Beyond apoptosis: evidence of other regulated cell death pathways in the ovary throughout development and life

Apoptosis

Apoptosis is the programmed, controlled death of a cell; which involves degradation and fragmentation of protein and DNA, and engulfment of the collapsed cell by neighbouring cells and/or phagocytes in a non-inflammatory manner. It occurs in all multicellular organisms throughout life from foetal development onwards and is an essential homeostatic mechanism to maintain healthy cell populations in tissues and organs. There are two major apoptotic pathways: the intrinsic and extrinsic pathways (Fig. 2). The intrinsic (mitochondrial) pathway is activated from within the cell and is predominately regulated by mitochondria. On the other hand, the extrinsic (death receptor) pathway is triggered from outside the cell, typically in response to conditions and factors within the extracellular environment.

Autophagy and autophagic cell death

Derived from the Greek language, meaning ‘self-eating’, autophagy is a tightly-regulated process whereby cells degrade and recycle their own cytosolic components inside lysosomes, which can lead to cell death (Galluzzi et al., 2018; Schwartz, 2021). Unlike apoptosis, this form of regulated cell death occurs in the absence of chromatin condensation and phagocytes. Autophagic cell death manifests in the accumulation of large numbers of autophagic vesicles containing cytoplasmic material for degradation by lysosomes, and results in early degradation of organelles and late degradation of cytoskeleton, which is the reverse for apoptotic cells (Cuervo, 2004; Thorburn, 2008).

Autophagy is mediated by dozens of autophagy-related (ATG) proteins that regulate expanding ‘isolation membranes’, which encapsulate and enclose proteins/organelles into a double-membrane structure called the autophagosome (Li et al., 2021b). These autophagosomes then fuse with liposomes to degrade the internal components. Once fused, acidic hydrolases in the lysosome can degrade the autophagic cargos, and salvaged nutrients are released to the cytoplasm to be recycled by cells. Genetic models have demonstrated that some autophagic machinery is essential for regulated cell death (e.g. ATG1, reviewed by Schwartz, 2021); however, it has been suggested that it might be more appropriate to name the process ‘autophagy-mediated cell death’ (Kroemer and Levine, 2008). Although, more recent literature suggests that autophagy-dependent cell death, that is independent of apoptosis or other regulated cell death pathways, can occur (Bialik et al., 2018; Denton and Kumar, 2019; Kriel and Loos, 2019) (Fig. 2). Indeed, there is strong evidence that apoptotic and autophagic machineries are highly interconnected during developmental regulated cell death (Zhang and Baehrecke, 2015).

Necroptosis

Necrosis is morphologically distinct from apoptosis and characterized by a gain in cell volume, organelle swelling, plasma membrane rupture, and loss of intracellular contents (Galluzzi et al., 2018). Unlike apoptosis, necrosis provokes an inflammatory response by spilling the cell’s cytosolic constituents into the extracellular space through the damaged plasma membrane. During apoptosis, however, these products are safely isolated by membranes and then consumed by phagocytes. Necroptosis is a programmed, regulated form of necrosis that is initiated by various extracellular and intracellular stressors, including viral infection (Guo et al., 2018), inflammation (Pasparakis and Vandenabeele, 2015), and factors detected by specific death receptors (e.g. FAS, TNFR1) or pathogen recognition receptors (Galluzzi et al., 2018; Frank and Vince, 2019) (Fig. 2). Importantly, these death receptors can also activate the extrinsic apoptosis pathway (Grootjans et al., 2017; Frank and Vince, 2019). At the molecular level, necroptosis critically depends on activation of the receptor-interacting serine/threonine-protein kinase (RIPK) 1/3 necrosome and mixed lineage kinase domain-like pseudokinase (MLKL), in the absence of caspase-8 activation, to ultimately cause cell membrane rupture (Galluzzi et al., 2018). Indeed, necroptosis is generally observed as a fall-back regulated cell death mechanism that is triggered when apoptosis is hindered, such as during pathogen infection (Brault and Oberst, 2017; Naderer and Fulcher, 2018).

