Macrophages: an indispensable piece of ovarian health

Abstract

Macrophages are the most abundant immune cells in the ovary. In addition to their roles in the innate immune system, these heterogeneous tissue-resident cells are responsive to tissue-derived signals, adapt to their local tissue environment, and specialize in unique functions to maintain tissue homeostasis. Research in the past decades has established a strong link between macrophages and various aspects of ovarian physiology, indicating a pivotal role of macrophages in ovarian health. However, unlike other intensively studied organs, the knowledge of ovarian macrophages dates back to the time when the heterogeneity of ontogeny, phenotype, and function of macrophages was not fully understood. In this review, we discuss the evolving understanding of the biology of ovarian tissue-resident macrophages, highlight their regulatory roles in normal ovarian functions, review the association between certain ovarian pathologies and disturbed macrophage homeostasis, and finally, discuss the technologies that are essential for addressing key questions in the field.

Macrophages exhibit different origins, heterogeneous phenotypes and activation states, and are actively involved in multiple aspects of ovarian physiology and pathologies.

Introduction

In the late 19th century, Russian zoologist Élie Metchnikoff discovered a group of white blood cells with phagocytic activities in starfish larvae. He noticed with great interest that these phagocytes were not only capable of attacking and eliminating foreign particles from the host, but were also involved in removing dying host cells during developmental processes such as the metamorphosis of tadpoles [1, 2]. This group of phagocytes was later named macrophages.

As a central component of innate immunity, macrophages are known to perform several major functions during infections and inflammation: (1) they engulf and breakdown pathogens during infection, and present antigens on their surface to helper T-cells [3–5]; (2) they produce various biochemical signals, including cytokines, to coordinate and recruit other immune cells [6], as well as growth factors that regulate cell proliferation and angiogenesis [7]; and (3) during wound healing and tissue repair, they secrete various enzymes and extracellular matrix (ECM) components that facilitate tissue remodeling [7]. In addition to their significance in the immune response, as noted by Metchnikoff, the functions of macrophages are also interwoven into the development and maintenance of various host tissues in vertebrates [8].

It has become increasingly clear in recent decades of research that macrophages are phenotypically and functionally heterogeneous in nature, and exhibit high levels of plasticity in their functions when exposed to different environments (Figure 1) [9, 10]. Traditionally, in in vitro studies, macrophages that are exposed to pro-inflammatory signals such as lipopolysaccharide (LPS) and interferon γ (INF-γ) can be classically activated into M1 macrophages [11, 12] that express high levels of inducible nitric oxide synthase (iNOS) and generally specialize in eliminating intracellular pathogens, whereas macrophages that are exposed to anti-inflammatory signals, such as interleukin (IL)-4 and IL-13, can be alternatively activated into M2 macrophages [11, 13] that are involved in tissue remodeling and repair, as well as the resolution of inflammation. The capability of responding to various stimuli reflects the highly plastic property of macrophages, which allows local macrophage populations to adapt to tissue-specific environments and functions [14, 15] and become an integral part of local biological processes. Although the field has gained tremendously from the M1/M2 paradigm, many studies have revealed the high complexity of tissue-resident macrophage biology, which cannot be simply applied to the M1/M2 model.

Figure 1.

Macrophage ontogeny and polarization. Macrophages are derived from multiple origins. Tissue-resident macrophages in adult organs are originated from the embryonic yolk sac and fetal liver progenitor cells during embryogenesis. Monocyte-derived macrophages arise from the bone marrow monocytes. In vitro macrophage polarization paradigm highlights that macrophages are classically activated and polarize to M1-like macrophages when exposed to pro-inflammatory signals like LPS and IFN-γ, whereas M2-like macrophages are alternatively activated when exposed to IL-4 and/or IL-13. In mouse ovary, both embryonic and bone marrow-derived macrophages contribute to the ovarian macrophage pool. M1- and M2-like macrophages are revealed in ovarian macrophage subpopulations.

Ontologically speaking, macrophages in adult tissues are known to originate from multiple sources (Figure 1). Tissue-resident macrophages in adult organs arise from precursor cells that emerge from the embryonic yolk sac and fetal liver during embryogenesis [16, 17]. Monocyte-derived macrophages arise from circulating monocytes that originate from bone marrow [17]. Both types of macrophages are capable of self-renewal within their tissue niches [18].

