for female aging, health, and longevity. They delineate existing model systems and methods that can be used to study ovarian aging, highlighting the potential of modern approaches to further advance our understanding of the complexities of human aging and women's health.

Abstract

Ovarian aging is a critical yet understudied driver of systemic aging in female bodies, with profound implications for female health and longevity. Despite its significance, we still know little about ovarian aging and its systemic effects on aging trajectories. With new efforts over the past few years, interest in the field has been growing and there is momentum to address these questions. This review highlights the importance of leveraging modern tools and approaches to better understand ovarian aging and its impact on health span. Specifically, we believe it will be useful for both aging researchers looking to go into research on ovarian aging and reproductive researchers looking to adopt more modern toolkit. We focus on menopause—a key marker of ovarian aging—as a lens through which to examine the current state of the field, identify limitations in existing research, and outline goals for future progress. By emphasizing cutting-edge techniques and emerging models, we seek to illuminate new pathways for research that could lead to improved strategies for managing ovarian aging and enhancing overall female health.

The cohort of human supercentenarians (i.e., individuals aged >110 years) reveals a surprising predictor for achieving such an exceptional longevity: being female (https://www.grg-supercentenarians.org/world-supercentenarian-rankings-list). Despite overall life expectancy increases for both women and men over the last century, human longevity remains highly sex-dimorphic, with women's life expectancy systematically and robustly exceeding men's (Austad and Fischer 2016; Yan et al. 2024). However, women also tend to age in poorer health than men, a phenomenon dubbed the mortality–morbidity paradox, whereby women generally live longer but experience greater frailty than age-matched men (Kulminski et al. 2008). Late-life changes in ovarian endocrine function may largely drive this paradox.

In this review, we discuss ovarian aging in humans, its systemic consequences, the state of the field, and how to rigorously study its health impacts in preclinical models. Here, we designate individuals with ovaries as female, using the biological (i.e., sex) rather than the societal (i.e., gender) definition of this word.

Key terminology and concepts

Ovarian aging is a critical yet understudied driver of organismal female aging with profound implications for female health and longevity, with special relevance to humans due to their long postreproductive survival (Shadyab et al. 2017; Fu et al. 2022).

To clearly explain the impact of ovarian aging and how best to study it, we first establish the definition and relevance of key concepts as follows:

“Reproductive aging” is an imprecise term often used to describe age-related changes affecting fertility. Because of its vagueness, this term could apply to male or female individuals and may also encompass aging of the entire female reproductive tract in addition to the ovary. Thus, in the context of research on female health span (irrespective of fertility), the notion of “ovarian aging” provides more precision.

“Ovarian aging” more accurately describes how the female gonad ages and its potential systemic impacts. Ovaries have two (partially) separable functions throughout female life span: (1) gamete production (i.e., oocytes) and fertility and (2) endocrine production of hormones and overall health promotion. Disruption of normal ovarian function at any age profoundly impacts long-term health (Hart and Doherty 2015; Shadyab et al. 2017; Fu et al. 2022; Baker and Benayoun 2023). Ovaries age at almost twice the rate of other tissues in female bodies (InterLACE Study Team 2019; Sirard 2022), which significantly impacts female health span independent of fertility.

“Menopause” is clinically defined as the time occurring 12 months after a woman's last menstrual cycle. Women typically experience menopause around age 51 years, but it can occur as early as in a female's 20s for a small percentage of the population (∼0.1%; known as premature ovarian insufficiency [POI] or premature ovarian failure [POF], defined as menopause before age 40 years) (Davis and Baber 2022). The menopause transition brings numerous systemic symptoms (e.g., hot flashes, heart palpitations, anxiety, etc.) (Santoro et al. 2015). Menopause, or the cessation of menstruation, is only one downstream consequence of ovarian aging, which uncovers risks for various health issues at any age.

Importance of studying ovarian aging

Menopause is accompanied by broadly increased health risks

Although menopause has historically been reduced to loss of fertility (its most obvious consequence), menopause also coincides with a broad acceleration of age-related functional decline (Baker and Benayoun 2023). Importantly, menopause accelerates so-called “biological” aging in women (Levine et al. 2016). Later age at menopause strongly predicts longevity (Ossewaarde et al. 2005; Hong et al. 2007; Shadyab et al. 2017) and lower incidence of many age-related comorbidities (Muka et al. 2016, 2017). Conversely, postmenopausal women face higher risk for almost every chronic age-related disease (e.g., osteoporosis, neurodegeneration, and cardiovascular diseases) (Gubbels Bupp 2015). Vasomotor symptoms (e.g., hot flashes and night sweats) commonly accompany the menopausal transition in humans associated with increased risk of diabetes and high blood pressure (Szmuilowicz et al. 2011; Gray et al. 2018). Thus, ovarian aging could be one of the underlying drivers of the mortality–morbidity paradox (Fig. 1).

Figure 1.

Overview of the lifelong impact of ovarian function on various health systems.

