Bidirectional communication in oogenesis: A dynamic conversation in mice and Drosophila

Abstract

In most animals, the oocyte is the largest cell by volume. The oocyte undergoes a period of largescale growth during its development, prior to fertilization. At first glance, tissues that support the development of the oocyte in different organisms have diverse cellular characteristics that would seem to prohibit functional comparisons. However, these tissues often act with a common goal of establishing dynamic forms of two-way communication with the oocyte. We propose that this bidirectional communication between oocytes and support cells is a universal phenomenon that can be directly compared across species. Specifically, we highlight fruit fly and mouse oogenesis to demonstrate that similarities and differences in these systems should be used to inform and design future experiments in both models.

Oocyte and support cell communication

The ultimate goal of oogenesis is to create a developmentally competent female gamete that has all of the necessary biological products to undergo fertilization and initiate embryonic development prior to zygotic genome activation. An essential part of this preparation is the accumulation of select molecules that can only be obtained from the local environment of oocytes. This surrounding cellular environment is made up of support cells that synthesize and provide these necessary products to oocytes throughout their development.

Although the Drosophila oocyte was once thought to be a passive recipient of products from attached sister cells, called nurse cells, we now know that these two cell types engage in bidirectional communication similarly to mouse oocytes and their supporting granulosa cells. The embryological origins are different between Drosophila nurse cells and mouse granulosa cells because the nurse cells are germline cells whereas granulosa cells are not. The Drosophila equivalent of the granulosa cells, embryologically speaking, are the follicle cells. Nevertheless, we liken Drosophila nurse cells to mouse granulosa cells because the oocyte is already specified by the time it interacts with these cell types. Moreover, the nurse cells and granulosa cells have similar functions in supporting the oocyte throughout oogenesis while it increases in volume by orders of magnitude and is halted in prophase of meiosis I during this period of growth [1,2].

Experimental manipulation of the oocyte and its support cells reveals that growth coordination between these cell types is necessary for proper growth and maturation of not just the oocyte but also the support cells themselves. In Drosophila, genetic manipulation of growth regulators in the nurse cells causes concurrent delays in oocyte growth and specification [3]. When Drosophila oocyte specification is lost through genetic manipulation, the nurse cells and the entire egg chamber ultimately undergo apoptosis [4]. In the mouse, where direct tissue manipulation is more feasible, physical separation of the oocyte from the granulosa cells followed by independent culturing results in stalled growth of both cell types [5–7]. This coordination is facilitated by bidirectional communication and functions to balance the needs of oocytes with the biosynthetic capacity of their support cells at any given time. If more nutrients are available, the system can grow together faster as a whole [8,9].

Although examples of bidirectional communication between oocytes and support cells were first characterized in Drosophila over 20 years ago and in mouse over 40 years ago, the mechanisms by which bidirectional communication coordinates growth of the germ cell and its companion support cells are still being deciphered [10–14]. They are likely a multifactorial combination of cellular processes including paracrine signaling, metabolic control, and dynamics of the local mechanical environment, which together create feedback loops sensed by both the oocyte and the support cells. Here, we discuss what is currently known about bidirectional communication in the mouse and in Drosophila and how we can advance this knowledge to fuel further discovery regarding oogenesis in these essential model organisms.

Introduction to Mus musculus oogenesis

In the fetal ovary, the primordial germ cell pool expands through mitotic divisions with incomplete cytokinesis, forming germ cell cysts [15]. The germ cells within cysts then synchronously initiate meiosis in a radial wave followed by an anterior-posterior wave in response to retinoic acid [16–22]. Subsequently, the germ cells, now called oocytes, progress through meiosis and arrest at diplotene of prophase I. This cell cycle arrest is maintained until the time of ovulation, which does not begin until puberty [16]. Germ cell number peaks during mid-to-late gestation and then declines until birth and throughout the reproductive lifespan [23–25]. Right before birth, the germ cell cyst breaks down and while some oocytes will die, others will be enclosed by a complete or incomplete layer of squamous granulosa cells, forming primordial follicles and thus, the ovarian reserve (Figure 1A–C). Throughout the reproductive lifespan of mice, primordial follicles experience 3 fates: remain in a non-growing state, undergo atresia prior to or during folliculogenesis, or become activated and recruited into an approximately 3-week maturation period during which oogenesis and folliculogenesis take place to ultimately produce a mature egg that is ovulated (Figure 1D–G) [24].

Figure 1. Mouse follicle formation and folliculogenesis.