As a major reproductive and endocrine organ that goes through cyclical changes, the functionality of mammalian ovary is especially reliant on the activities of macrophages [19]. During each reproductive cycle, ovarian tissue goes through substantial remodeling to accommodate the development of the follicle, ovulation, and corpus luteum (CL) formation and regression. The phagocytotic and secretory functions of the ovarian macrophages have been associated with the regulation of many of these processes. The dysfunction of ovarian-resident macrophages, on the other hand, has been increasingly implicated in the inflammation and abnormal biochemical microenvironment associated with pathological conditions in ovarian tissues, including aging [20], polycystic ovary syndrome (PCOS) [21], metabolic disorders [22], and cancer [23]. Thus, a deep delineation of the function of ovarian macrophages bears significant importance to the understanding of the physiology of the ovary, as well as the development of novel therapeutic strategies for ovarian pathologies. The goal of this review is to provide an updated summary of important scientific findings on tissue-resident macrophages in the ovary, as well as their roles in ovarian functions and pathogenesis. We will discuss the biology of ovarian macrophage and how they regulate follicle development, ovulation and CL dynamics, explore the association of ovarian macrophage dysfunctions with ovarian pathologies, and discuss several novel approaches for studying the biology of ovarian macrophages.

Macrophages and ovarian function

Macrophages are highly abundant in the mammalian ovary. Studies in humans and other animals have reported a wide distribution of macrophages in ovarian substructures, with dense localization in regressing CL and atretic follicles, and sparse to moderate presence in the thecal layer of developing follicles [24, 25]. These cells have also been found in abundance in human follicular fluid aspirates, and they make up 5–10% of the total cells in patient follicular aspirates from in vitro fertilization cycles [26]. The presence of macrophages exhibits variations during the ovarian cycles [27], as the number of ovarian macrophages increases during the preovulatory period and diestrus [28, 29].

Consistent with tissue-resident macrophages in other organs, ovarian macrophages also display high level of heterogeneity and multiorigination during the development. A new study by Jokela et al. [30] identified multiple distinct macrophage and monocyte subpopulations in the mouse ovary using single-cell mass cytometry (CyTOF) in combination with fate-mapping experiments. They observed three macrophage clusters with distinct expression levels of major histocompatibility complex II (MHCII), MER tyrosine kinase (MerTK), F4/80, and CD11b. These macrophage groups show overlapping distribution in the ovarian stroma, and the proportions of these macrophage groups exhibit dynamic changes from birth to sexual maturity [30].

The specific localization pattern, heterogeneity, and dynamic presence of macrophages in the ovary imply their active involvement in ovarian homeostasis and sensitivity to hormonal control. This hypothesis has been supported by studies in human and animal models recently, and it is increasingly clear that macrophages, with their pleiotropic functions, play critical roles in multiple aspects of ovarian functions.

Folliculogenesis

At the time of sexual maturity, the mammalian ovary harbors hundreds of thousands of prophase-arrested stage oocytes enclosed in primordial follicles. At each ovarian cycle, some of these follicles are recruited and undergo folliculogenesis in response to local growth factors and hormonal control.

Although there is no evidence indicating a direct interaction between macrophages and primordial follicles, macrophages are commonly found in the theca layer of growing follicles [25, 28]. It is speculated that these leukocytes participate in the regulation of follicular growth through their large panel of secretory mediators. Studies in rodent models have established that the presence of macrophages in the ovarian tissue is essential for the maintenance of growing follicles. Conditional depletion of ovarian macrophages in adult CD11b-diphtheria toxin (DT) receptor (DTR) mice by DT injection leads to hemorrhaging and widespread cell death in developing follicles [19], implicating an indispensable role of macrophages in the maintenance of follicle structure and integrity. Furthermore, mouse primary and secondary follicles revealed significantly better survival and growth if cocultured with macrophage and theca cell-enriched stromal cells in vitro [31]. Consistently, Fukumatsu et al. [32] reported an improved proliferation of rat granulosa cells in the presence of peritoneal macrophages in the culture. Interestingly, in a recent study by Ono et al. [33], it is shown that differing macrophage populations appear to have distinct impacts on the process of folliculogenesis, as the depletion of classically activated M1-like macrophages (CD11c+) causes follicular impairment and hemorrhaging, whereas the depletion of alternatively activated M2-like macrophages (CD206+) does not.