Menopausal hormonal therapy: a missed opportunity for women's health

The Women's Health Initiative trial was designed to determine the efficacy and safety of menopausal hormone therapy (MHT) to treat age-related phenotypes in postmenopausal women (Rossouw et al. 2002). The original trial was terminated prematurely by the Data Safety Monitoring Board due to nonsignificant trends toward increased risks of gynecological cancers in the MHT group, likely driven by flawed design (i.e., lack of power to stratify data analysis based on time since menopause) (Rossouw et al. 2002; Baker and Benayoun 2023). Thus, fallout from this initial study (Rossouw et al. 2002; https://www.whi.org) created resistance from the public and poorly informed medical personnel to broadly adopt MHT as standard of care for postmenopausal women (Lobo 2013; Gurney et al. 2014; Bluming et al. 2023; Stute et al. 2023). However, more recent analyses have shifted the narrative, revealing MHT's net beneficial effect in menopausal women, with increased preservation of cognition and bone density (Gambacciani and Levancini 2014; Gurney et al. 2014). This is especially the case when MHT is initiated early in the menopausal transition and in the absence of specific risk factors (Lobo 2013). Moreover, MHT shows promising impact on cardiovascular disease risk as a prophylactic treatment, with significant reduction in cardiovascular events, if started within 10 years of age at natural menopause (ANM) (Schierbeck et al. 2012). This effect is especially important for aging women because commonly used prophylaxis strategies against cardiovascular disease (i.e., aspirin and statins), which work well in men, show minimal efficacy in women (Ridker et al. 2005; Petretta et al. 2010). Thus, failure to broadly adopt MHT as a health-promoting strategy for postmenopausal women is a missed opportunity for women's health.

Factors regulating menopause onset in humans

Oocyte quality and quantity are important factors in setting the onset of menopause, with a drastic decline in women's mid-30s (Wallace and Kelsey 2010). However, no causal relationship has been established, and it is not clear whether the decline in egg number and quality is a cause or a consequence of ovarian aging. This decline has been discussed extensively elsewhere (Moghadam et al. 2022; Park et al. 2022; Charalambous et al. 2023). Although it is important in the context of fertility, the dynamics of oocyte loss does not provide insights into the systemic impacts of menopause or the therapeutic strategies to regulate its onset. Thus, we focus on known genetic and lifestyle factors that are associated with menopause onset.

Genetic regulation of age at natural menopause

Multiple studies support the genetic basis of ANM, including premature menopause, with heritability estimates ranging from 44% to 85% (te Velde and Pearson 2002; Murray et al. 2011). Genome-wide association studies (GWASs) and Mendelian randomization (MR) analyses have highlighted genetic loci associated with ANM in humans (He et al. 2009; Stolk et al. 2009; Day et al. 2015; Ruth et al. 2021; Yazdanpanah et al. 2024). For instance, a GWAS meta-analysis of 70,000 women revealed 54 significant genetic loci that, combined, explain ∼6% of ANM variance (Day et al. 2015). Interestingly, genes linked to these loci were significantly enriched for DNA damage response (e.g., BRCA1) and included key players of the hypothalamic–pituitary–gonadal (HPG) axis (e.g., FSHB) (Day et al. 2015). A recent study identified 209 significant loci associated with ANM in a cohort of >200,000 European women and was validated across diverse ethnicities (Ruth et al. 2021), reinforcing a link between genes involved in DNA damage response and ANM. A recent analysis of rare protein coding variants in >100,000 women from the UK Biobank study revealed new potential genes with a large predicted effect on the rate of ovarian aging (i.e., ETAA1, ZNF518A, PNPLA8, PALB2, and SAMHD1) (Stankovic et al. 2024).

Population-level genetic association studies for POI have been less successful, likely due to a lack of statistical power and/or the impact of rare mutations (Fortuño and Labarta 2014; Louwers and Visser 2021). However, classical genetics approaches have identified many genes whose mutations lead to POI under various presentations (syndromic or nonsyndromic) and severities (ranging from primary amenorrhea to early menopause). Intriguingly, many X-linked genes are associated with other forms of syndromic POI, including forms of Turner's syndrome (Fortuño and Labarta 2014). Mutations in several autosomal genes involved in the HPG axis (e.g., FSHR), steroid hormone biosynthesis (e.g., STAR), or DNA damage response (e.g., ATM) also cause POI (Fortuño and Labarta 2014). For example, mutations in FOXL2 (the master regulator of ovarian fate) (Uhlenhaut et al. 2009; Boulanger et al. 2014) were first identified as causal for a syndromic form of POI, type I blepharophimosis ptosis epicanthus–inversus syndrome, using positional cloning (Crisponi et al. 2001), though hypomorphic FOXL2 variants have also been identified in nonsyndromic POI patients (Laissue et al. 2009). Candidate gene approaches for ANM regulators have focused on genes involved in the HPG axis, sex hormone signaling, and genes associated with POI. For instance, variants in FSHB and ESR1 are significantly associated with ANM (Louwers and Visser 2021). Together, genetics studies show that ANM is regulated by a complex genetic architecture involving both common and rare genetic variants, with a strong involvement of genes related to DNA repair, mitochondrial function, immune response, and HPG axis regulation.

Lifestyle, diet, and environmental factors

Beyond genetics, lifestyle, reproductive, and environmental factors also influence ANM. Specifically, smoking, a high-fat Westernized diet, and intense physical activity seem to lead to earlier ANM (Ceylan and Özerdoğan 2015; Appiah et al. 2021). In contrast, moderate alcohol consumption, light physical activity, and higher BMI are associated with later ANM (Ceylan and Özerdoğan 2015; Appiah et al. 2021). Intriguingly, several reproductive-linked factors have been found to be significantly associated with ANM. For example, nulliparity is linked to earlier ANM, whereas multiparity is associated with later ANM (Ceylan and Özerdoğan 2015; Scime et al. 2024). In addition, women using oral contraceptives tend to experience later ANM and a longer reproductive span (Appiah et al. 2021). Socioeconomic and demographic factors can also influence ANM, with Black/Hispanic women experiencing earlier menopause, while women with higher years of education experience later menopause (Appiah et al. 2021).

Menopause onset can also be accelerated by necessary medical interventions, specifically in the context of chemotherapy and/or use of selective estrogen receptor modulators (SERMs) for cancer treatment (Goodwin et al. 1999).

The modern toolkit to study ovarian aging and its systemic impacts

In this section, we discuss the current toolkit to study ovarian aging and its impact on health span, as well as caveats and limitations. Our focus is specifically on applying modern tools to answer these important questions.