(A) In the fetal ovary, incomplete cytokinesis results in the formation of germ cell cysts. The germ cells are connected through bridges that act as intercellular transport elements and synchronize meiotic initiation of germ cells (blue). (B, C) Right before birth, the cysts start to break down, and approximately 33% of these germ cells will be enclosed by granulosa cells (yellow) establishing the primordial follicle pool. Although signaling occurs between germ cells and granulosa cells at the primordial follicle stage, whether this is true bidirectional communication has not been well established. (D-F) Stages of oogenesis that are the focus of functional comparison with Drosophila (shaded). (D) At the primary stage, the zona pellucida (purple) is assembled as well as the transzonal projections (yellow filopodia) to ensure oocyte-granulosa cell communication, and this communication is maintained until the oocyte is ovulated. Following follicle activation, the oocyte undergoes a significant increase in transcription and translation [98,99]. (E) Secondary follicles represent the active growth phase of the oocyte, and the mRNAs and proteins that are actively synthesized and accumulate in the oocyte cytoplasm at this stage and throughout oogenesis are likely important for oocyte maturation and future embryonic development. Theca cells (blue) recruited to the antral follicle produce androgens which are converted into estrogen by the granulosa cells. (F) The antral follicle is the last stage of folliculogenesis and is gonadotropin-dependent; it comprises an oocyte and differentiated granulosa cells. The antrum, a fluid-filled cavity, separates the cumulus cells from the mural granulosa cells. Within large antral follicles, de novo transcription is silenced in fully grown oocytes and this correlates with the transition of the chromatin from a non-surrounded nucleolus configuration (euchromatin) into a surrounded nucleolus configuration (heterochromatin) [100]. The transition to transcriptional quiescence occurs in preparation for oocyte maturation. (G) At the time of ovulation, the oocyte undergoes meiotic maturation and the cumulus oocyte complex (COC) is released from the antral follicle into the fallopian tube. The transzonal projections retract at this point in development.

When a follicle is activated, the oocyte starts to grow, and the granulosa cells change their morphology from squamous to cuboidal. These follicles, known as primary follicles, are surrounded by a single complete layer of granulosa cells (Figure 1D). During the active growth phase, the oocyte synthesizes the zona pellucida, a glycoprotein matrix that separates the oocyte from the granulosa cells, and produces and stores mRNAs, mitochondria, proteins and other organelles which are necessary for later development [26–28]. Granulosa cells continue to proliferate, generating several layers around the oocyte and eventually, theca cells are recruited outside the follicle basement membrane in a highly vascularized region. Theca cells produce androgens that will be converted to estrogens by the granulosa cells and are essential for follicle development [29]. These follicles are classified as secondary follicles (Figure 1E) [30–32]. Until this stage, the follicles are not dependent on gonadotropins for development [33].

The formation of the antral follicle stage is defined by the establishment of a fluid-filled cavity in the follicle, known as the antrum, and the differentiation of the granulosa cells into two populations, mural granulosa cells and cumulus granulosa cells (Figure 1F). Mural granulosa cells, which form the outer wall of the follicles, are in contact with the follicular basement membrane. These cells respond to the Luteinizing Hormone (LH), which triggers follicle growth, meiotic resumption, and ovulation. On the other hand, cumulus granulosa cells directly surround the oocyte and provide nutrient and metabolic support for the growth and maturation of the oocyte [7].

Oogenesis culminates with ovulation. Within the periovulatory antral follicle the oocyte is surrounded by multiple layers of cumulus cells that separate it from the antral fluid in what is called the cumulus oocyte complex (COC). At the time of ovulation, which is hormonally regulated by the rapid increase in LH produced by the anterior pituitary gland (commonly known as the LH surge), the oocyte resumes meiosis and the COC is released from the ovary into the oviduct when the follicle wall ruptures (Figure 1G). Prior to ovulation the physical connections between the oocyte and the cumulus cells are lost, marking the last time when the oocyte and the companion cumulus cells have direct communication [34].

Granulosa cell to oocyte communication

The continuous communication between the granulosa cells and the oocyte is important for follicle development because if coupling between the oocyte and granulosa cells is abolished, follicles do not reach the antral stage [35]. In the follicle, gap junctions are one of the main mediators of intercellular communication. These structures are intercellular channels that allow exchange of small molecules (<1 kDa) between adjacent cells. Within the ovarian follicle, there are two primary gap junction proteins: gap junction protein alpha 4 (GJA4) and gap junction protein alpha 1 (GJA1). These proteins are commonly known as connexin 37 (CX37) and connexin 43 (CX43), respectively. In the adult ovary, GJA4 primarily mediates communication between the oocyte and the granulosa cells while GJA1 mediates granulosa-granulosa coupling [36,37]. Despite these cell type specific functions, the ectopic expression of GJA1 in the oocytes of Gja4-null mice mutants restores oocyte-granulosa bidirectional communication, oocyte growth, maturation and fertility, suggesting that these connexins are physiologically interchangeable [37]. Although connections between the oocyte and the granulosa cells are present within the primordial follicle, GJA4 and GJA1 depletion only affects folliculogenesis at later stages, suggesting that gap-junction mediated bidirectional communication may not be functional in primordial follicles [35,38].