Pyroptosis

Exogenous insults also extend to infection, and regulated cell death is proposed to contribute to immune defence against infections (Jorgensen et al., 2017). Pyroptosis is mediated by the cleavage of gasdermins, caspases (namely caspase-1), or granzymes; leading to the formation of pores in the cell membrane, lysis of the cell, and the release of inflammatory molecules (Frank and Vince, 2019) (Fig. 2). This type of regulated cell death is primarily observed in inflammatory cells, such as macrophages, and occurs most frequently upon infection with intracellular pathogens (Xia et al., 2019). As such, it is likely to form part of the host response to control bacterial, viral, fungal, or protozoan pathogens (Xia et al., 2019). While the sterile inflammatory response is required for organ development and tissue repair, dysregulation of this process may lead to inflammatory disease, for example asthma, Type 2 diabetes, and inflammatory liver diseases (Rock et al., 2010). Indeed, there is accumulating evidence that pyroptosis and inflammasome dysregulation may contribute to sterile inflammatory diseases, gynaecological diseases, autoimmune diseases, neuronal diseases, and even cancer (Li et al., 2021a; Yu et al., 2021). Moreover, cytokine dysregulation resulting in a pre-inflammatory phenotype that occurs with age, known as ‘inflammageing’ (Rea et al., 2018), has also been associated with pyroptosis (Mejias et al., 2018).

Parthanatos

Parthanatos is a poly[ADP-ribose] polymerase 1 (PARP1)-dependent and apoptosis-inducing factor (AIF)-mediated, caspase-independent cell death pathway, which is distinct from apoptosis, necroptosis, or other known forms of regulated cell death (Fig. 2). Parthanatos is associated with various diseases including several retinal diseases, Parkinson’s disease, stroke, heart attack, and diabetes (David et al., 2009). Parthanatos is triggered by an excessive ROS response, which leads to an accumulation of poly[ADP-ribose] (PAR) polymer and translocation of AIF from mitochondria to the nucleus, resulting in chromatin condensation and nuclear fragmentation (David et al., 2009; Wang et al., 2011). Parthanatos does have shared characteristics with necroptosis, including loss of membrane integrity and depletion of cellular energy stores (NAD and ATP) (Yu et al., 2003). However, cells undergoing parthanatos experience regulated chromatolysis without swelling and rupturing of cell membranes (Andrabi et al., 2008), as occurs during necroptosis. Key morphological features of parthanatos include shrunken, condensed nuclei, and membrane disintegration (Andrabi et al., 2008).

Ovarian reserve establishment

Germ cell loss occurs throughout the process of primordial follicle formation, predominately via apoptosis (McClellan et al., 2003). However, our analysis of studies performed in mice reveals some discrepancies in the exact timing of apoptosis. One of the earliest characterizations was by Bakken and McClanahan (1978), who identified germ cell degeneration (death) by their condensed nuclei containing rounded clumps of densely stained chromatin. This study reported increased germ cell degeneration during the last mitotic divisions, as oocytes become arrested in the first meiotic prophase (Bakken and McClanahan, 1978). At this time, germ cell nests begin to break down and facilitate primordial follicle assembly, culminating in loss of up to 98% of germ cells (Pepling and Spradling, 2001). Data examining germ cell loss during the early stages of meiosis (embryonic day 13.5–17.5 in mice; approximately mid-gestation in humans) are variable. Some reports highlight little to no germ cell loss and few TUNEL-positive-stained germ cells at this time (Pepling and Spradling, 2001). Meanwhile, others have identified a continuous decline and increased proportion of apoptotic germ cells during the first meiotic prophase compared to oogonia undergoing mitosis (Coucouvanis et al., 1993; McClellan et al., 2003). Thus, it is still debated by the field as to whether there are specific windows of germ cell loss during foetal development, or if this loss occurs continuously in the lead up to nest breakdown and follicle formation. Irrespective of the timing of loss, it is well-established that the key anti-apoptotic proteins (e.g. BCL-XL and BCL-2) and pro-apoptotic proteins (e.g. BAX, BMF, and PUMA) are clearly required for this process (Rucker et al., 2000; Flaws et al., 2006; Ke et al., 2013; Myers et al., 2014; Vaithiyanathan et al., 2016).