Established and emerging models for studying ovarian aging

Across taxa, very few species undergo natural menopause (namely, humans and a few species of whales) (Johnstone and Cant 2019), which has severely limited our ability to model and understand this phenomenon in preclinical settings. In addition, most studies of ovarian function in existing preclinical models thus far approach the question through the restrictive lens of fertility. Together, this has created a significant gap in our understanding of the ovary's role in overall health and the identification of relevant therapeutic targets. In this section, we discuss current models and their advantages and drawbacks (Table 1).

Table 1.

Animal and human models to study ovarian aging

Species/clade	Model	Advantages	Drawbacks	Reference
Mus musculus	Physiological aging	No intervention required Age-related decline in ovarian function	Estropause occurs later in life Persistent low E2 is never truly achieved Estrous patterns differ from human menstrual cycle	Van Kempen et al. 2011
	Ovariectomy (OVX)	Easy to determine time of “ovarian failure”	No hormonal transition period Missing postreproductive ovarian tissue Potential impact of surgical shock/prophylactic antibiotic treatment Generally has been performed in barely sexually mature animals, eliminating the impact of the aged milieu	Laszczynska et al. 2008; Van Kempen et al. 2011
	VCD injections	Easy to determine time of “ovarian failure” Existence of transitional hormonal states Achievement of low persistent E2 in the posttreatment state	Generally has been performed in barely sexually mature animals, eliminating the impact of the aged milieu Estrous patterns differ from human menstrual cycle	Van Kempen et al. 2011
	Fshr +/−	Mutations of the gene are involved in human POI Clear mechanistic explanation for a core gene of the HPG signaling access Some features of menopause were reported to be conserved in this model	Model was not reproducible across studies using different alleles/genetic backgrounds	Danilovich et al. 2000, 2002; Danilovich and Sairam 2002; Sairam et al. 2006; Mehalko et al. 2023
	Foxl2 +/−	Mutations of the gene are involved in human POI Clear mechanistic explanation through core function in ovarian granulosa cells and pituitary gonadotroph cells Animals are fertile and develop ovarian dysfunction in an age-dependent manner	Characterization of haploinsufficient model is incomplete Potential off-target effects outside of the HPG axis are unclear	Zlotogora et al. 1983; Crisponi et al. 2001; Harris et al. 2002; Gersak et al. 2004; Schmidt et al. 2004; Uda et al. 2004; Laissue et al. 2009; Uhlenhaut et al. 2009
	Cd38 KO/78c treatment (CD38 inhibitor)	Clear mechanistic understanding through conserved age-related pathways	Delayed ovarian failure extends preclinical studies Systemic effects are expected through nonovarian CD38 expression (not downstream from sustained ovarian function)	Perrone et al. 2023; Yang et al. 2024
	Ifng ARE−/−	Models immune drivers of ovarian aging	Estrous patterns differ from human menstrual cycle Systemic elevated inflammation can interfere with determination of causal impact of ovarian dysfunction	D'Souza et al. 2021; Bafor et al. 2024
Nonhuman primates	Physiological aging	Closely phylogenetically related to humans Similar age-related neuroendocrine profiles	Oocyte loss occurs at a slower rate than humans Long life span makes longitudinal studies difficult and expensive	Bellino and Wise 2003; Downs and Urbanski 2006; Alberts et al. 2013; Chaffin and Vandevoort 2013; Cloutier et al. 2015; de Jesus Rovirosa Hernandez et al. 2017; Colman 2018
Whales	Physiological aging	Rare species with natural menopause	Not tractable for preclinical studies and experimental interventions Biology not as well known Extremely long life span makes longitudinal studies difficult	Ellis et al. 2018, 2024
Acomys cahirinus	Physiological aging	True menstruation Human-like endocrine profiles Longer gestational periods	Distinct pattern of follicular depletion Estrous cycling-like pattern, distinct from the human one	Bellofiore et al. 2017, 2021
Homo sapiens	Aromatase inhibitor (AI) therapy	Human context leads to best applicability to human biology Rapid onset of menopausal symptoms and accelerated ovarian aging	Sudden disruption of hormone signaling different from gradual menopausal transition AI is given to breast cancer patients, precluding clean association of phenotypes regardless of cancer and associated treatment plan (e.g., chemotherapy and radiotherapy, which can mimic accelerated aging phenotypes on their own)	Hong et al. 2017
	POI	Human context leads to best applicability to human biology POI allows tracking of other systemic premature aging phenotypes in a causal manner	Genetic variants may have pleitropic effects outside of the ovary, potentially confounding conclusions	Fortuño and Labarta 2014; Louwers and Visser 2021
	Ovarian organoids	Human cells and human biology Defined, controlled environment limits interference from unknown variables	Aging process difficult to recapitulate in organoids Ex vivo aspect precludes influence of systemic factors (e.g., cytokines and immune cells) Lack the complex endocrine signaling networks (e.g., HPG axis)	Del Valle and Chuva de Sousa Lopes 2023; Dipali et al. 2024; Nason-Tomaszewski et al. 2024; Dadashzadeh et al. 2025
	Microfluidic systems	Human cells and human biology “Organ on a chip” modeling of ovarian function and aging in vitro Ability to maintain ovarian tissue alive for an extended time	Ex vivo aspect precludes influence of systemic factors (e.g., cytokines and immune cells) Lack the complex endocrine signaling networks (e.g., HPG axis), including cyclicity/rhythmicity	Xiao et al. 2017; Campo et al. 2023

Modeling disruption of ovarian function in rodents

Hallmarks of ovarian aging in rodent models have been comprehensively reviewed elsewhere (Balough et al. 2024). Below, we discuss a few widely used and emerging new rodent models to study ovarian aging and its systemic impacts.