Oocytes within primary follicles synthesize and assemble the zona pellucida (ZP), an extracellular matrix that surrounds the oocyte. Although the ZP physically separates the oocyte and granulosa cells, transzonal projections (TZPs) cross through the ZP to enable granulosa-oocyte communication (Figure 1D). TZPs are specialized filopodia that emanate from the granulosa cells, penetrate the zona pellucida, and contact the oocyte to mediate intercellular communication within the follicle [39]. Two different TZPs subpopulations have been identified with the most abundant population having a central core of actin and the less abundant one containing tubulin [39]. The functional relevance of these different TZP populations is not understood.

Mouse oocytes receive many small molecules from the granulosa cells through TZPs. Although communication occurs between the granulosa and oocytes at the primordial stage, in this review we will focus on molecules that are exchanged after the follicle is activated and the TZPs are assembled. The nature of this transport has been studied within actively growing preantral follicles, including primary and secondary follicles. However, the bulk of what is known about exchanged materials comes from COCs within antral follicles because isolation and functional studies of COCs are easier to perform. Metabolic coupling refers to the process in which granulosa and cumulus cells provide the oocyte with products to compensate for its metabolic deficiencies, including sugars, amino acids, nucleotides, electrolytes and cholesterol. For example, oocytes carry out glycolysis poorly, and therefore, they require products of glycolysis such as pyruvate from the granulosa cells [40–42]. In addition, granulosa cells provide alanine, histidine, and leucine to the oocyte because the oocyte lacks the sodium-coupled neutral amino acid transporter SLC38A3 that has a substrate preference for these amino acids [43]. The oocyte expresses the cholesterol biosynthetic enzymes at low levels, and therefore, cholesterol is also obtained from the cumulus cells. Due to its size, cholesterol is probably not transported through gap junctions but instead it might be transferred across the plasma membrane of granulosa cells that lie in close apposition to the oocyte [31,44]. A similar metabolic coupling approach is needed to regulate oocyte pH. Oocytes do not have the machinery to control pH levels and require the acidosis-correcting mechanisms of the granulosa cells to maintain a homeostatic status [45–47]. Why mouse oocytes are transcriptionally active but deficient in certain key molecules and metabolic pathways may reflect the role of its support cells. Indeed, it has been proposed that outsourcing catabolic metabolism to granulosa cells keeps the levels of oxidative stress in the oocyte low, acting as a mechanism to protect the germ cell (Table 1) [48]. Organelles such as mitochondria, endosomes, and lysosomes may also be translocated from granulosa cells to the oocytes through TZPs, but more studies are needed to confirm these observations and to understand how this active transport is driven [49]. In addition to organelles, large molecules, including mRNAs and long noncoding RNAs, can transit to the oocyte from cumulus cells via transzonal projections [50,51]. These RNA trafficking studies have been performed in the bovine model so the broad applicability of these observations across species is unknown.

Table 1.

Known molecules exchanged between oocytes and granulosa cells in the mouse

Follicle stage	Known molecules exchanged	Biological process regulated	Reference
Preantral follicles	Electrolytes	Oocyte pH regulation	[45–47]
	GDF9	Folliculogenesis: TZP connections, granulosa cell proliferation, cell junction establishment	[39,57–61,66,101]
	BMP15
	R-spondin2
Antral follicle and COC	Pyruvate	Metabolic coupling - glycolysis	[13,40–42,102]
	Amino acids	Metabolic coupling - protein synthesis	[43]
	Cholesterol	Metabolic coupling - cholesterol biosynthesis	[44]
	Mitochondria, endosomes and lysosomes	Organelle transport	[49]
	GDF9	Cholesterol biosynthesis, glycolysis, amino acid uptake, folliculogenesis (TZP connections, granulosa cell proliferation and expansion, HA deposition, luteinization)	[5,6,39,52,53,65,103]
	BMP15
	FGF8

