Diverse roles of stress-responsive RNP granules in oogenesis and infertility

Abstract

Effectively responding to cellular stress (e.g., nutrient deprivation, oxidative stress) is essential for cell and organismal survival. A protective mechanism is especially critical in developing oocytes, where a prolonged quiescent state and the inability to divide render oocytes highly susceptible to accumulating stress that can result in cell death if unaddressed. Despite the common view that stress granules are the primary stress-responsive ribonucleoprotein granule, accumulating evidence shows that in ovaries, other ribonucleoprotein granules also uniquely mediate gene regulation in response to stress. Here, we review recent insights into ribonucleoprotein granule dynamics and ribonucleoprotein granule protein function during stress in the context of oogenesis among both invertebrates and vertebrates, with an emphasis on insights from Drosophila and mice. We also discuss roles for stress-responsive ribonucleoproteins in maintaining stem cell populations and complicating fertility treatments. By exploring how stress-induced ribonucleoprotein dynamics can impact oogenesis, both positively and negatively, we can better understand how stress contributes to reduced fecundity and infertility. We conclude by offering key research questions that can drive the next generation of insights.

Introduction

In humans, oocytes are limited in number and remain in a quiescent state for decades while being arrested in meiosis [1]. During this prolonged duration, exposure to cellular stressors can significantly impact oocyte quality and reproductive outcomes. Like most cells, oocytes can experience stress from a variety of insults [2–13]; however, unlike other quiescent stem cells, which can differentiate to avoid stress [14], or divide and undergo mitotic lysosomal exocytosis (a process hypothesized to eliminate toxic molecules) [15, 16], oocytes are unable to do so to escape stress. Consequently, chronic stress in oocytes or oocyte-associated cells typically results in apoptosis or autophagy and presents as ovarian failure, polycystic ovary syndrome (PCOS), or endometriosis-associated infertility in a variety of human studies and model organisms [4, 17–20].

Oogenesis

To begin to understand how stress impacts oogenesis, it is important to understand how oocytes develop with the assistance of both germ cell–derived nurse cells and somatic follicle cells. We have described these events in the context of Drosophila and mammalian oogenesis below.

Drosophila oogenesis begins in the germarium, where germ stem cells undergo asymmetric cell division to generate a stem cell and a partially differentiated daughter cell called a cystoblast. The cystoblast undergoes four rounds of mitotic division to form a 16-cell cyst, where all 16 cells are connected by cytoplasmic bridges. One of these cells initiates meiosis to become the oocyte, while the remaining 15 exist as nurse cells that serve to supply the transcriptionally silent oocyte with RNAs and proteins necessary for early embryonic development [21]. Between germarium stage 2a/2b, each cyst becomes encased by somatic follicle cells to make up an egg chamber that then proceeds through 14 stages of development. Follicle cells supply yolk proteins to the oocyte and promote vitellogenesis starting at egg chamber stage 8 [22, 23]. Starting at stage 11, nurse cells dump their cytoplasm into the oocyte and undergo nonapoptotic programmed cell death mediated by follicle cells [24, 25]. Following nurse cell death, follicle cells secrete proteins necessary for the chorion and also undergo cell death, presumably by autophagy, though the exact mechanisms are not fully understood [26].

Mammalian oogenesis has a similar progression to Drosophila oogenesis. Primordial germ cells initially undergo mitotic divisions to form up to 32 cell cysts, which then fragment into four to six groups [27]. Within each divided group, all cells initiate meiosis, but it is not clear how many become oocytes [28]. Those that do not become oocytes act as nurse-like germ cells [29]. Transfer of cytoplasmic material from nurse-like germ cells to oocytes occurs via membrane breakdown and programmed cell death of the nurse cells prior to primordial follicle formation [29, 30]. The growing oocyte is then encased by somatic follicle cells called granulosa cells that form gap junctions with the oocyte, allowing for exchange of nutrients and small molecules between granulosa cells and the oocyte [31]. With assistance from these granulosa cells, the oocyte matures through three follicle stages of growth (primary, secondary, and tertiary) prior to ovulation. Unlike in Drosophila, the mammalian oocyte is transcriptionally active during initial follicle growth but ceases transcription just prior to the oocyte resuming meiosis in the large antral follicle [32, 33]. At the antral stage, granulosa cell transcription increases and some differentiate into cumulus cells to assist in meiosis, ovulation, and fertilization [34, 35].

Two important commonalities exist between oocytes across species: (1) the maturing oocyte depends on nutrients, messenger RNAs (mRNAs), proteins, and small molecules from both nurse cells and somatic follicle cells, and (2) the oocyte stores translationally repressed mRNAs long term in preparation for development. How does stress impact these activities?

Roles of the integrated stress response in oocyte development

Cellular response to stress is largely carried out by the integrated stress response (ISR), a highly conserved eukaryotic pathway that manages cellular homeostasis by inhibiting translation and coordinating with other stress pathways [36–40]. Oogenesis relies on an efficient ISR for oocyte survival even in the absence of obvious stress. In mammalian ovaries, the ISR is naturally most active in mature oocytes, granulosa cells, and oocytes of primary follicles, where expression of all four ISR kinases, PERK, HRI, GCN2, and PKR, are upregulated compared to earlier staged primordial follicles and non-ovarian tissues including brain and mouse embryonic fibroblasts (Figure 1) [41, 42]. Each ISR kinase is activated by a specific stressor: in mammals, PERK responds to misfolded proteins and endoplasmic reticulum (ER stress), HRI responds to heme deprivation and oxidative stress, GCN2 responds nutrient deprivation, and PKR responds to viral infection [36]. Drosophila only have Pek (PERK homologue) and Gcn2, although Pek appears to compensate for the lack of an HRI homologue in response to oxidative stress [43]. Evidence suggests the ISR partakes in at least two roles in oocyte development. In mature oocytes, an active ISR likely inhibits translation in unfertilized oocytes [42], while in granulosa cells and oocytes of primordial and primary follicles, the ISR regulates follicle growth and cellular repair [41]. In this later role, a balance is necessary for functional oogenesis: too much ISR activation in granulosa cells can lead to granulosa cell apoptosis and follicle death, which can present as endometriosis-associated infertility and PCOS [4, 17–19], while too little ISR activation in primordial follicles can be problematic as well and lead to reduced activation of primordial follicles to primary follicles, similar to that seen in primary ovarian insufficiency [41, 44]. How does environmental stress impact this balance? Many of the primary non-genetic factors that decrease female fecundity, in theory, should directly and chronically activate the ISR through one of its four kinases. This includes obesity and insulin resistance in type 2 diabetes and increased temperatures (ER stress) [4, 13, 45, 46]; malnutrition; anorexia; and relative energy deficiency in sports (nutrient deprivation) [3], alcohol, tobacco and aging (oxidative stress) [5–8], and viral infection [10–12, 47]; yet, the impact of these factors on the ISR in ovaries has been largely unexplored.