Physiological aging

Prior to menopause, human women experience regular hormonal fluctuations, including alternating phases of increased estrogen versus progesterone production that regulate the menstrual cycle. Similar cyclic fluctuations are observed across mammals, termed “estrous cycle” in rodents (Hong and Choi 2018). The simplest rodent model to understand how ovarian aging may impact overall aging in humans is “physiological aging,” where female mice age without external intervention, experiencing the endogenous effects of ovarian aging. Although the physiological aging paradigm shows a transitional period with fluctuations in the levels of circulating ovarian hormones, it lacks a true “menopause”-like state because mice never achieve low to undetectable estrogen levels. Instead, mice enter an “estropause” state with low but persistent estrogen levels (Van Kempen et al. 2011).

Surgical models

One of the most commonly used models of “menopause” in mice is surgical removal of ovaries (ovariectomy [OVX]) in a unilateral or bilateral fashion. The rationale for this use is that surgical removal of ovaries eliminates hormonal cycling and makes exposure to systemic estrogens virtually inexistent (Van Kempen et al. 2011). However, OVX fails to recapitulate two crucial aspects of human menopause: (1) a transitional period of erratic hormonal fluctuations and (2) the presence of postreproductive ovarian tissue (Van Kempen et al. 2011). The OVX model thus fails to address the fact that, although it no longer produces substantial amounts of estrogens, the postmenopausal ovary still produces androgens such as testosterone and androstenedione (Laszczynska et al. 2008), which exert systemic effects and can be converted to estrogens. Thus, although OVX may be useful in studies that investigate whether a phenotype depends on ovarian inputs, it provides a poor model of menopause and should never be used as such.

Chemical models

Repeated injections with 4-vinylcyclohexene diepoxide (VCD) can promote selective atresia of primordial and primary ovarian follicles (Van Kempen et al. 2011). Within ∼100 days of a 15 day injection regimen, mice undergo accelerated ovarian failure with hormonal states reminiscent of perimenopause and postmenopause, including estrous acyclicity and fluctuating and then undetectable systemic estrogen levels (Van Kempen et al. 2011). This state closely mimics that of ovary-intact postmenopausal women. However, thus far, VCD treatment generally has been performed in very young animals (e.g., 2–3 month old mice and recently fertile juveniles). This creates an important modeling shortcoming, as menopause in humans occurs in an aged organism, not barely after puberty. Ample evidence shows that aging leads to numerous differences in the systemic milieu that can affect stem cell renewal, tissue regeneration, and brain function (e.g., heterochronic parabiosis) (Conboy and Rando 2012; Bieri et al. 2023). Thus, implementation in middle-aged animals, matching and modeling the natural timeline of human ovarian aging, is likely to provide additional insights. Accordingly, a recent study showed that initiation of VCD injections at up to 10 months of age (early middle age) does lead to premature ovarian failure, with unique additional features that may help model age at menopause in a preclinical setting (Kim et al. 2025). Notably, potential off-target effects of chemical exposure still represent a potential shortcoming of chemical exposure-driven models.

Genetic models

Several genetic models have been proposed to model human menopause in mice, inspired by knowledge of ovarian function control (e.g., Fshr knockout), genes associated with POI in humans (e.g., Foxl2 knockout), and organismal aging-related pathways expressed in ovaries (e.g., Cd38 knockout and an Ifng ARE^−/− model). Fshr haploinsufficiency was described as a promising genetic mouse model of menopause >20 years ago (Danilovich and Sairam 2002). In these studies, Fshr^+/− mice underwent progressive fertility decline, estrous cycle irregularity, and increased follicular atresia, culminating in reproductive arrest at age 7–9 months (Danilovich et al. 2000, 2002; Danilovich and Sairam 2002). The arrest was accompanied by high levels of FSH, LH, and androgens (Danilovich et al. 2000, 2002), low to undetectable estradiol levels (Danilovich et al. 2000), and metabolic defects at middle age (Danilovich and Sairam 2002; Sairam et al. 2006). Although these were promising features, these findings of menopause-like transition phenotypes could not be replicated in mice with an independent Fshr^+/− knockout allele (Mehalko et al. 2023).

As discussed above, germline mutations of FOXL2 are associated with both syndromic (Zlotogora et al. 1983; Crisponi et al. 2001) and isolated (Harris et al. 2002; Gersak et al. 2004; Laissue et al. 2009) POI in humans. Consistently, Foxl2 full knockout mice show abnormal ovarian physiology and granulosa cell biology (Schmidt et al. 2004; Uda et al. 2004; Uhlenhaut et al. 2009). Although in-depth phenotyping previously focused on full knockout Foxl2^−/− animals, haploinsufficient Foxl2^+/− females were noted to be subfertile (Schmidt et al. 2004; Uda et al. 2004) and could represent a powerful genetic model of menopause in mice. Indeed, a recent study found that Foxl2 haploinsufficient mice present with a number of phenotypes consistent with premature ovarian aging (i.e., increased reproductive latency, early immune infiltration of ovarian tissue, and a prematurely aged transcriptomic signature across cell types) (Kim et al. 2025).

Leveraging knowledge about pathways involved in systemic aging has also provided important information about genetic drivers of ovarian aging and dysfunction. For example, expression of CD38, an NADH scavenging enzyme, increases with aging in the ovary, concomitant with decreased NAD⁺ levels, which leads to increased levels of inflammatory gene expression in the ovary across cell types (Yang et al. 2024). Conversely, genetic ablation and chemical inhibition of CD38 counteracted ovarian aging phenotypes (Perrone et al. 2023; Yang et al. 2024), suggesting that CD38 modulation can be used to understand molecular drivers of ovarian aging.