Oocyte to granulosa cell communication

During mouse oogenesis, the oocyte drives bidirectional communication to outsource metabolic functions to the cumulus cells to compensate for its deficiencies in glycolysis, cholesterol biosynthesis, and amino acid uptake [48,52]. Oocyte-mediated control of these processes was demonstrated through elegant experiments involving oocytectomy (surgical removal of the oocyte, OOX) from the cumulus oocyte complex [43,44,53]. Upon OOX, specific gene expression and metabolic processes were reduced in the cumulus cells. These functions in OOX cumulus cells could be rescued by co-culturing with fully grown oocytes or with a defined cocktail of oocyte-derived paracrine factors, including GDF9, BMP15, and FGF8. Together these results demonstrate the central role of the oocyte in controlling somatic cell function. Such oocyte-mediated regulation of granulosa cell function may be a mechanism by which the oocyte controls the rate of follicle development. In fact, when oocytes isolated from secondary follicles are cultured with granulosa cells isolated from primordial follicles in a heterochronic aggregation model, folliculogenesis is accelerated [8]. In contrast to granulosa to oocyte communication, which occurs mainly through the gap junctions located at the TZPs, it is thought that the oocyte communicates with the granulosa cells by secreting specific factors into the intercellular space that will bind to their receptors on the granulosa cells. Just as cell processes extend from the distal layers of cumulus cells to the oocyte, it is likely that processes extend to granulosa cell layers from the oocyte. Interestingly, two recent studies addressed this question by injecting fluorescent proteins that are too large to fit through gap junctions directly into the oocyte and found that they can transit from the oocyte to the granulosa cells [54,55]. This has been shown to occur via exocrine-like vesicle exchange between microvilli-like structures [55] and possibly through other connections [54]. For example, recent electron microscopy studies revealed the presence of putative channel-like structures between granulosa cells as well as between oocyte and granulosa cells, which may mediate communication. These connections are thought to be structurally different from TZPs because the cell membranes apparently fuse directly without the typical TZP tip on the oocyte surface [54] and they may more closely resemble the membrane gaps characterized in the germ cell cysts [56] or Drosophila ring canals (see Drosophila section). However, the existence of these structures must be interpreted with caution as the observed cytoplasmic continuity detected using electron microscopy may simply be an artifact caused by the orientation of the plane of the membrane where the oocyte process contacts the granulosa cell. In fact, a merged cytoplasm would obviate the need for gap junction-based communication and would negate well characterized patterns of cell type specific gene expression. Further work is needed to characterize macromolecule exchange between the oocyte and granulosa cells, the physical means of this communication, and the mechanisms that govern directionality. This understanding may reveal additional similarities to Drosophila oogenesis where cytoplasmic connections facilitate sharing of structures as large as organelles (see Drosophila section) among several cell layers. Nevertheless, these recent discoveries show that the oocyte has established complex and elegant mechanisms to communicate with the granulosa cells.

GDF9 (growth differentiation factor 9), a member of the transforming growth factor-beta superfamily, is a well characterized oocyte secreted factor that mediates oocyte to granulosa cell communication [57,58]. GDF9 is expressed in the oocyte within growing follicles beginning at the primary follicle stage [57]. In Gdf9-null mutant mice, follicles reach the primordial and primary follicle stages but not beyond. In these mutant mice, the oocyte increases in size but granulosa cell proliferation is impaired such that the follicles never become multi-layered. The number and morphology of the TZPs are also severely compromised in these mutant mice. These results demonstrate that GDF9 is required for granulosa cell proliferation, TZP assembly, and folliculogenesis [39,58,59].

BMP15 is another member of the transforming growth factor (TGF)β superfamily and also a well-characterized oocyte-secreted factor [60]. In contrast to Gdf9, Bmp15 knockout mice are subfertile, with decreased incidences of ovulation and fertilization [61]. Both GDF9 and BMP15 are closely related paralogs but they signal through different pathways in granulosa cells. In particular, GDF9 binds to a BMPR2-ACVR1B receptor dimer and signals through SMAD2/3, whereas BMP15 bind to a BMPR2-BMPR1B receptor dimer and signals through SMAD1/5/8. The SMADs move to the nucleus where they activate target genes that will increase the glycolytic and tricarboxylic acid (TCA) cycle activity, cholesterol biosynthesis, amino acid transport, and mTOR activity. Products synthesized in the somatic cells to meet the metabolic demands of the oocyte will then be transported to the oocyte [39,43,44,62,63]. These signaling pathways also activate target genes that regulate the number of TZPs and mediate cumulus expansion. In addition, GDF9 and BMP15 can form heterodimers that are more active oocyte-derived regulators of granulosa cell function compared to GDF9 and BMP15 homodimers and may thus be the physiologically functional version [63,64]. Less explored is the role of the fibroblast growth factors (FGFs) as oocyte-secreted signaling molecules, but it is known that FGF receptors are expressed in mouse cumulus cells and that the synergistic effect of BMP15 and FGF8 promotes the expression of glycolytic genes in these cells [52,65]. Finally, growing oocytes also secrete R-spondin2, which promotes granulosa cell cycle progression, proliferation and establishment of cell junctions by activating the WNT-β-catenin (CTNNB1) signaling pathway in the granulosa cells. In the absence of R-spondin2, follicles arrest at the primary stage, similarly to the Gdf9-null mutant mouse model (Table 1) [66,67].