The ISR and associated stress response pathways function in part through ribonucleoprotein (RNP) granules, membraneless organelles that form via liquid–liquid phase separation of RNA-binding proteins and their associated RNAs [48]. While the roles of RNP granules in oogenesis and development have been extensively studied, models to explain how both external and cellular stressors impact these roles are lacking. In most somatic cells, the ISR is mediated by transient RNP granules called stress granules (SGs), which facilitate translational repression and mRNA sequestration during stress [49]. However, to date, canonical SGs have been largely unobserved in oocytes, which begs the question: how do oocytes respond to, and combat, stress? Understanding these molecular responses is critical for addressing the top non-genetic causes of reduced fecundity across species and will further clarify our understanding of the mechanisms that limit successful fertility treatments today.

This review synthesizes results from molecular and cellular studies across species—with a primary focus on Drosophila and mice—and expands on current models to explain how cellular and environmental stress can impact oogenesis by regulating a variety of RNA granule dynamics and functions. We focus on four classes of RNP granules, all of which have been shown to change in size, abundance, and/or function in response to environmental stressors: processing bodies (P-bodies), which participate in mRNA decay, transport, and translational repression; nuage, which regulates PIWI-interacting RNA (piRNA) and micro RNA (miRNA) biogenesis; U bodies (with accompanying Cajal bodies and nuclear speckles), which regulate spliceosomal maturation; and SGs, which inhibit translation and regulate mRNA stability in response to stress. We detail their RNA and protein compositions, known roles in development, and how stress impacts those functions to regulate or halt oogenesis. Next, we explore priming roles of active stress pathways and RNP granules in stem cell populations and identify likely parallels to cellular events in oocytes. Finally, we end by discussing the implications of RNP dynamics in fertility treatments. Throughout, we revisit foundational studies on stress responses in ovaries and reframe them in the context of more recent studies on oocyte RNP dynamics in development. We conclude by offering five key research questions that could serve to motivate future research in the field. Our goal is to rekindle a timely interest in investigating stress response pathways in oocytes and further our understanding of the role of RNP granules in fecundity and fertility.

Role of ribonucleoprotein granules in stress and development

Despite evidence of an active ISR in oocytes across species, the presence of classic SGs has been largely unobserved in this cell type. In contrast, multiple other RNP granule types have been identified in oocytes across species, and many have been shown to have unique responses to stress, including changes in size, abundance, and/or function. Thus, we argue that most oocyte RNP granules can be considered stress-responsive RNPs. The following sections detail what is known about stress-responsive RNP granule composition in Drosophila and mammals, roles in both oogenesis and early development, and how stress mediates their dynamics and function in these processes.

P-bodies

P-body components and functions in oogenesis

P-bodies are cytoplasmic RNP granules with proposed roles in mRNA translational repression, storage, transport, and decay [50–53]. P-bodies are distinct from SGs in that they are regular fixtures in developing oocytes; however, their composition, size, and location change in response to both developmental and stress events. They are one of the more well-studied RNP granules in ovaries, and their composition and roles during development are well documented across species. P-body assembly is initiated by RNA helicase, DDX6/Me31B, and eIF4-binding protein, eIF4-T/Cup, which sequester translationally repressed mRNAs through the 5′ cap-binding protein eIF4E/eIF4E1 [54–57]. In Drosophila germline cells and mouse oocytes, these initial aggregates primarily serve to transport and translationally repress mRNAs but can also recruit mRNA decay proteins, including decapping enzymes (Dcp1/dDcp1 and Dcp2/dDcp2), exonucleases (Xrn1/Pacman and Xrn2/Rat1), and RNA-binding proteins (DHX9/Mle and HNRNPU/Hnrnp-k) [53, 58, 59–63].

During mammalian oogenesis, P-body formation is dynamic and both their formation and dissolution are required to progress to later stages in oogenesis (Figure 1). In mouse ovaries, P-bodies form in the oocyte prior to primordial follicle formation and are required for maintaining primordial follicle quiescence during oocyte maturation [64, 65]. Loss of DDX6 prevents P-body formation and results in abnormally large primordial follicles that fail to progress and are eventually lost, leading to infertility [64, 65]. Similarly, loss of the 4E-BP, eIF4ENIF1, in humans, can lead to POI [66, 67]. In later stages of oogenesis, P-bodies transport and translationally repress maternal mRNAs into oocytes prior to fertilization. In mature oocytes, DDX6, along with YBX2 and CPEB, form subcortical aggregates with maternal mRNAs that dissolve at the start of metaphase II to allow for their translation (Figure 1) [68]. Other mRNAs, like the one that encodes the luteinizing hormone receptor, are degraded in P-bodies of mature follicles to help promote ovulation [69]. P-bodies therefore serve as essential regulators of development throughout oogenesis in mice.