Finally, genetic deletion of a regulatory element in the 3′ UTR of Ifng ARE^−/− leads to stabilization of Ifng mRNA, impaired IFNg regulation, elevated blood IFNg levels, and overall systemic inflammation, reminiscent of normal aging (D'Souza et al. 2021). Intriguingly, ARE^−/− mice show features of accelerated ovarian aging, including decreased fertility, immune infiltration in ovarian tissue, and pituitary–gonadal axis dysfunction (Bafor et al. 2024). Thus, due to the direct impact of immune infiltration on ovarian functions displayed by the model, ARE^−/− mice could be used to model the immune drivers of ovarian aging.

Overview of other mammalian models

A major challenge in studying menopause is the lack of models that fully recapitulate human ovarian aging. Even our closest evolutionary relatives show striking differences in ovarian aging patterns, necessitating careful consideration of each model's strengths and limitations. Beyond rodents, several mammalian models provide valuable insights into ovarian aging and menopause, each with distinct advantages and limitations for studying reproductive longevity. Invertebrate models fall outside the scope of this discussion and are reviewed in depth elsewhere (Tatar 2010; Scharf et al. 2021; Athar and Templeman 2022).

Nonhuman primates

Due to their evolutionary proximity, nonhuman primates (NHPs) provide the closest phylogenetic model of human reproductive aging biology (Colman 2018). Changes in the neuroendocrine profiles of aging macaques share key features with the human menopause transition, making them valuable for mechanistic studies (Bellino and Wise 2003; Downs and Urbanski 2006; Chaffin and Vandevoort 2013).

However, critical differences exist even between humans and our closest relatives. A detailed comparison of follicular loss patterns revealed that, although humans and chimpanzees show similar depletion rates until age 35, humans experience a markedly more precipitous decline thereafter (Cloutier et al. 2015). This accelerated loss pattern appears unique to humans and suggests fundamental differences in the pathways controlling follicular depletion. Furthermore, although aging macaques show some parallel features in neuroendocrine profiles, they do not experience the same dramatic rise in FSH or prolonged postreproductive life span characteristic of human menopause (Alberts et al. 2013; de Jesus Rovirosa Hernandez et al. 2017). These differences, combined with ethical considerations and practical limitations, necessitate careful interpretation of findings from NHP studies.

Whales

Among mammals, only humans and four species of toothed whales experience natural menopause followed by an extended postreproductive life span (Ellis et al. 2018, 2024). This shared characteristic makes whales particularly interesting for understanding the evolutionary basis of ovarian aging. However, the mechanisms driving reproductive cessation likely evolved independently in these species, as evidenced by their divergent evolutionary histories and different ecological pressures. Although this convergent evolution of postreproductive survival provides valuable evolutionary insights, whales’ aquatic nature, protected status, and significant differences in reproductive physiology from humans prevent detailed mechanistic studies and limit their practical relevance as experimental models.

Spiny mice

The common spiny mouse (Acomys cahirinus) has emerged as a promising new model that shares several key reproductive characteristics with humans. Unlike traditional laboratory rodents, this species exhibits true menstruation, endocrine profiles more closely resembling human patterns, and longer gestation periods (Bellofiore et al. 2017, 2021). These features make it particularly suitable for studying certain aspects of female reproductive physiology.

However, important differences remain. Spiny mice do not exhibit the same pattern of accelerated follicular depletion in midlife characteristic of human ovarian aging or experience a similar length of postreproductive life span. Their reproductive cycling, while more similar to humans than traditional laboratory rodent models, still differs significantly in duration and hormonal patterns. Although these animals may prove valuable for studying specific aspects of reproductive physiology, they cannot fully model human ovarian aging.

Human clinical conditions

Clinical conditions that accelerate ovarian aging provide unique windows into aging mechanisms while also highlighting the complexity of natural ovarian aging. Breast cancer patients receiving aromatase inhibitor therapy experience rapid onset of menopausal symptoms and accelerated ovarian aging (Hong et al. 2017), offering opportunities to study both the mechanisms of ovarian aging and potential protective interventions. However, the sudden disruption of hormone signaling in these cases differs significantly from the gradual changes observed in natural reproductive aging, limiting the generalizability of findings.

In vitro models

Recent technological advances have enabled the development of sophisticated in vitro systems for studying ovarian function, though each has important limitations in modeling the full complexity of ovarian aging.

Ovarian organoids

Three-dimensional ovarian organoids represent a significant advance in modeling tissue-specific aging processes. These structures can recapitulate key aspects of ovarian architecture and function, including follicular organization and hormone production (Del Valle and Chuva de Sousa Lopes 2023; Dipali et al. 2024). Recent developments in extracellular matrix composition and growth factor combinations have enhanced their physiological relevance (Nason-Tomaszewski et al. 2024; Dadashzadeh et al. 2025).

However, current organoid systems lack the complex endocrine signaling networks that regulate ovarian function, particularly the hypothalamic–pituitary–ovarian axis interactions. Although organoids derived from malignant patient cells have provided insights into cancer biology, their utility for studying normal aging processes remains limited by their inability to reproduce the decades-long process of natural ovarian aging or the influence of systemic factors.

Microfluidic systems

Advanced microfluidic “organ on a chip” technologies have revolutionized the ability to model ovarian function and aging in vitro. These systems enable precise recreation of the complex ovarian microenvironment, including spatiotemporal gradients of hormones and growth factors (Xiao et al. 2017; Campo et al. 2023). The ability to maintain viable ovarian tissue with proper cellular organization and function for extended periods allows for detailed studies of aging processes.

However, neither organoids nor microfluidic systems can fully recapitulate the complexity of in vivo aging. They lack the broader physiological context, including the immune system interactions, systemic metabolism, and long-range hormonal signaling that influence reproductive aging. Although they excel at modeling specific cellular interactions, their artificial nature and inability to model long-term aging processes require careful consideration when extrapolating findings to menopause and ovarian aging.