A variety of macrophage-derived growth factors and cytokines, including vascular endothelial growth factor (VEGF) [34], basic fibroblast growth factor (bFGF) [35], macrophage migration inhibitory factor (MIF) [36], epithelial growth factor (EGF) [37], and hepatocyte growth factor (HGF) [38] have been shown to influence folliculogenesis through paracrine and autocrine signaling. VEGF and bFGF are angiogenesis-promoting factors [35, 39]. In both human and animal models, overexpression or treatment of VEGF and bFGF, alone or in combination, has been shown to improve follicle survival and growth [34, 40, 41]. MIF is a cytokine known to be produced by T-cells and macrophages that prohibits the migration of macrophages. Inhibition of MIF signaling with antibodies in mice leads to a reduced number of growing follicles with abnormal morphology and decreased granulosa cell proliferation [36]. EGF is a growth factor known to promote cell proliferation and differentiation. EGF is actively secreted by macrophages in developing follicles [42]. Treatment of EGF increased granulosa cell proliferation and reduced apoptosis in porcine preantral follicle cultures [37]. Lastly, HGF is also shown to have anti-apoptotic effects on cultured rat granulosa cells [38]. However, it is worth noting that macrophages are not the sole source of the factors mentioned above. Nonimmune cells including granulosa cells, theca cells, and oocytes are also capable of producing some or all of these factors [43]. The cellular sources of these factors for folliculogenesis are still unclear. Whether the crosstalk between nonimmune cells and macrophages is essential for the production of key factors during folliculogenesis merits further investigation.

Some pro-inflammatory cytokines that can be produced by macrophages, like tumor necrosis factor α (TNF-α) [44, 45] and IL-6 [46], are shown to exert complex regulatory effects on follicle growth and selection. Although TNF-α appears to promote mouse granulosa cell proliferation in vitro through c-Jun signaling [44], TNF-α knockout mice show increased granulosa cell proliferation and reduced apoptosis [45]. These results suggest that the regulatory effect of TNF-α on follicular cells may be phase specific and context dependent.

In addition to the regulation of granulosa cell proliferation, evidence has also indicated that macrophages may contribute to the vascularization of growing follicles. Ono et al. [33] showed that the elimination of CD11c+ cells (M1-like macrophages and dendritic cells (DCs)) in CD11c-DTR mice by DT injection leads to a reduction in the numbers of pericytes and endothelial cells surrounding the follicles, an indication of impaired angiogenesis. Interestingly, this vasculature impairment is not observed in mice with conditional depletion of CD206+ M2-like macrophages, suggesting a role of M1-like, but not M2-like, macrophages in the vascular integrity in the ovary. Macrophages are known to produce platelet-derived growth factor B (PDGF-B), which mediates the recruitment of pericytes [47, 48]. Ono et al. [33] demonstrated that CD11c+ cell depletion leads to the elimination of PDGF-B signal from ovarian stromal lesions and a reduction of PDGF receptor β-positive pericytes near the follicles, pointing to PDGF-B as a mediator of M1-like macrophages on maintaining ovarian vasculature integrity. However, it should be noted that CD11c is expressed in both DCs and M1-like macrophages. Using CD11c-DTR mice could not exclude the role of DCs in vasculature formation during folliculogenesis. Whether and how M1-like macrophages are definitively involved in the regulation of follicle vascularization represent a significant knowledge gap for future research.

In addition to providing an environment for oocyte development, the growing follicles are also production centers for various signaling molecules and steroid hormones, which communicate with the hypothalamus and pituitary gland to orchestrate the reproductive cycle. Evidence suggests that macrophages may also be influential to the secretory profiles of follicular cells. Altered levels of macrophage-derived cytokines in PCOS model rats have been associated with the decreased production of estradiol [49]. In late stage follicles, follicle-stimulating hormone (FSH) signaling stimulates the aromatization of androgenic steroids into estradiol in granulosa cells [50]. In vitro experiments show that exposure to TNF-α can attenuate the stimulatory effect of FSH on cultured rat granulosa cells [51]. As a major producer of TNF-α, local macrophages may thus represent a paracrine mechanism that modulates gonadotropin signaling in developing follicles [51]. However, experiments in mice appear to reflect a more complex interplay between ovarian macrophage activities and steroid hormone production, as the ablation of CD11c+ cells (M1-like macrophages and DCs) actually reduced the serum estradiol levels of affected mice, whereas depletion of M2-like macrophages exerts no significant impact on estradiol levels [33]. More explicit in vitro and in vivo experiments are still needed to further elucidate the influence of macrophages on the secretory activities of the follicular cells.