In Drosophila ovaries, P-bodies are similarly critical for oocyte maturation, and P-body proteins are active prior to meiosis. For example, Cup is required for meiotic chromosome segregation. Mutant Cup ovaries display aberrant chromosome segregation and arrested egg chambers [70]. Later in oogenesis, Drosophila P-bodies translationally repress, store, and transport mRNAs from nurse cells into the developing oocyte, and multiple proteins found in P-bodies have been linked to regulation of key maternal mRNAs oskar (osk), bicoid (bcd), nanos (nos), and gurken (grk) [62, 71–76]. Small P-bodies in nurse cells travel down microtubules into the mature oocyte where they take on a heterogeneous foci pattern within the cytoplasm. Similar to mice, Me31B foci in mature Drosophila oocytes rapidly disperse upon egg activation, which correlates with translation of several axis-patterning maternal mRNAs, including bcd [59, 77]. Thus, P-body proteins ensure proper transport and spatiotemporal expression of key oogenesis maturation genes during oogenesis and maternal effect genes during early development [56, 78].

P-body formation and activity are regulated by nutrient-sensing mechanistic targeting of rapamycin/targeting of rapamycin

P-bodies have consistently been shown to be stress-responsive granules in ovaries. P-body formation and localization are mediated primarily by the mechanistic targeting of rapamycin/targeting of rapamycin (mTOR/TOR) nutrient stress response pathway. P-body formation in Drosophila is dependent on an inactive TOR pathway specifically in the follicle cells [79]. When the TOR pathway is inhibited by starvation, rapamycin treatment, or overexpression of the TOR antagonist Tuberous Sclerosis 1 in follicle cells, P-bodies increase in size and number in Drosophila germ cells [79]. Conversely, knocking down TOR in germ cells specifically does not induce P-body aggregation, suggesting follicle cells, but not oocytes, are primarily involved in sensing nutrient deprivation [79]. Across species, the formation and dissolution of P-bodies during oogenesis are anticorrelated with mTOR/TOR, eIF4E, and eIF4E1B activity, which all transiently decrease prior to the onset of meiosis, then gradually increase throughout oocyte maturation [80–84]. While the direct mechanism between mTOR/TOR activity and P-body size has not been definitively shown, it is known that mTOR/TOR inactivation allows unphosphorylated 4E-BPs to bind to eIF4E and prevent reassociation with eIF4G [57] and that 4E-T binding to eIF4E in HeLa cells occurs in P-bodies, increases P-body size, and suppresses translation [78]. Thus, it can be inferred that during mTOR/TOR inactivation, 4EBPs bind and sequester eIF4E-bound mRNAs into P-bodies. When mTOR/TOR is active, 4EBPs are phosphorylated and unable to compete with eIF4G binding to eIF4E, allowing translation to resume.

P-body response to nutrient deprivation stalls oogenesis

Both nutrient and thermal stress inhibit mTOR/TOR and consequently all cap-dependent translation [85–87]. In Drosophila ovaries, dDcp1-positive P-bodies exist primarily in nurse cells under basal conditions but dramatically increase in size and associate around nurse cell nuclei upon nutrient deprivation and thermal stress (Figure 2A and B) [59, 63, 79, 88]. This enlarged state is due to stalled P-body transport from nurse cells to oocytes, increased sequestration of eIF4E1-bound mRNAs, and recruitment of decapping enzymes [79, 88]. Nutrient deprivation also initiates two apoptotic events in Drosophila ovaries—one within the germarium 2a/2b region, at the onset of meiosis, and the other in stage 8 egg chambers, at the onset of vitellogenesis—but maintains stages 1–7 between these two checkpoints [23]. These events are largely regulated by P-body dynamics and function. Drosophila loss of function Tor mutant ovaries have premature meiotic entry and germline-derived cells within egg chambers undergo autophagy that results in infertility [80]. Nutrient deprivation similarly inhibits TOR activity, and thus, we hypothesize nutrient deprivation–induced apoptosis of germline-derived cells within the 2a/2b region of the Drosophila germarium is likely due in part to premature P-body formation and subsequent early meiotic entry of germline cells.

This is consistent with observations in mouse ovaries showing that P-body formation is essential for initiating meiosis (Figure 1) [64, 65]. While inhibiting mTOR is necessary for P-body formation and meiosis to occur, premature meiotic entry due to excessive mTOR inhibition can trigger apoptosis [64, 65]. Similar observations were recently made in mouse ovaries exposed to the widely used type II pyrethroid pesticide, fenvalerate (FEN), that activates the ISR [89, 90]. Fenvalerate disrupts follicle growth by inhibiting primordial follicle formation and induces a primary ovarian insufficiency–like phenotype in mice. Fenvalerate enhances P-body formation indirectly by elevating m6A-mRNA levels and their reader, YTHDF2, which recruits P-body proteins (e.g., DDX6) and stress granules protein G3BP1 to m6A modified RNAs [89]. Similar to TOR inhibition, FEN exposure could either promote premature P-body formation and/or sequester P-bodies from their usual targets and disrupt normal oogenesis.

In Drosophila, apoptosis of stage 8 oocytes, which occurs immediately prior to vitellogenesis, ensures oocytes do not mature during conditions where yolk production is inadequate (e.g., nutrient deprivation) (Figure 2A and B) [23]. P-bodies facilitate yolk production. Within stage 6–7 egg chambers, P-body proteins normally package and transport osk mRNA down microtubules from nurse cells to the posterior pole of the oocyte [88]. dDcp2 further regulates actin filament assembly to correctly position osk mRNA in the posterior pole [91]. However, during nutrient deprivation, osk mRNA localization in stage 6–7 egg chambers shifts from the posterior pole of the oocyte to P-bodies within the nurse cell. This relocalization is due to the dramatic reorganization of microtubules and serves to regulate Osk protein activity [79, 88]. As Osk protein helps promote endocytosis of yolk from somatic cells into the oocyte, relocalizing P-bodies prevents osk mRNA from translocating into the oocyte and being translated [92–94]. Thus, stalled P-body transport ensures eggs will not progress to stage 8 and prematurely promote endocytosis of yolk until nutrients are available again.

During prolonged stress, inhibition of P-body functions is reinforced by additional stress response pathways. In both mouse embryonic fibroblasts and Drosophila S2 cells, the mTOR/TOR pathway can be inhibited by ISR activation through GCN2/Gcn2, which utilizes ATF4/Atf4 to upregulate 4E-BPs and mTORC1/Tor inhibitory genes [95–97]. If stress cannot be mitigated, both the ISR and mTOR/TOR communicate with apoptotic pathways to eliminate the damaged cell and autophagy pathways to recycle cellular components for energy maintenance [36, 38]. Whether similar mechanisms occur in ovaries has yet to be explored.