Extensive germ cell loss occurs during nest breakdown. It has been proposed that these dying cells could be acting as nurse-like cells that transfer important organelles (including centrioles, Golgi, and mitochondria) and the cytoplasm to a select few oogonia, before undergoing regulated cell death (Lei and Spradling, 2016). The remaining oogonia that acquired these additional cellular components become progressively larger, mature, and form the primordial follicle pool. Apoptosis is fundamental to orchestrating nest breakdown and primordial follicle assembly (Lei and Spradling, 2016). However, autophagy likely also plays a key role in this process (recently reviewed by Bhardwaj et al., 2022). Interestingly, inhibiting autophagy mediators using 3-methyladenine delays follicle formation and results in an increased number of cyst oocytes (Rodrigues et al., 2009; Zhihan et al., 2019). In support, germ cell survival immediately prior to and during follicle formation is severely impaired in mice with genetic loss of autophagy-related genes Atg7 and Becn1 (Gawriluk et al., 2011; Song et al., 2015). Lysosome amplification—a hallmark of autophagy—is also observed in oocytes upon birth and is most apparent in primordial follicle oocytes (Rodrigues et al., 2009). Lysine-specific demethylase 1 (LSD1)—a critical repressor of autophagy—is highly expressed in mouse foetal ovaries, but sharply reduces during the period of follicle assembly. Expression data, together with functional studies in vitro, indicate that LSD1 is an indispensable regulator of oocyte death during ovarian reserve establishment, via regulation of autophagy (He et al., 2020). During the foetal-to-neonatal transition in mammals, which is associated with a transformation in nutrient supply from the maternal–foetal blood interface to lactation, there is a disruption in nutrient supply to the ovary (Kuma et al., 2004). Interestingly, autophagy has been shown to play a critical role in nutrient stress adaptation to prevent excessive germ cell loss during this period in mice (Song et al., 2015; Sun et al., 2018). Together, these studies imply key roles for autophagy mediators in regulating oocyte survival, particularly during nest breakdown and primordial follicle formation.

Once established, primordial follicles enter meiotic arrest. Deemed ‘non-growing follicles’, the oocytes remain arrested at the diplotene stage of meiotic prophase 1, and the surrounding granulosa cells have low mitotic potential (Hartshorne et al., 2009). These primordial follicles represent the stockpile of oocytes available to females for their reproductive life. As such, these primordial follicle oocytes are some of the longest living cells in the mammalian body, and may remain arrested for decades in humans (Pelosi et al., 2015). These follicles will ultimately leave this period of dormancy, either by activation to continue through stepwise maturation via follicle development, or to undergo follicle atresia and oocyte loss. Regulation of this process and a balance between growth, survival, and atretic factors are essential in maintenance of normal reproductive function.

Puberty

During puberty, rising gonadotrophin levels cause dynamic physiological changes in the ovary, including the development of antral follicles to the Graafian (pre-ovulatory) stage and the onset of ovulation. Strikingly, this window also coincides with a significant spike in the loss of the ovarian reserve of primordial follicles, by approximately half during the adolescent/young adult period (ages 13–25 years) in humans (Wallace and Kelsey, 2010) and roughly two-thirds in mice (Allan et al., 2006; Bristol-Gould et al., 2006). This loss is thought to be gonadotrophin-mediated, though the precise mechanisms remain unknown (Liew et al., 2017). Notably, overexpression of LH in juvenile mice has been shown to trigger depletion of the primordial follicle reserve (Flaws et al., 1997). Conversely, LH and FSH suppression prevented follicle loss in a more recent study (Liew et al., 2017). Importantly, this primordial follicle loss is regulated by apoptosis, as demonstrated by the essential requirement for the pro-apoptotic BH3-only protein, BMF, in this process (Liew et al., 2017).