Follicular atresia

Among the vast number of primordial follicles that reside in the ovary at menarche, only a very small portion eventually achieve ovulation. Greater than 99% of the follicles will go through atresia during development [52]. Although macrophages only localize to the theca layer of nonatretic follicles, the cells are found to invade granulosa cell layers in large numbers at the advanced stage of atresia subsequent to the disruption of the basement membrane [27, 53, 54], suggesting a role of macrophages in follicle atresia.

The biological process that is most commonly associated with macrophages in atretic follicles is the phagocytosis of apoptotic cells in these follicles [53]. Evidence suggests that the recruitment of macrophages to the site is likely mediated by IL-33 [55, 56], a cytokine mostly produced by endothelial cells around the follicles. The expression pattern of IL-33 in the ovary varies throughout the estrous cycle and is correlated with the fluctuation of macrophages during the wave of follicle atresia [55]. The IL-33−/− mice fail to recruit macrophages to the atretic follicles and exhibit abnormal accumulation of tissue wastes in ovarian tissue [56].

In addition to the removal of dead cells, macrophages are also thought to actively participate in the degradation of atretic follicles by secretion of matrix metalloproteinases (MMPs), as macrophage-associated MMPs, like stromelysin (MMP-3), are upregulated in atretic follicles after macrophage invasion [57]. However, the specific activities of MMPs in ovarian macrophages, as well as whether they contribute to the progression of follicle atresia, remains largely unknown and merits further studies.

Lastly, macrophages may be able to facilitate granulosa cell apoptosis through the pro-apoptotic effect of their secretions, such as TNF-α [45]. However, direct evidence has yet to be found to support this hypothesis.

Ovulation

During ovulation, preovulatory follicles at the surface of the ovary will rupture at the apex and extrude an oocyte. The rupture of the follicle involves the breakdown of ovarian epithelium, underlying connective tissue, the basement membrane, and all the theca and granulosa cell layers at the apical site of the follicle [58]. Evidence suggests that macrophages are involved in the ovulatory process, as the colony stimulation factor-1 (CSF1) mutant mice with depleted macrophages show severely reduced ovulation [59]. Consistent with that, intrabursal injection of liposome-encapsulated clodronate in mice with ablated ovarian macrophages caused reduced ovulation [60]. In contrast, supplementation of leukocytes to perfused rat ovaries increases the number of ovulated oocytes in vitro [61].

The process of ovulation is set off by a surge of luteinizing hormone (LH) released by the pituitary gland. The rise of LH stimulates the local synthesis of numerous pro-inflammatory molecules and cytokines by the granulosa cells, including chemoattractants that recruit monocyte/macrophages [58] like monocyte chemotactic protein-1 (MCP-1) [62, 63], C–C-motif ligand-20 (CCL-20) [62, 64], chemokine (C-X-C motif) ligand 10 (CXCL-10) [65], IL-1 [66], and IL-6 [67]. As a consequence, macrophages infiltrate preovulatory follicles at the time of ovulation initiation [25, 28, 68]. Once onsite, as the major producers of various cytokines and chemokines, macrophages are thought to be involved in amplifying the recruitment of leukocytes through the secretion of cytokines like MCP-1, TNF-α, IL-1β, and IL-6 [58]. Though the molecular mechanisms by which macrophages regulate the ovulation process are still unclear, studies in different animal models have provided some potential candidates. Some pro-inflammatory cytokines are capable of directly promoting ovulation or facilitating ovulatory events. For example, IL-1β treatment has been shown to stimulate ovulation and promote oocyte maturation in rats and mares [69, 70]. In vitro experiments show that IL-6 can promote cumulus cell expansion through the mitogen-activated protein kinase pathway [71]. TNF-α treatment, when combined with LH, can drastically increase the ovulation rate in perfused rat ovaries [72]. Neutralization of TNF-α using an anti-TNF-α antibody in preovulatory follicles inhibits ovulation in ewes [73], indicating a pivotal role of TNF-α in the orchestration of follicle rupture.

In addition to cytokines, macrophages are also a major producer of MMPs. Studies in rodents and bovine showed that the expression of some MMPs are elevated following the LH surge, including MMP-1 (collagenase) [74] and MMP-19 [57], both of which are known to be expressed by macrophages [75]. It is thus reasonable to speculate that, besides secreting immune mediators that target other cells, ovarian macrophages could directly participate in the breakdown of ECM, as well as the regulation of tissue remodeling upon ovulation through the activities of MMPs. However, a study of macrophage-specific MMP expression and activities upon ovulation is necessary to draw definitive conclusions.