Nuage and PIWI ribonucleoproteins

PIWI-interacting protein components and functions in oogenesis

PIWI proteins are implicated in at least three different roles in the developing oocyte: piRNA biogenesis and retrotransposon silencing in nuage, retrotransposon silencing and mRNA regulation in P-bodies, and germ cell identity maintenance in germ granules (Polar Granules in Drosophila) [98–102]. In Drosophila, nurse cell perinuclear nuage contains both P-body proteins and PIWI-interacting proteins including the piRNA-processing proteins, Aubergine (Aub) and Argonate 3 (Ago3), RNA helicases, Vasa (Vas) and Spindle-E (Spn-E), multiple tudor domain–containing proteins, Tudor (Tud), Krimper (Krimp), Tejas (Tej), and Vrenteno (Vret), and Maelstrom (Mael) [98, 99]. Here, the proposed main function of PIWI-interacting proteins is to generate piRNAs required to block transposable element (TE) retrotransposition [100, 103–105]. In P-bodies, a subset of PIWI-interacting proteins, including Aub, Ago3, and Krimp, facilitates retrotransposon mRNA degradation and Aub is specifically implicated in osk mRNA translation in the oocyte [102, 106]. In germ granules, another subset of PIWI-interacting proteins including Aub, Tud, and Vas is critical for specifying and maintaining germ cell identity [103, 107–110]. The ability of these PIWI proteins to carry out various gene expression regulation activities is dependent on piRNAs. PIWI-interacting RNAs are small (23–29 nt) non-coding RNAs with varying degrees of complementarity to TEs, mRNAs, and cellular mRNAs that are produced primarily in germ line cells [111, 112]. PIWI-interacting RNAs act as guides and recruit PIWI proteins Aub and Ago3 to complementary mRNAs to initiate mRNA degradation and translational control [102]. In the case of TEs, piRNAs can also employ several other mechanisms to inhibit retrotransposition, including disrupting splicing of TE mRNAs [102, 112].

Drosophila PIWI mutants display defects in oogenesis including patterning defects, abnormalities in oocyte polarity, and sterility [105, 113, 114]. Sterility phenotypes are often due to an inability of mutant PIWI proteins to effectively hinder retrotransposition, as observed in cases of hybrid dysgenesis [108, 112]. Additional phenotypes are due to functional loss of these components within P-bodies and germ granules (polar granules in Drosophila), where piRNAs and Aub assist P-body constituents to facilitate maternal mRNA translation, degradation, or localization [103, 106, 107, 115]. PIWI-interacting RNAs can bind with maternal mRNAs with reduced specificity and recruit Aub to degrade them. Alternatively, Aub can also promote mRNA translation, when it interacts with factors that enhance mRNA stability and translation, including poly(A)-binding protein (PABP), eIF3, eIF4E, and Wispy [103, 107, 115]. Toward the end of Drosophila oogenesis, piRNAs bound to PIWI proteins are stored in mature oocytes, where they govern maternal mRNA localization and stability during embryogenesis [114–116]. In early embryos, a piRNA “ping-pong cycle” utilizes Aub and Ago3, where Aub-mediated mRNA degradation generates fresh piRNAs employed by Ago3 for further mRNA targeting. The lack of maternally inherited ping-pong amplification in germ cells leads to infertile offspring [108]. These examples are likely just the tip of the iceberg of PIWI protein and piRNA-mediated mRNA regulatory activities in Drosophila.

In mammals (including humans, mice, macaques, pigs, and cows), PIWI proteins have been primarily implicated in piRNA biogenesis [100, 101, 117]. The majority of PIWI mammalian orthologues, including PIWIL1/MIWI, PIWIL2/MILI and PIWIL3 (Aub orthologues), VASA/DDX4/MVH, MAEL, TDRD1, TDRD7, TDRD8, and TDRD9 are expressed in developing mammalian oocytes starting in the primordial follicle stage [100, 101, 117]. It is during this primordial follicle stage that mouse oocyte nuage containing multiple PIWI proteins has been observed [100]. In mouse oocytes, loss of Mili disrupts nuage formation, while loss of Mvh, Mili, Miwi, Gasz, or Pld6 disrupts piRNA biogenesis and repression of some retrotransposons [100, 118]. The primary activities of the nuage appear restricted to the primordial follicle stage, as nuage dissipates by the primary follicle stage. However, as follicles mature, piRNA biogenesis increases suggesting these proteins are still active when diffused in the cytoplasm [101, 117].

An activated integrated stress response disrupts transposable element suppression

Most TEs, namely, retrotransposons, become active after germ cell differentiation (e.g., in nurse cells within Drosophila ovaries), though some transposons (e.g., P-elements) can become active in undifferentiated germ cells [112, 119]. In Drosophila ovaries, retrotransposons are made primarily in nurse cells but exclusively infiltrate the Drosophila oocyte nucleus. Depletion of Aub and Ago3 exacerbates this event but does not lead to TE insertion into nurse cell nuclei. Thus, PIWI proteins within the nurse cell likely serve to block retrotransposon mobilization from nurse cells to the oocyte [119].

During ISR activation induced by heat shock and oxidative stress, TEs exhibit an increase in both number and activity [120–122]. Work in multiple cell types has found that TEs are able to do this by disrupting stress response pathways themselves or by bypassing compromised PIWI machinery. Some TE proteins can disrupt translational repression locally within SGs. One particular TE protein, ORF1p, is expressed in mouse spermatocytes and interacts with AGO3 and numerous SG factors, including G3BP2, FMR1, PABPC1, and CAPRIN1, to facilitate the translation of TE transcripts within SGs [123, 124]. ORF1p can also bind to the PKR regulator, PACT, resulting in reduced eIF2alpha phosphorylation and heightened TE retrotransposition [123]. Whether similar interactions occur within ovaries remains to be established. In other scenarios, PIWI machinery that normally prevents retrotransposition can become compromised during ISR activation. In human retinal pigment cells, PIWIL4, a protein responsible for piRNA-mediated DNA methylation of TEs, relocates to the nucleus in response to acute oxidative stress, aiding in the suppression of Alu mRNAs [125–127]. However, under chronic oxidative stress, PIWIL4 becomes sequestered within SGs, leading to enhanced Alu element retrotransposition due to inefficient targeting [127]. It remains to be seen whether other PIWIL proteins respond to stress similarly on oocytes.