Introduction to Drosophila melanogaster oogenesis

Drosophila oogenesis takes place in an assembly line fashion within a structure called an ovariole. The formation of the egg begins with the asymmetric division of a germline stem cell in a region called the germarium at the tip of the ovariole, creating a cystoblast. The cystoblast then enters a modified mitotic cycle that lacks the final pinching-off step of cytokinesis. Through four rounds of synchronous division, a 16-cell interconnected cyst is thus formed. This division pattern occurs the same way for every cyst and creates a lineage tree where each cell is uniquely identifiable based on the pattern of connections [68]. The 15 stalled cytokinetic furrows, known as ring canals, are reinforced over the course of oogenesis and allow for the sharing of many different cytoplasmic components [69,70]. Ring canals undergo constant remodeling and can grow up to 20 μm in diameter during oogenesis [71]. Large molecules, even organelles, can be exchanged among cells of the cyst via the ring canals [70,72–74].

One of the two cells resulting from the first division of the cystoblast becomes specified as the oocyte and is maintained in prophase of meiosis I until a later stage when meiosis resumes, pauses again at metaphase I, and then completes meiosis during fertilization [75,76]. The other 15 cells in the cyst enter a modified cell cycle of alternating G and S phases, the endocycle, and become polyploid nurse cells that will grow to over 100 times their original size as oogenesis progresses [4]. After the cyst is formed, it comes in contact with somatic stem cells that initiate the process of cyst envelopment by a monolayered epithelium called the follicle cells. The encapsulation of the germline cyst by follicle cells is one major checkpoint in egg chamber development. It is thought that if the proper number of follicle cells do not surround the cyst due to nutrient deprivation, cell death is triggered as a control mechanism [77]. The 16-cell cyst surrounded by the follicle cells constitutes an egg chamber (Figure 2A).

Figure 2. Drosophila egg chamber formation and mechanics of bidirectional communication.

(A) Within the germarium, one cell divides 4 times with incomplete cytokinesis to create a 16-cell cyst. One of the two cells from the original division is then specified as the oocyte and the other 15 cells become nurse cells. After encapsulation by a layer of the somatic follicle cells, an egg chamber is formed. Drosophila egg chambers develop through 14 morphologically distinct stages, resulting in a mature egg. Stages 3–10 (shaded) are the focus of functional comparison with the mouse (see shaded region of Figure 1). (B) Prior to stage 7, the microtubule network in the germline cyst is anchored in a microtubule organizing center located within the oocyte. During stages 7 and 8, the microtubule network reorganizes. Asymmetric actin baskets encircle the ring canals. (C) Products that are required in the oocyte are produced in the nurse cells and are localized to the oocyte on a cytoskeletal-based network. The oocyte-synthesized CDK inhibitor Dacapo diffuses from the oocyte to the nurse cells. Overall, there is thought to be cytoplasmic mixing of general growth regulators between the oocyte and nurse cells [61].

After its formation, an egg chamber progresses through 14 morphologically distinct stages of development during an approximately 3-day period before becoming a fully mature egg awaiting fertilization. During stages 1–6, the oocyte maintains a constant size in relation to the nurse cells [78]. After this period, the cytoskeleton undergoes a major reorganization and nurse cell endocycling pauses before resuming again [79,80]. Stages 7–9 are marked by another major checkpoint in egg chamber development. As the process of yolk uptake from the follicle cells, known as vitellogenesis, is initiated during stage 8, it is thought that this checkpoint senses nutrient availability prior to committing to this major metabolic change [77]. Egg chambers that pass this checkpoint elongate and the oocyte begins a period of largescale growth, eventually becoming equal in volume to the nurse cells. During stage 11 the remaining contents of the nurse cells are ejected into the oocyte in a process known as nurse cell “dumping.” The nurse cells then undergo nonapoptotic programmed cell death [81] and the oocyte finalizes its maturation and awaits fertilization (Figure 2A).