Methods to evaluate ovarian function throughout life

Limitations of broadly used methods to measure ovarian function

Accurately assessing ovarian function faces significant technical challenges that have historically limited our understanding of ovarian aging. Traditional methods each present distinct limitations that must be carefully considered when interpreting results.

Rodent E2 measurement issues

A major limitation in ovarian function assessment lies in the measurement of estradiol (E2), particularly in rodent models. The presence of equol, a metabolite produced by gut bacteria from soy-derived isoflavones, can substantially interfere with E2 measurements in conventional immunoassays (Barkley et al. 1985; Thompson et al. 1985). Unfortunately, soy is a major component of the standard rodent chow used in most research settings. This interference stems from structural similarities between equol and E2, leading to cross-reactivity that produces artificially elevated readings (Barkley et al. 1985; Thompson et al. 1985; Brown and Setchell 2001; Allred et al. 2005; Ward et al. 2005). Addressing this issue may raise further challenges because the alternative is to use expensive soy-free synthetic diets, which may themselves impact ovarian and organismal health beyond the removal of interfering equol.

An additional critical challenge in quantifying steroid hormones stems from their binding to carrier proteins in circulation. In both humans and rodents, the majority of steroid hormones, including estradiol, progesterone, and testosterone, circulate bound to proteins such as sex hormone-binding globulin (SHBG) and albumin (Hammond 2016). These protein–hormone complexes effectively mask the hormones from direct detection in many assay formats (Bikle 2021). Consequently, accurate measurement requires sample processing steps that disrupt these binding interactions, such as organic solvent extraction or enzymatic treatment, adding complexity to the workflow and introducing potential sources of variability.

Importantly, the inherently small sample volumes available from rodent models present another significant obstacle. Mice typically yield only 50–100 µL of plasma per collection, which severely restricts the number of analytes that can be measured unless from terminal collection and often precludes technical replicates. This volume limitation becomes particularly problematic when comparing with human samples, where 1–5 mL of plasma or serum may be readily available. The restricted volume compounds the challenge of detecting steroid hormones that naturally circulate at low concentrations in rodents. For example, basal estradiol levels in female mice (5–10 pg/mL) (Haisenleder et al. 2011) are significantly lower than those in humans (30–400 pg/mL, depending on cycle phase) (Frederiksen et al. 2020).

Finally, many commercially available immunoassays lack sufficient sensitivity to detect physiological concentrations of steroid hormones in rodent samples. The lower limits of detection for standard ELISA kits often approach or exceed the normal concentrations found in mouse models, particularly for estradiol (Haisenleder et al. 2011; Brouillard et al. 2025). Although mass spectrometry-based methods offer improved sensitivity and specificity, they require specialized equipment and expertise not widely available in reproductive biology laboratories (Nilsson et al. 2015). These technical limitations significantly constrain our ability to accurately characterize the hormonal milieu associated with ovarian aging in rodent models.

Luteinizing hormone and follicle-stimulating hormone assays

Recent studies have identified a concerning loss of antibody recognition in commonly used luteinizing hormone (LH) assays. This issue appears to stem from changes in glycosylation patterns of LH that occur with aging, potentially leading to underestimation of LH levels in older subjects (Wide and Eriksson 2013). The reduced antibody binding affects the reliability of LH measurements, particularly in studies examining age-related changes in reproductive function.

Similar challenges plague the measurement of follicle-stimulating hormone (FSH), which exists as multiple glycoforms with differential bioactivity (for review, see Das and Kumar 2018). The carbohydrate moieties attached to the FSH protein backbone vary considerably, resulting in multiple isoforms with distinct physiological actions at the receptor level. Most commercial immunoassays use single antibodies that recognize only specific epitopes, potentially missing important glycoform variants. This limitation is particularly problematic in aging studies, as the glycosylation pattern of FSH changes significantly across the reproductive life span, with documented shifts in sialic acid content and branching complexity (Bousfield et al. 2018). Consequently, standard FSH assays may detect only a subset of circulating FSH molecules, leading to inaccurate quantification and potentially obscuring important biological trends. This glycoform-specific detection issue introduces considerable uncertainty when interpreting FSH levels as markers of ovarian reserve or aging.

Monotarget ELISA kits

Traditional monotarget ELISA kits present significant limitations for comprehensive assessment of ovarian function. These assays can measure only one analyte at a time, making it difficult to capture the complex interplay of multiple hormones and proteins involved in ovarian function. Furthermore, the sequential nature of single-target measurements can introduce temporal variations that complicate data interpretation, especially when studying dynamic processes like follicular development.

Traditional histological evaluation

Traditionally, the progression of ovarian aging has been evaluated by direct quantification of follicles on postmortem histological evaluation of ovarian tissue to evaluate ovarian reserve. However, conventional histological evaluation methods present significant constraints. Standard sectioning and staining techniques provide only two-dimensional snapshots of three-dimensional structures, making it challenging to accurately assess follicle numbers and stages. The physical sectioning process can introduce artifacts and may lead to double counting or missing follicles entirely (Adeniran et al. 2021). Traditional approaches often fail to capture the spatial relationships between different ovarian compartments and cell types, limiting our understanding of age-related changes in tissue architecture.

Modern tools for modern research!

Advanced imaging techniques

Recent technological advances have revolutionized our ability to visualize and analyze ovarian tissue architecture and function across multiple scales. High-content microscopy enables systematic, large-scale analysis of structural and molecular changes in ovarian tissue with unprecedented resolution and throughput. This approach has been particularly valuable for quantifying follicle dynamics and identifying cellular hallmarks of reproductive aging (Folts et al. 2024).