It remains unclear why such vast numbers of primordial follicles are eliminated at the time of sexual maturation. One explanation for this may be the fact that there are two separate populations of primordial follicles in the ovary that each have distinct functional roles. Class 1 primordial follicles, which are localized to the ovarian medulla, will activate and grow during pre-pubertal life but never be ovulated. These are fast growing, taking 19–21 days to reach maturity in mice (Gilchrist et al., 2001). They will contribute to the first wave of follicle activation, which is in turn likely required for the establishment of the hypothalamic–pituitary–ovarian axis, and puberty onset. Class 2 primordial follicles, located at the ovarian cortex, are slower growing (∼47 days) and are thought to represent the source of all mature ovulatory oocytes for fertilization (Mork et al., 2012; Zheng et al., 2014). Therefore, it is possible that the spike in oocyte loss observed at the transition of puberty involves the clearance of any remaining Class 1 primordial follicles in the medulla. While the two classes of follicles have not been clearly identified in human ovaries, the growth pattern of the first wave of activated follicles during foetal development is conserved in humans (Lintern-Moore et al., 1974; Peters et al., 1975). Moreover, in human ovaries, nearly 20%, 5%, and 0% of the primordial follicles in pre-pubertal, pubertal, and adult ovaries, respectively, are classed as morphologically abnormal (Anderson et al., 2014). Cultures of human cortical tissue containing primordial follicles and isolated pre-antral follicles from pre-pubertal and pubertal girls exhibited low activation rates and compromized oocyte growth respectively, compared to adult samples (Anderson et al., 2014). Thus, it can be speculated that these abnormal primordial follicles are either eliminated or preferentially activated before puberty onset and the commencement of ovulation in humans (Anderson et al., 2014). This provides an intriguing parallel to the two distinct types of primordial follicles identified by Mork et al. (2012) in mice (Anderson et al., 2014).

Follicle atresia throughout reproductive life

Atresia is a complex process that naturally occurs to regulate the follicle pool across the lifespan (Regan et al., 2018). It affects all stages of follicular development and involves multiple forms of regulated cell death. The highest incidence of follicular degeneration is observed when follicles become dependent on FSH, at the early antral follicle stage (Chun et al., 1996). Atresia is essential for maintaining ovarian homeostasis, and the dysregulation of atresia contributes to reproductive disorders, including PCOS and POI (Duncan, 2014).

Ovarian ageing

Reproductive ageing coincides with a decline in ovarian follicle number, leading to loss of fertility and endocrine function, and eventually menopause. Notably, emerging evidence indicates that the ovary naturally transitions to a low-level inflammatory microenvironment with advancing maternal age, termed ‘inflammageing’ (Lliberos et al., 2021; Camaioni et al., 2022; Umehara et al., 2022). The NLRP3/apoptosis-associated speck-like protein (ASC) inflammasome, which activates caspase-1, appears to be central to this process, raising the possibility that pyroptosis might contribute to age-associated follicle loss. Although pyroptosis has not been directly studied during ovarian ageing, this hypothesis is supported by recent reports demonstrating that genetic loss or pharmacological inhibition of NLRP3 or ASC reduces caspase-1 levels in the ovary, increases oocyte number, and delays ovarian ageing (Lliberos et al., 2020; Navarro-Pando et al., 2021). Furthermore, the granulosa cells of women with POI exhibit elevated NLRP3, caspase-1, and IL-1α levels (Navarro-Pando et al., 2021). Further investigations of how pyroptosis is regulated in the ovary will provide valuable therapeutic targets to potentially delay natural ovarian ageing.

Some new evidence suggests that parthanatos may also contribute to premature ovarian ageing. Cumulus granulosa cells collected from women with diminished ovarian reserve showed increased levels of nuclear purified PAR and AIF (Batnasan et al., 2020), suggesting a role for PARP-dependent cell death in diminished ovarian reserve pathophysiology. Therefore, inhibition of parthanatos, amongst other regulated cell death pathways, may prove useful for patients with diminished ovarian reserve.