Nurse cell to oocyte communication

From stages 1–10 of Drosophila oogenesis, specific mRNAs, proteins, and organelles from the nurse cells arrive in the oocyte after being actively transported along a microtubule network that passes through the ring canals (Table 2) [73,79]. This delivery of nurse cell products enables essential biological processes from maintenance of oocyte quiescence to axis specification [82–85]. From stage 1 until stage 7, microtubules nucleated by a microtubule organizing center within the oocyte extend through ring canals into the nurse cells. After stage 7, the microtubule network reorganizes and microtubules are nucleated at the oocyte cortex [79]. Recent work identified microtubules that traverse the ring canals with their minus ends pointed towards the oocyte [73]. The minus end directed motor Dynein associates with adaptor proteins Egalitarian and Bicaudal-D that are necessary to facilitate the loading and directed transport of molecules to the oocyte (Figure 2B) [73,86–88]. Actin also plays a role in facilitating directed movement of products from the nurse cells to the oocyte. Protrusions of actin that surround ring canals, called actin baskets, are asymmetrically enriched on the side of the ring canals pointing away from the oocyte (Figure 2B). These structures are necessary to properly polarize the microtubule network via interactions with the spectraplakin short stop [73], and are also thought to govern the passage of products through the ring canals. In one study, only 17% of Golgi that arrived at the ring canals passed through, though what regulates this possible gatekeeping mechanism is unknown [72].

Table 2.

Known molecules exchanged between oocytes and nurse cells in Drosophila

Egg Chamber Stage	Known molecules exchanged	Biological process regulated	Reference
Germarium	Oocyte specification factors	Balbiani body and oocyte formation	[85]
	Mitochondria		[74]
Stages 1–10	Oocyte specification factors	Maintenance of oocyte identity/Prevention of cell cycle entry	[4]
	Patterning factors	Polarization of oocyte and follicle cells	[12]
	CDK Inhibitor Dacapo	Nurse cell endocycle regulation	[90]
	Organelles	General metabolism	[72,73]
	Ribosomes
	Lipid droplets
	Yolk
Stage 11+	Cytoplasm	General metabolism	[89]
	Nuclear degradation products

While active transport creates localized pools of specific molecules in the oocyte, there is a uniform distribution of metabolic proteins, ribosomes, and other translational machinery throughout the germline cyst. It has been proposed that the two different modes of molecule redistribution, active transport-based and diffusion-based, reflect the needs of the different cell types within the system [70]. Specification factors are dedicated for the oocyte, while general growth regulators are required throughout the cyst (Figure 2C). This hypothesis, along with mechanisms of ring-canal “gatekeeping” warrants further study.

Nurse cell dumping is the final step in the nursing of the oocyte. During this time, the nurse cells squeeze their remaining contents into the oocyte, resulting in a final accumulation of all nurse cell cytoplasmic contents, notably lipids and glycogen to prepare for embryonic metabolism [84,89]. The three modes of nurse cell nursing behavior, one highly selective (active transport), one diffusive (general growth factors), and the other a wholesale process (nurse cell dumping), are indicative of the complex role of the nurse cells over the course of oogenesis.

Oocyte to nurse cell communication

Drosophila oocytes have generally been thought of as passive recipients of nurse cells products. However, two studies have documented movement of proteins from the oocyte to nurse cells [70,90]. One of these proteins, the CDK inhibitor Dacapo (Dap; CDKN1B or p27 in mammals), is produced in the oocyte and regulates nurse cell growth. dap mRNA transcribed in the nurse cells is actively transported to the oocyte, where it then translated. Photoconversion experiments with Dendra2-tagged Dap showed that Dap diffuses from the oocyte to the nurse cells [90] (Figure 2C). How translation of dap is controlled such that Dap is produced primarily in the oocyte remains to be determined. The retrograde movement of Dap has the effect of coordinating the endocycles within groups of nurse cells based on their proximity to the oocyte. These endocycling groups not only replicate their DNA together, they also grow together. Importantly, this mechanism allows for growth coordination of the nurse cells across distances of up to hundreds of microns, conferring collective regulation at the tissue-level [90]. Though this finding is the first example of the Drosophila oocyte sending information back to regulate processes in the nurse cells, it raises the possibility that additional instructions are similarly relayed.