Stress impacts on PIWI-associated chaperones, however, have been documented in Drosophila ovaries. (Figure 2C and D). Multiple chaperones, including Hsp83/Hsp90, Hop (hsp70/90-organizing protein), and Hsc70-4, play crucial roles in loading piRNAs onto Ago3 and aiding in the degradation of TEs [128, 129]. Their necessity in this process is fundamental for oocyte maturation and viability. Drosophila ovaries with mutant hsc70-4 and hop exhibit non-viable eggs and halted oogenesis even in the absence of stress [128, 130]. In contrast, the loss of heat shock response proteins Hsp70, Hsp83, and Hsf in ovaries only leads to oogenesis defects in the presence of stress [128, 131]. A curious unintended outcome of the heat shock response in oocytes is that numerous PIWI-associated chaperones become sequestered within Hsp70 cytoplasmic granules, away from the perinuclear nuage (Figure 2C and D). While this aids in facilitating general protein refolding during ER stress, it concurrently debilitates the PIWI complex, so it can no longer effectively silence TEs [128].

While increased piRNA expression during stress has been observed, this does not appear to be sufficient to combat against the transposition-favoring events above. In Drosophila ovaries, some dormant piRNA clusters can become transcriptionally active at higher temperatures to inhibit specific transposons; however, this process occurs at the epigenetic level and is relatively rare, occurring in only ~2% per generation [132]. Thus, TEs appear to win out under most stress conditions. Given the prominent presence of TEs in ovaries, understanding how they evade PIWI proteins and how stress exacerbates these events will be critical for assessing the impact of TE contributions to female fecundity during stress.

Stress impacts on PIWI complex–mediated mRNA regulation

There are limited data on how stress impacts the mRNA regulatory activities of PIWI proteins within P-bodies and germ granules. In Drosophila, elevated temperatures have been shown to downregulate expression of factors required for piRNA biogenesis in somatic ovarian cells, including Yb, Shu, and Hsp83. However, Aub and Ago3 levels increase and likely contribute to an enhanced ping-pong amplification cycle [122]. How this might impact oogenesis or early embryonic development is not clear.

Much of what is known about stress and PIWI protein–mediated mRNA regulation comes from other cell systems. In mouse testes, elevated temperatures lead to increased expression of a particular piRNA, piR-020492, predicted to target mRNAs involved in AMPK and insulin pathways [133]. Overexpressing this piRNA prevents germ stem cell proliferation and induces apoptosis. Whether this directly regulates translation of these genes within P-bodies or whether gene upregulation occurs in response to DNA damage from TEs in general is not known. Similarly, in bronchial epithelial cells, piRNA expression increases with ER stress, and piRNAs and associated PIWI proteins are required to induce apoptosis following unsuccessful unfolded protein response (UPR) in these cells [134]. A greater understanding of how heat impacts PIWI complex activity and piRNA processing of mRNAs in ovaries will clarify the extent to which these pathways contribute to, or impact, healthy oogenesis (Box 1).

U bodies, Cajal bodies, and nuclear speckles

U body, Cajal body, and nuclear speckle components and functions during oogenesis.

U small nuclear RNP (snRNP)-bodies (U bodies), Cajal bodies, and nuclear speckles assemble snRNPs and facilitate pre-mRNA splicing, a critical step in post-transcriptional RNA processing [50, 51, 135–137]. In Drosophila, cytoplasmic U bodies function as sites of snRNP assembly and storage prior to nuclear import [137]. In Drosophila oocytes, U bodies are comprised, namely, of spliceosome assembly proteins (survival of motor neurons [SMN] and Gemins), spliceosomal (Sm) proteins, and small nuclear RNAs (snRNAs): U1, U2, U4, and U5 [137]. Fully assembled snRNP complexes translocate to the nucleus into Cajal bodies to undergo snRNA modification and subsequent reassembly following splicing [138, 139]. This complex then merges with specific serine/arginine-richsplicing factors (SRSFs), and long non-coding RNA, MALAT1 (identified in human cells), in nuclear speckles where proteins are further modified and perform pre-mRNA splicing [140, 141].

Many U body components have critical roles in oogenesis and development. Gemin3, a DEAD-box RNA helicase involved in snRNP assembly and maintenance of both U bodies and Cajal bodies, is required for proper oocyte maturation and nurse cell endoreplication during oogenesis in Drosophila [136]. Mutating Gemin3/mel-46 results in aberrant ovarian morphology in mice, Caenorhabditis elegans, and Drosophila [136, 142, 143]. Gemin3 also interacts with the RNA-induced silencing complex (RISC), involved with miRNA processing and RNA degradation [144]. survival of motor neurons, a key player in snRNP biogenesis, is expressed in all stages of oogenesis and is critical for oocyte maturation and embryonic development. In Drosophila, SMN is diffuse in the cytoplasm of early staged egg chambers and becomes more concentrated in U bodies in stage 6 egg chambers that persist through stage 10 (Figure 1) [135]. Similarly, Cajal bodies are visible in nurse cells around stage 6 and fragment into multiple granules within each nucleus between stages 8 and 10, prior to disappearing [145]. Together, this indicates an increased requirement for assembled spliceosomes later in oogenesis. Like Gemin3 mutants, SMN mutants are unable to form Cajal bodies and have abnormally large P-bodies localized to the perinucleus. In Drosophila, SMN mutants also have similar defects in nurse cell endoreplication [145].