The development of tissue clearing techniques has transformed our understanding of three-dimensional ovarian organization (Soygur and Laird 2021). Methods like CLARITY and CUBIC allow visualization of intact ovaries while maintaining spatial relationships between different cell types and structures (Feng et al. 2017; McKey et al. 2020; Soygur et al. 2023). This has provided crucial insights into how the complex ovarian architecture changes during aging, particularly regarding follicle distribution patterns and vascular networks that may influence follicle survival and function.

In vivo imaging approaches have emerged as powerful tools for studying dynamic processes in living ovarian tissue. Intravital microscopy enables real-time observation of follicle development, ovulation, and age-related changes in the local microenvironment (Bochner et al. 2015; Feng et al. 2018). These techniques have revealed previously unappreciated aspects of ovarian aging, such as alterations in blood flow patterns and immune cell trafficking that may contribute to declining organ function.

Improved biomarker assessment

The field has largely moved beyond traditional monotarget approaches to embrace multiplexed analysis of ovarian aging biomarkers. Modern multiplex immunoassay platforms allow simultaneous quantification of dozens of proteins, metabolites, and other molecules from small sample volumes (Cai et al. 2023). This has enabled more comprehensive profiling of age-related changes in the ovarian microenvironment and identification of novel biomarker signatures associated with reproductive aging.

Although anti-Müllerian hormone (AMH) remains an important indicator of ovarian reserve, research has revealed numerous additional promising biomarkers. Inflammatory mediators have emerged as important indicators that may reflect the “inflammaging” of ovarian tissue, including from increased immune infiltration, upregulation of inflammatory genes, and multinucleated giant cells formed by macrophage fusion (Isola et al. 2024a; Converse et al. 2025). Metabolic markers provide insight into altered energy utilization patterns (Mani et al. 2024; Sun et al. 2024; Zeng et al. 2025), whereas oxidative stress markers suggest compromised cellular function (Yan et al. 2022). Recent studies have also identified novel miRNAs involved in follicular development and survival (Yerushalmi et al. 2018; Kamalidehghan et al. 2020). The integration of multiple biomarkers using machine learning approaches has improved our ability to predict reproductive aging trajectories and identify individuals at risk for accelerated ovarian aging.

“Omics” approaches to study ovarian landscape

Single-cell genomics—Leveraging single-cell approaches to understand cell type-specific gene regulation in the aging ovary provides unprecedented resolution and throughput to understand the heterogeneity of cellular responses. Notably, a major limitation of most droplet-based technologies (which are most readily available for high-throughput analysis; e.g., 10x Genomics) is that the required size-based exclusion for cellular transcriptome profiling necessarily leads to the loss of larger cells, such as oocytes and multinucleated giant cells, which are crucial in the context of the aging ovary. Indeed, over the recent years, pioneering studies have used this technology to evaluate molecular drivers of ovarian aging in mice (Ben Yaakov et al. 2023; Isola et al. 2024b; Wang et al. 2024; Kim et al. 2025), nonhuman primates (Wang et al. 2020), and humans (Wu et al. 2024; Zhou et al. 2024; Devrukhkar et al. 2025; Jin et al. 2025). These analyses revealed widespread remodeling of gene regulation across ovarian cell types with aging, with many features conserved across species, including activation of fibrotic pathways in stromal cells, increased stromal cell senescence, profound immune remodeling, increased DNA damage signaling, and loss of follicular cells (i.e., granulosa/theca cells) (Wang et al. 2020, 2024; Ben Yaakov et al. 2023; Wei et al. 2023; Isola et al. 2024b; Wu et al. 2024; Zhou et al. 2024; Jin et al. 2025). Conversely, based on these single-cell atlases, many features of ovarian aging may be species-specific; for example, increased immune infiltration with mouse ovarian aging (Ben Yaakov et al. 2023; Isola et al. 2024b; Wang et al. 2024). Importantly, these atlases have also revealed potential regulators of ovarian aging that may represent future therapeutic targets, including transcription factors FOXP1 and CEBPD, mTOR signaling, and macrophage pyroptosis.

Spatial transcriptomics—Studying gene expression profiles in situ in the context of undissociated tissue provides important information about niche interactions and tissue heterogeneity. This is all the more important for tissues with heterogeneous structure like the ovary, where the cortex and medulla host different cell types performing different key functions. This kind of study has been made possible in the past 6 years thanks to new technological developments, which collectively have led to the ability to profile tissue transcriptomes at higher throughout while preserving some spatial information (e.g., 10x Genomics Visium) (Williams et al. 2022; Vandereyken et al. 2023). However, a major caveat to these studies is that most commercially available spatial transcriptomics technologies do not have single-cell resolution. Recent advances using Slide-seq, which enables near-single-cell resolution with 10 mm beads, have revealed unprecedented details of ovarian structure across cycling and aging in mouse tissue (Lan et al. 2024). Because spatial transcriptomics is still in its infancy, only a few spatial studies of ovarian aging have been conducted, though data have been obtained in mice (Russ et al. 2022; Wei et al. 2022; Lan et al. 2024; Huang et al. 2025), nonhuman primates (Lu et al. 2024) and humans (Jones et al. 2024; Wu et al. 2024). These studies have revealed spatial heterogeneity in ovarian aging at the molecular and cellular levels, including signaling hubs likely driving senescent cell accumulation and fibrosis. By allowing subtissue analyses, targets that may be hidden in classical single-cell approaches are now made accessible to scientists, with the potential to reveal new conserved molecular targets against ovarian aging.

Proteomics and metabolomics—High-resolution mass spectrometry has revolutionized our ability to profile the molecular landscape of ovarian aging. Proteomics studies have revealed complex changes in protein expression and post-translational modifications associated with reproductive aging (Bomba-Warczak et al. 2024). Significant alterations have been observed in mitochondrial proteins, suggesting compromised energy metabolism with age. Changes in DNA repair factors indicate accumulated genetic damage, whereas shifts in inflammatory mediators reflect chronic tissue stress. Modifications in extracellular matrix components demonstrate altered tissue architecture during the aging process.