Functional similarities of mouse and Drosophila bidirectional communication

When comparing model organisms, one typically looks for similarities based on homologous structures or genes. For example, both mouse and Drosophila oocytes develop in cysts of interconnected germline cells. In addition, both mouse and Drosophila oocytes become surrounded by somatic epithelia, the granulosa and follicle cells, respectively. However, while the cyst structure is maintained during much of Drosophila oogenesis, it breaks down in the mouse prior to largescale oocyte growth. Mouse and Drosophila oogenesis thus seem to diverge, with unrelated cell types – the granulosa cells and nurse cells – supporting oocyte growth. Furthermore, while there are examples of gene homology that are relevant for general ovarian function [91], it is unclear whether those identified are relevant to bidirectional communication. While cell lineage or gene homology differences may exist between the two species, our premise is that there is an essential functional relatedness as it pertains to the common biological process of oocyte growth. Although the mechanisms and molecules differ, the granulosa cells and nurse cells are both programmed to increase their collective biosynthetic capacity and engage in bidirectional communication with the oocyte. Interestingly, despite their germline origin, Drosophila nurse cells more closely resemble somatic cells epigenetically [92], so the distinction of somatic vs germline-derived support tissues may be inconsequential. Regardless of whether the comparable support functions of granulosa and nurse cells result from convergent evolution or by divergence from an ancestral system of growth regulation, bidirectional communication is at the center of successful oocyte development in both organisms. We propose that further insight will be best gleaned by focusing on the functionally related processes that the systems share. For example, in both systems the oocyte is responsible for regulating growth of the support cells. This regulation must be maintained over the course of oogenesis as the support cells proliferate and have increasingly varied proximities to the oocyte. In return, both the granulosa and nurse cells are responsible for metabolically supporting the oocyte. It is thought that since the mouse oocyte is transcriptionally active during the growth phase and synthesizes many of its own macromolecules, it only relies on granulosa cells for metabolites and ions, which can fit through gap junctions. On the other hand, the largely transcriptionally quiescent Drosophila oocyte depends on macromolecules produced by the nurse cells and therefore the ring canals are large enough for macromolecules to pass through. However, the discovery of other types of connections between the oocyte and granulosa cells may disrupt this dichotomy.

Both model systems also present an opportunity to explore feedback mechanisms. The high rate of atresia in the mouse during normal follicle development and the existence of several apoptosis-inducing checkpoints during Drosophila oogenesis suggest that there are built-in control mechanisms to ensure oocyte quality. Bidirectional communication may play a role in sensing the progress of oocyte development to slow growth or trigger apoptosis/atresia if the needs of the oocyte are not being met. How this feedback would promote metabolic or other physiological changes in the oocyte and support cells is an important area for future investigation.

Finally, although outside the scope of this review, signaling via local mechanical properties, known as mechanotransduction, between the oocyte and supporting tissues may also function as a form of bidirectional communication in both systems. For example, there is evidence that physical compression prevents primordial follicles from activating [93]. In Drosophila, a mechanical steady state of actin tension has been proposed to coordinate growth of the somatic follicle cells and the nurse cell-oocyte complex [94]. Changes in tissue biomechanics are also relevant in the context of ovarian aging as it was recently reported that ovarian stiffness significantly increases with age in mice, potentially impacting oocyte quality [95]. Understanding whether age-associated changes in tissue biomechanics influence oocyte-support cell communication and oocyte quality becomes increasingly relevant as the average maternal age rises [96].

Concluding remarks

Despite structural and physiological differences between oogenesis in Drosophila and the mouse, there are commonalities in how the two systems achieve the same goal: creating a healthy egg. Since bidirectional communication has emerged as a fundamental feature of oogenesis among animals, it will be important to interrogate features that are well-established in one system but may not have been considered in another (see Outstanding Questions Box). Manipulating or ablating the physical structures that mediate connections between oocytes and supporting cells will provide important insights into the mechanisms and roles of bidirectional communication during oogenesis across model systems. Such approaches will become more feasible as the components comprising each structure are further elucidated which will inform targeted genetic and pharmacologic approaches. Experimentally, the strengths of each system can be leveraged to fill gaps left by weaknesses in the other. For example, Drosophila oocytes can only be cultured ex vivo for short periods [97] and physical manipulation of nurse cells causes the egg chamber to rupture. Neither of these limitations apply in the mouse. Conversely, while the timescale of genetic manipulation in the mouse has greatly improved in recent years with better CRISPR technologies, the speed, cost, and breadth of genetic tools available to Drosophila researchers allows for the construction of animals with complex genetics backgrounds comparatively easily. Parallel studies in multiple model organisms will provide a faster route toward fundamental insights that may benefit the search for future diagnostics and therapies to improve gamete quality and reproductive outcomes.

Outstanding Questions.

What are the evolutionary pressures that have forced the mouse oocyte to outsource the production of some metabolites?

Why are there two different populations of transzonal projections (microtubule vs actin)? And what are their roles?