Cajal and U body response to stress alters splicing during oogenesis

Nutritional stress in Drosophila induces SMN-positive U bodies in early staged egg chambers (stages 1–7) and enlarges SMN-positive U bodies in mid-staged egg chambers (stages 8–10) (Figure 2E and F). Prolonged starvation causes Drosophila ovaries to shrink in size, primarily due to a loss of mid-stage oocytes (see P-body Response to NutrientDeprivationStalls Oogenesis). Mid-stage oocytes that persist often display abnormal fusion events and have large U bodies, suggesting that U body-mediated RNA processing events are particularly sensitive to nutrient stress at this stage of oogenesis [135]. Cajal bodies in mid-staged egg chamber nurse cells also increase in both size and number with nutrient stress (Figure 2E and F) [146]. Enlarged U bodies and Cajal bodies could reflect a block in spliceosomal assembly and modification and thus serve as a means to halt pre-mRNA processing and mRNA export. Indeed, ISR activation in HeLa cells has been shown to inhibit spliceosome nuclear import by sequestering nuclear import factor importin-β1 into SGs [147]. However, in this study, both U bodies and Cajal bodies decreased in size, suggesting that ovaries may have alternative mechanisms in place to regulate these RNP functions during stress.

An intriguing explanation for U body enlargement in ovaries during stress involves the role of cytoskeletal-mediated “cytoplasmic stirring.” During oogenesis, a dynamic cytoskeleton creates cytoplasmic stirring that promotes nuclear RNP granule condensation in the oocyte [140, 148]. In mice, both Cajal bodies and nuclear speckles increase in size and decrease in number as an oocyte grows and the ability to form larger aggregates depends on actin assembly within the cytoplasm. The stirring forces applied by actin cytoskeleton assembly in mouse oocytes stabilize spliceosome complexes within nuclear speckles and Cajal bodies, and this process is required to complete meiosis. Disrupting actin assembly dramatically alters splicing patterns and oocytes fail to undergo the first meiotic division [140]. In Drosophila, microtubules primarily facilitate oocyte cytoplasmic stirring, while actin acts more as a shield. Inhibiting actin assembly in this model promotes cytoplasmic stirring, which enhances oocyte nuclear speckle condensation and accelerates oocyte maturation (Figure 2E-F) [140]. Microtubules are also very dynamic in nurse cells starting at stage 6 and move rapidly through ring canals from nurse cells to oocytes to facilitate transport of mRNAs and proteins [149]. Here, microtubules do not participate in stirring but rather serve to transport RNP granules from nurse cells into oocytes [150]. Thus, nurse cell Cajal and U bodies are relatively small in unstressed cells. Nutrient deprivation in Drosophila dramatically reorganizes microtubules within both nurse cells and oocytes, primarily repositioning them along the cortex of egg chambers (Figure 2F) [88]. Stalled microtubule dynamics impede cargo flow from nurse cells to oocytes and disrupt cytoplasmic stirring in Drosophila oocytes. The former creates enlarged Cajal and U bodies within nurse cells due to reduced RNP dynamics, while the latter reduces Cajal and nuclear speckle size within the oocyte (Figure 2F) [140, 146]. Both events can inhibit U body translocation to the nucleus, and nuclear speckle and Cajal body processes. These events, in turn, could alter splicing patterns in both nurse cells and oocytes.

This model is supported by observations that multiple ISR stressors alter alternative splicing in other cell types. Nutrient deprivation decreases splicing of ribosomal genes in yeast, while sodium arsenite (oxidative stress) and thapsigargin (ER stress) disrupt alternative splicing in HeLa cells [147, 151]. Studies in Tibetan sheep give us some insight into the impact of ISR activation on splicing in ovaries. Tibetan sheep live at high altitudes where they are subjected to chronic oxidative stress and have reduced fertility compared to those raised at lower altitudes [152]. One of the most striking differences between these two populations of sheep is extensive changes in alternative splicing patterns in ovaries, characterized primarily as intron-retaining events. These events disrupt splicing of transcription factors, RNA-processing enzymes, and even splicing factors themselves [152]. Together, these observations suggest stress in ovaries impacts cytoskeletal movement and oocyte cytoplasmic stirring. This, in turn, impacts nuclear import of splicing factors and nuclear RNP granule assemblies, both of which disrupt splicing. More direct experiments to determine whether environmental stressors do indeed alter splicing by disrupting cytoskeletal dynamics and RNP formation would solidify this model and provide insight into the impact these alternative splicing events have on oogenesis (Box 1).

U body interaction with piRNAs

Drosophila U bodies also exhibit a response to heat shock, by increasing in size in mid-stage egg chambers. However, in early egg chambers, the U body protein SMN localizes within the perinuclear nuage [111, 135]. Recently, piRNAs were found to promote aberrant splicing of P-element mRNAs as a means to render them non-transposable [112]. While this process does not appear to be temperature dependent, it suggests collaboration between U bodies and piRNA processing enzymes. As heat exacerbates TE abundance, relocating U bodies to perinuclear nuage may act as a protective mechanism to ensure P-elements are spliced to prevent transposition.

Stress granules

Stress granule components and functions in oogenesis

Stress granules (SGs) are dynamic RNA–protein complexes that are directly induced by and mediate the ISR. Stress granules form subsequent to eIF2alpha phosphorylation and ribosomal runoff and, in mammals, are composed of translationally stalled mRNAs, 40s ribosomal components, translation initiation factors eIF3, eIF4A, eIF4B, eIF4G, and eIF4E, and RNA-binding proteins (RBPs) including G3BP1, G3BP2, FMRP, PABP, ATXN2, TIA-I, UBAP2L, and Caprin [49, 153–156]. Stress granules serve as sites for mRNA sequestration, sorting, and localization during cellular stress and allow cells to prioritize essential transcripts and protect them from degradation. Stress granule formation is a conserved process, with Drosophila counterparts including Rox8 (TIAR homologue), eIF3, eIF4E, Rasputin (Rin) (G3BP1 homologue), dFMRP, and dPABP, readily assembling SGs in response to sodium arsenite and heat shock in S2 cells [43, 157, 158].