Distinguishing between oocyte versus somatic cell death in ovarian follicles

Granulosa cells are specialized somatic cells in the ovary, vital for oocyte survival and female fertility. When follicles are formed, oocytes not surrounded by granulosa cells are eliminated. Impaired granulosa cell function also dysregulates oocyte growth and causes POI, characterized by early loss of fertility and hormone production (Uda et al., 2004). Recent endeavours to derive human or mouse oocytes from induced pluripotent stem cells rely exclusively on mouse granulosa cells to support the germ cell-like cells (Sarma et al., 2019; Stringer and Western, 2019). Collectively, these observations reveal that granulosa cells have unique and essential functional properties that cannot be replaced by other somatic cell types.

In response to activation signals, primordial follicles give rise to large, hormone-producing follicles, and mature ovulatory oocytes. The founding population of ∼5–8 granulosa cells present in a primordial follicle undergoes clonal divisions to eventually produce >2000 granulosa cells that support a mature oocyte (Hirshfield, 1991). After primordial follicle formation, there is no evidence of new granulosa cell formation from other somatic cell types (Zhang et al., 2014). Thus, all mature granulosa cells appear to be clones of primordial follicle granulosa cells. As oocytes cannot survive without these essential granulosa cells, granulosa cell apoptosis invariably causes follicle atresia. However, the possibility that primordial follicle granulosa cells are susceptible to ageing and exogenous insults has not been investigated. Understanding the regulated cell death pathways utilized by each cell type comprising ovarian follicles is important to better understand the effects of maternal ageing, and to develop appropriate fertility preservation strategies for female cancer patients.

Role of the ovarian environment in regulated cell death of ovarian follicles

Studies of the ovarian environment, including the stroma and other supporting somatic cell types, have emerged as a recent area of interest (Kinnear et al., 2020). Some studies have established that fibrosis is one early hallmark of the aging ovarian stroma (Briley et al., 2016; Amargant et al., 2020; Umehara et al., 2022). It has thus been proposed that this altered microenvironment could contribute to the age-associated decline in oocyte number and/or quality. However, a direct link between the ovarian stroma and regulated cell death in ovarian follicles or other somatic cell types is lacking, and should be the focus of further investigation.

Targeting regulated cell death pathways to protect fertility

Inhibiting key mediators of oocyte death is a promising strategy for protecting primordial follicles from anti-cancer treatment and age-mediated depletion. Such strategies require characterization of the pathways and specific mediators involved in order to identify suitable targets. Of relevance to cancer treatment, Bax^–/– mice have prolonged fertility after chemotherapy (Kujjo et al., 2010). Similarly, irradiated TAp63^–/– (Stringer et al., 2020), Puma^–/– (Kerr et al., 2012b), and checkpoint kinase 2 (Chk2)^–/– mice (Bolcun-Filas et al., 2014; Rinaldi et al., 2017) remain fertile after treatment with radiation or chemotherapy, unlike their wild-type counterparts, demonstrating the potential of this strategy for fertility preservation. Translation of this information is currently limited by the availability of effective small-molecule inhibitors to these proteins. However, an inhibitor of CHK2 does exist, and its transient use reduced irradiation-mediated primordial follicle loss in mice, leading to the generation of healthy offspring (Rinaldi et al., 2017).

Targeting other regulated cell death pathways, independently or in concert with apoptosis, may also provide novel fertility preservation strategies. Indeed, investigations on the use of AChE and necroptosis inhibitors to improve folliculogenesis (Urra et al., 2016; Du et al., 2018), or inflammasome inhibitors to delay ovarian ageing (Navarro-Pando et al., 2021; Umehara et al., 2022), are exciting potential avenues for fertility preservation. However, more studies are needed to understand when and how these pathways are activated in the ovary, in order to determine the best timing and suitability of these therapies.