What is the difference between the bidirectional communication that is mediated by TZPs versus the bidirectional communication that is mediated by paracrine factors secreted into the extracellular space?

In addition to TZPs, do structures exist during mouse folliculogenesis that are more analogous to Drosophila ring canals? If so, what molecules are being exchanged? How does their structure compare to TZPs?

What molecular mechanism controls whether molecules diffuse or are actively trafficked through ring canals?

Are there other molecules besides Dacapo that the oocyte produces to communicate with nurse cells?

Does the Drosophila oocyte communicate with nurse cells though paracrine factors like in mice?

Are there species-specific differences in the molecules that are being exchanged between cell types?

Does the oocyte engage in bidirectional communication with cell/tissue types?

What role does bidirectional communication play in aging?

Could defects in bidirectional communication be linked to infertility conditions such as primary ovarian insufficiency?

Highlights.

Bidirectional communication between oocytes and support cells is essential for oocyte development and maturation. This dynamic is a common feature of both mouse and Drosophila oogenesis and is likely conserved in other animal systems.

Physical cell structures facilitate the exchange of nutrients, growth factors, metabolites, and organelles between the oocyte and support cells in both mice and Drosophila. In the mouse, communication occurs through transzonal projections and the secretion of factors within the extracellular space. In Drosophila, factors are exchanged through ring canals.

Although it is established that the mouse oocyte drives bidirectional communication with granulosa cells, the Drosophila oocyte until recently was considered a passive recipient of nurse cell products. Emerging studies show that the Drosophila oocyte also engages in communication with these support cells.

Glossary

Actin basket

actin that protrudes from the perimeter of ring canals and is thought to direct transport through ring canals

Antrum

the fluid-filled cavity within the antral follicle that surrounds the cumulus oocyte complex

Cumulus cells

granulosa cells that immediately surround the oocyte within the antral follicle and provide nutrient and metabolic support for the growth and maturation of the oocyte

Cumulus-oocyte complex

the oocyte surrounded by several layers of cumulus cells within the antral follicle

Cystoblast

single daughter cell arising from the division of a germline stem cell

Drosophila germline cyst

cluster of 15 interconnected nurse cells and an oocyte that is created from 4 rounds of incomplete division without cytokinesis from the cystoblast

Egg chamber

germline cyst surrounded by a layer of the epithelial follicle cells

Endocycle

modified cell cycle whereby cells replicate their DNA without dividing, resulting in polyploidy

Follicle

the functional unit of the mouse ovary, composed of an oocyte surrounded by supporting granulosa cells

Granulosa cells

the somatic cells that surround the oocyte in the follicle. Granulosa cells communicate with the oocyte to support its growth and development within the follicle.

Meiotic spindle

barrel shaped microtubule-based structure that segregates chromosomes during the meiotic cell divisions

Mural granulosa cells

granulosa cells that form the outer wall of the antral follicle and are in contact with the basement membrane. These cells respond to the Luteinizing hormone (LH) surge and are involved in meiotic resumption and ovulation.

Mouse germline cyst

nest of germ cells that are connected through cytoplasmic bridges and surrounded by pre-granulosa cells. Mouse germline cysts are found in the fetal ovary.

Mouse germline cyst break down

the process by which the cyst composed of germ cells is fragmented. While most of these germ cells undergo apoptosis, some will be surrounded by squamous granulosa cells to form primordial follicles which dictate the ovarian reserve.

Nurse cells

group of 15 polyploid Drosophila germline cells with connections to each other and to the oocyte that produce maternal gene products and support the growth and development of the oocyte

Nurse cell dumping

the process by which the nurse cell contents are transferred to the oocyte at the end of oogenesis

Oocyte

female germ cell (arrested in prophase of meiosis I in mouse)

Ovarian reserve

number of primordial follicles in the ovary that dictates reproductive lifespan

Ovariole

subunit of the Drosophila ovary containing multiple egg chambers at different stages of oogenesis

Ring canal

stalled cytokinetic furrow that is structurally reinforced and directly connects the cytoplasm of two cells

Theca cells

stroma-derived cells that surround the external part of the follicle basement membrane and whose main function is to produce androgens which are converted to estrogens by the granulosa cells. Theca cell recruitment occurs in the early growing follicle at the late primary to early secondary stages of development.

Transzonal projections

specialized filopodial projections derived from granulosa cells which pass through the zona pellucida and make contact with the oocyte to mediate intercellular communication

Zona pellucida

glycoprotein matrix that is produced by the oocyte beginning at the primary follicle stage and separates the oocyte from the granulosa cells. The transzonal projections emanating from the granulosa cells cross through the zona pellucida to make contact with the oocyte.