Metabolomics approaches have provided complementary insights into altered biochemical pathways during ovarian aging. Researchers have observed significant changes in energy metabolism and mitochondrial function, along with shifts in lipid metabolism and membrane composition. Studies have also revealed alterations in amino acid utilization and protein homeostasis, as well as modifications in oxidative stress responses and antioxidant systems. The integration of proteomic and metabolomic data has revealed novel molecular networks and potential therapeutic targets for maintaining ovarian function during aging.

Computational approaches

Together with the advent of new, more powerful “omic” profiling techniques, new computational and machine learning approaches have been developed to better understand patterns associated with complex phenotypes like aging. For instance, leveraging knowledge about ligand/receptor signaling and accumulating predictive algorithms to elucidate potential changes in cell-to-cell communication from single-cell and spatial transcriptomics data sets (Dimitrov et al. 2022; Lan et al. 2024). Indeed, this has already been used in studies of ovarian aging using single-cell RNA-seq (Ben Yaakov et al. 2023; Isola et al. 2024b; Wang et al. 2024; Jin et al. 2025) and spatial transcriptomics (Lan et al. 2024; Wu et al. 2024), revealing changes in cellular communication in the ovary driving fibrosis; for example, TGFb signaling. Emerging studies have been leveraging machine learning methods to build “clocks” of aging based on stereotypical transcriptional changes in specific cell types in mouse and human aging tissues (e.g., adipose, blood, brain, liver, lungs, and skeletal muscle) (Buckley et al. 2023; Yu et al. 2023; Zhu et al. 2023; Neumann et al. 2024). To date, algorithms based on elastic net regression seem to outperform others in that space, with the benefit of model interpretability. Although this has not yet been applied to the aging ovary, it is likely that application of such methods will provide important insights into molecular drivers and biomarkers of ovarian aging.

Future directions and conclusions

The field of ovarian aging research stands at a critical juncture where the integration of multiple approaches may help overcome the limitations of individual models and methods. The complexity of human aging necessitates innovative strategies that can bridge the gaps in our current understanding.

Integration of multiple modern approaches

Advances in single-cell genomics combined with spatial transcriptomics now enable mapping of molecular changes during ovarian aging with unprecedented resolution. These technologies have revealed widespread remodeling of gene regulation across ovarian cell types with aging, identifying features conserved across species, including activation of fibrotic pathways in stromal cells, increased stromal cell senescence, profound immune remodeling, increased DNA damage signaling, and loss of follicular cells. Integration of in vivo imaging with molecular profiling provides dynamic views of tissue-level changes, whereas artificial intelligence approaches enable analysis of these complex data sets. However, the challenge remains in translating findings from these high-resolution snapshots into mechanistic understanding of the decades-long process of human ovarian aging.

Translation to clinical applications

The development of clinically relevant interventions requires careful consideration of the unique features of human reproductive aging. Noninvasive diagnostic tools that inform women about both their reproductive span trajectory and rate of ovarian aging are on the horizon (Zaniker et al. 2024). For interventions that impact fertility, early intervention points are particularly crucial given the accelerated decline in oocyte numbers that occurs in women's mid to late 30s (Wallace and Kelsey 2010). The creation of personalized treatment strategies must account for individual genetic variation in both initial oocyte stockpile and depletion rates. Although prophylactic approaches show promise in certain clinical scenarios, such as fertility preservation before cancer treatment, developing interventions for ovarian aging presents distinct challenges due to the complex interaction between ovarian function and systemic health.

Expanding research to other female-specific organs

Understanding ovarian aging provides a template for investigating aging in other female-specific organs, though each tissue presents unique challenges. The uterus, which undergoes dramatic remodeling throughout reproductive life, requires distinct approaches for studying age-related changes in both structure and function (Tinelli et al. 2023). Fallopian tube aging may influence both fertility and disease susceptibility but presents significant technical challenges for in vivo study (Weigert et al. 2025). Aging of vaginal tissue and its microbiome demonstrates the importance of considering both host and microbial factors in reproductive health (Shardell et al. 2021). Each of these tissues requires specialized approaches that can account for their distinct biology while considering their integration within the broader reproductive system.

Toward precision medicine in women's health

The advancement of precision medicine in women's health reveals both the promise and challenges of personalizing interventions for ovarian aging. Finding and developing aging biomarkers is essential to account for the significant individual variation in ovarian reserve and depletion rates. The integration of reproductive health data with broader health metrics is particularly crucial given the systemic effects of ovarian function, yet current healthcare systems often fail to capture these connections adequately, and critical metrics like age at natural menopause (ANM) are rarely recorded.

These challenges are exemplified by the current lack of proper medical coding for many female-specific procedures. Essential interventions such as ovarian detorsioning lack CPT codes, whereas complex procedures like endometrial excision for a uterus >250 g lack insurance billing codes, often leading to unnecessary hysterectomies. These systemic barriers highlight how advancing scientific understanding must be coupled with healthcare system changes to improve outcomes.

The development of clinically relevant interventions requires careful consideration of the unique features of human ovarian aging. Noninvasive diagnostic tools and early intervention points are particularly crucial given the accelerated decline in oocyte numbers that occurs in women's mid to late 30s. The creation of personalized treatment strategies must account for individual genetic variation in both initial oocyte stockpile and depletion rates. Although prophylactic approaches show promise in certain clinical scenarios, such as fertility preservation before cancer treatment, developing interventions for natural reproductive aging presents distinct challenges due to the complex interaction between ovarian function and systemic health.

Competing interest statement

The authors declare no competing interests.