Detection of bonafide SGs in oocytes has been limited primarily to Drosophila and C. elegans. In C. elegans germ cells, SG proteins have been identified in “P granules,” RNP granules that localize to the nuclear periphery, which comprise proteins found in SGs, P-bodies, U bodies, and perinuclear nuage in other species [159]. TIAR-1 forms granules induced by both heat and starvation in C. elegans oocytes, which co-localize with the P-body component CGH-1/DDX6 [160]. Likewise, other SG proteins, TIA-1 and PABP, have been observed to interact with C. elegans P granules [161]. In Drosophila, the SG proteins dFMRP, PABP, and eIF4E all exhibit localization to SGs induced by arsenite and heat in somatic follicle cells [157]. While SGs have been identified in spermatogonia, it remains uncertain whether SGs serve as distinct RNP granules in female germ cells or whether various SG proteins simply coalesce with other RNP bodies [162].

Despite the lack of evidence for SGs in oocytes across species, several pivotal SG proteins exhibit significant enrichment in germ line cells and play critical roles in oogenesis and early development. During human fetal development, one of the most extensively studied SG proteins, FMRP, is exclusively expressed in germ line cells [163]. While FMRP is primarily recognized as a translational repressor, it can also act as a translational activator and may further have roles in transcriptional regulation, chromatin modification, RNA modification, and RNA editing [164]. In Drosophila, dFMRP is required for germ stem cell maintenance and proliferation and, together with Caprin, regulates the maturation of follicle stem cells [165–167]. Both FMRP/dFMRP and Caprin are essential for proper oogenesis. GGC nucleotide repeat expansions within the 5’UTR of FMR1 reduce FMRP protein levels and can lead to Fragile-X associated premature ovarian insufficiency [168, 169]. Moreover, the loss of Caprin in Drosophila ovaries prolongs mitosis in follicle stem cells, diminishing their overall count and resulting in early defects in egg chambers [166]. Similarly, other SG proteins are critical for oogenesis. Loss of poly(A)-binding protein nuclear 1-like (PABPN1L), a protein involved in maternal mRNA decay, causes sterility in mice [170]. Loss of the SG nucleator, rin, causes female sterility in Drosophila and results in faulty egg chambers marked by chromatin compaction and actin cytoskeleton abnormalities [153, 171–174]. In C. elegans, TIAR-1 mutants exhibit modest infertility and display meiotic defects during oogenesis [160]. Are these phenotypes due to disruptions in SG formation, general loss of RNA processing and metabolism, or both?

Role of the integrated stress response and stress granules in stem cell maintenance and priming

Numerous SG proteins localize within cytoplasmic foci during meiosis. In humans, FMRP forms cytoplasmic granules alongside G3BP and the P-body protein GW182 during the transition from primordial germ cells to meiotic germ cells (Figure 1) [163]. Similarly, cytoplasmic granules containing SG proteins PAB-1 (PABP) and TIA-1 have been observed in meiotically arrested oocytes in C. elegans [161].

What role do these granules play in oogenesis? Numerous RNA-binding proteins, such as Rin, Gemin3, Smn, and Cup, play pivotal roles in preserving the genomic integrity of oocytes [70, 145, 171, 175]. Oocytes are especially susceptible to DNA damage due to their prolonged lifespan, leading to the potential accumulation of damage over time [176]. However, recent studies suggest that unlike in mitosis, DNA damage repair processes are active in meiotic cells, enabling the successful repair of damaged DNA in both meiosis I and meiosis II oocytes [177–179]. One hypothesis is that meiotic stress-like granules may act as sensors of DNA damage. This is consistent with observations that low level ISR activation slows cellular metabolism, decreases the generation of stress-inducing molecules such as ROS, and provides protection against DNA damage [180, 181]. Integrated stress response activation may further contribute to overall oocyte growth and maintenance over time. mRNA transcripts with upstream open reading frames (uORFs)—translated specifically during ISR activation—are disproportionately transcribed in granulosa cells during follicle development [41]. These uORF-containing genes assist in follicle growth and maintenance, by allowing follicles to overcome cellular damage. Thus, these meiotic granules could assist the ISR in repairing oocytes and promoting further oocyte maturation.

A second (not mutually exclusive) hypothesis is that these granules, with assistance from the ISR, help maintain oocytes in a totipotent, undifferentiated state and prime them for cellular differentiation. Indeed, much like in oocytes, eIF2alpha is phosphorylated and global translation is significantly reduced in other stem cell populations [14, 42, 182–186]. Drosophila intestinal stem cells exhibit high expression of SG proteins, forming compact RNP granules that readily transform into larger SGs following arsenite and rapamycin treatment. In contrast, surrounding terminally differentiated cells have low to undetectable levels of SG proteins [180]. These progenitor populations also harbor P-bodies, which merge with SGs upon induction of stress, playing a vital role in maintaining progenitor cells within a stem cell state [180]. Similar phenomena have been observed in quiescent stem cell populations. Mouse skeletal muscle stem cells exhibit elevated levels of p-PERK, p-eIF2alpha, and cytoplasmic granules enriched with both SG and P-body proteins [14, 187]. Integrated stress response activation in these cells triggers translation of genes that promote stemness via non-canonical initiation at uORFs, while simultaneously inhibiting the canonical translation of genes required for muscle differentiation through mRNA sequestration into cytoplasmic granules [14, 187]. Dephosphorylation of eIF2alpha in these cells causes them to differentiate and lose these cytoplasmic granules [14].

P-bodies and the mTOR pathway, which also serves as a stress-responsive translation regulator, have similarly been linked to self-renewal in quiescent stem cells [188]. In adult neuronal stem cells, activation of mTOR, which promotes cap-dependent translation, allows cells to differentiate [189]. In mouse ovaries, P-body protein ELAVL2 promotes quiescence in primordial follicles [64]. In contrast, RNP protein DDX6, found in both P-bodies and stress granules, appears to promote differentiation of neural progenitor cells into neurons, while loss of DDX6 maintains stemness in this cell population [190]. The role of DDX6 and P-bodies in regulating pluripotency is context dependent: DDX6 promotes differentiation of human pluripotent stem cells, intestinal stem cells, and neuronal progenitors but promotes pluripotency and stemness in mesodermal lineages [190]. This dual function is driven in part by P-body translational regulation genes encoding pluripotency driving transcription factors including OCT4 and NANOG, and chromatin regulators including histone methyltransferase KDM4B, highlighting a unique role for P-bodies in regulating chromatin remodeling events during cellular differentiation [190]. Similar events were recently observed in mouse ovaries where the loss of P-body protein LSM14B prevented proper oocyte maturation [191]. In the absence of LSM14B, DDX6 and 4E-T are mislocalized to the nucleus and depleted in the cytoplasm resulting in enhanced translation of KDM4B, which alters global transcription in follicles [191]. How stress impacts these activities has not been investigated. Collectively, these observations support a fundamental role for stress response pathways in regulating oogenesis and preserving stemness. Further defining the contributions of the ISR to oogenesis will be critical for predicting the impact that environmental stressors will have on development (Box 1).

Importance of considering stress-responsive ribonucleoproteins in fertility treatments

Understanding the roles of stress pathways and RNP contributions to oogenesis is similarly critical for improving fertility treatments. Some commonly used fertility treatments themselves can either inhibit or induce stress response pathways, but how these treatments impact RNP activities has been largely ignored. Given the relatively low success of fertility treatments, we argue these impacts should be more formally investigated. For instance, metformin has been used in numerous clinical trials to assist in in vitro fertilization (IVF) treatment of individuals with PCOS. Metformin is typically used to treat type 2 diabetes mellitus, a common comorbidity in PCOS. A popular hypothesis is that lowering insulin levels prior to egg retrieval could assist IVF of women with PCOS [192–194]. However, these studies ignore the fact that metformin indirectly inhibits mTOR [195] and consequently should stabilize P-bodies in ovaries and prevent their dissolution required to progress through the final stages of oogenesis. At the cellular level, infertility in PCOS is caused by a failure of follicles to mature past the early antral stage [196]. Inhibiting mTOR via metformin would exacerbate this issue. Not surprisingly, despite multiple clinical trials, the effect of metformin on improving IVF and live births has been negligible or inconclusive for individuals with PCOS [192–194].

Metformin has been successful in scenarios where the underlying cause is accelerated oogenesis. Broiler hens have increased numbers of follicle cells and double ovulation that leads to inefficient egg laying over time [197, 198]. These events could be explained by premature P-body dissolution, which would promote early oocyte maturation. Feeding hens metformin leads to more stable egg laying throughout their lifetime, resulting in higher yields and more viable eggs [198]. While the focus of this study was on hormone-level alterations, an alternative model is that metformin enhances P-body formation in early oocytes and slows egg production. This is observed as a more linear and compact ovarian follicular hierarchy. Consistent with this, metformin has been shown to assist with ovarian hyperstimulation syndrome induced during IVF [193]. As P-bodies are largely responsible for timing and location of gene expression events within the developing oocyte, understanding how fertility drugs impact P-body function will likely aid in more effective treatments (Box 1).

Gonadotropins are another class of commonly used fertility drugs that speed up egg development and ovulation. Recently, however, they have also been shown to decrease levels of piRNA biogenesis genes Tdrd1, Tdrd9, and Mael in mice [199]. This is potentially problematic, as it has been shown in Drosophila, that one of the ways in which oogenesis combats TEs is by favoring germ stem cell proliferation over differentiation and diluting the pool of stem cells that contain deleterious TEs [200]. Gonadotropins tip the scale toward more differentiation and thus could be enhancing the potential of TEs while reducing the protective mechanisms to combat TEs. Overstimulation by gonadotropins can also activate the ISR. In mouse studies, ovaries treated with pregnant mare serum gonadotropin had elevated levels of GCN2-P in mature oocytes, while in humans, gonadotropin treatment can lead to ovarian hyperstimulation syndrome (OHSS) and is associated with ER stress in cumulus cells. Administration of the ER stress inhibitor tauroursodeoxycholic acid reduces OHSS in rat models by decreasing vascular endothelial growth factor A (VEGFA) levels artificially inflated by gonadotropins [42, 201]. In contrast, studies in mice with galactosemia-induced primary ovarian insufficiency show an opposite role for the ISR [44]. These mice exhibit a compromised ISR. When these mice are pharmacologically treated with an ISR activator, they have improved follicle counts and fertility. Together, these data suggest that the underlying cause of infertility and its relationship to the ISR should be considered during treatment. A clearer understanding of how stress contributes to different infertility conditions, and how fertility treatments themselves can impact stress events within ovaries, is necessary to improve best practices in this field (Box 1).

Five key questions for the future

Given the complex and interactive nature of stress-responsive RNPs on the short- and long-term development of oocytes across species, we now briefly outline five key research questions that could motivate future research in the field (Box 1).

Concluding remarks

By their very dynamic nature, RNPs are able to balance well-orchestrated gene regulation with quick responses and adjustments to various insults. This is precisely why stress impacts so many RNA processing events detailed above. Increasing evidence also suggests that RNP granules are highly communicative. P-bodies associate with SGs, U bodies, and PIWI proteins [103, 107, 108, 135, 159, 202] and likely facilitate the coordination of RNA metabolism and regulate gene expression within the cytoplasm. How stress pathways help promote or deter these events is still largely understudied. Among the factors that should motivate us to look further into these connections are the observations that: (1) our species is experiencing a gradual increase in average maternal age over the last few decades that correlates with an increased demand for fertility treatments [203], and (2) impacts of global climate change and loss of habitat are directly impacting fertility rates in pollinator species [204]. Given increasing evidence that stress pathways are active throughout oogenesis and that oocytes are subjected to a multitude of insults over their lifetime, exploring how these factors influence this pivotal stage of development is both increasingly relevant and critical to combating threats to female infertility.

Author contributions

This review was initially conceived by MRG and then designed, written, and edited by MRG & CN. CN created all figures with input from MRG.

Conflict of interest: The authors have declared that no conflict of interest exists.

References