Corpora lutea formation and regression

After the extrusion of the oocyte, a ruptured follicle is left behind. This remaining structure will go through substantial remodeling, which involves the luteinization of granulosa cells, vascularization, and leukocyte infiltration, and become a CL [76]. As a transient endocrine organ, the CL is the production center for progesterone (P4), a crucial steroid hormone for the sustainment of uterine development and, in the case of fertilization, embryo implantation [76].

Macrophages are mainly found in the theca–lutein layer of the CLs [77]. The migration of macrophages into the CL has been examined by several studies in human and other models [25, 77–80]. In general, the number of macrophages in the CLs exhibits an increasing trend throughout the luteal phases, with fast accumulation in the late luteal phase in preparation for the subsequent luteolysis [77, 80]. This accumulation is thought to be mediated by chemokine MCP-1/CCL2, the expression of which rises upon luteolysis in many species [80–84]. Notably, in humans, if pregnancy is to occur, human chorionic gonadotropin (hCG) released by trophoblasts will prevent the accumulation of macrophages in the late luteal phase, and thus leads to the persistence of the CL [77].

Macrophage infiltration appears to be crucial for several aspects of CL development and function. First, conditional depletion of macrophages in mice causes significant hemorrhage and loss of endothelial cells in the CL [19, 85], indicating an essential role of macrophages in promoting vascularization and maintaining vascular integrity. Macrophages have been implicated as key regulators of angiogenesis through the production of angiogenic factors such as VEGF, bFGF, and MMPs. It is reasonable to speculate that ovarian macrophages are the key cellular source of those factors. Direct experimental evidence supporting this hypothesis, however, is still missing.

In addition, macrophages also thought to regulate steroidogenesis in CLs. In vitro coculture experiments demonstrated that the presence of macrophages significantly increases progesterone production in cultured luteal cells [79, 86]. Evidence suggests that macrophages and their activation status are crucial for the normal steroidogenic capability of CLs. Histological studies of human ovaries revealed that CL macrophages go through morphological alterations during the cycle, from smooth-edged in the early and late luteal phases to dendritic-like in the midluteal phase [78]. This morphological alteration is consistent with a change in activation status from inactive in the early and late luteal phases and active in midluteal phase. Moreover, the activation status of macrophages appears to be regulated by granulocyte-macrophage colony stimulation factor (GM-CSF), as GM-CSF deficiency is associated with a significant reduction of MHC class II (a marker for M1-like macrophages) expressing macrophages in the developing CL [87]. This disruption of macrophage activation leads to impaired luteinization and reduced steroidogenesis in the CLs [87]. In comparison, mice with deficient transforming growth factor β 1, a key factor that promotes macrophage alternative activation (M2), produce on average 75% less P4 in early pregnancy [88]. These studies highlight a role of macrophages in regulating P4 synthesis, which in turn could lead to establishment of the pregnancy.

In the late luteal phase, the CLs go through regression (luteolysis). In most mammalian species, this process is triggered by prostaglandin F2α (PGF2α) [89]. It is evident, at least in ruminants, that the uterus is the main source of PGF2α [89, 90]. PGF2α signaling inhibits P4 synthesis by blocking the uptake of cholesterol by the mitochondrial membrane [90]. Evidence suggests a role of macrophages in regulating PGF2α synthesis, as depletion of these cells leads to premature luteolysis due to the elevated expression of prostaglandin-endoperoxide synthase 2, an essential enzyme for PGF2α production [87]. Macrophages also produce TNF-α, which is expressed in the CLs, and has been shown to promote the synthesis of PGF2α in bovine CLs [91]. As TNF-α is a characteristic product of M1-like macrophages [92], it is thought that changes of macrophage activation status in the mid-to-late luteal phases are relevant to the onset of luteolysis. Although M2-like macrophages are more predominant in developing CLs, pro-inflammatory M1-like macrophages are associated with CL regression [93]. In line with this hypothesis, macrophages have been shown to express P4 receptors and respond to P4 treatment. Miller and Hunt [94] demonstrated that P4 signaling exerts an inhibitory effect on the production of TNF-α in M1-activated macrophages in vitro. Consistently, the expression of M1 macrophage marker MHCII is significantly elevated in bovine CLs in the late luteal phase [95]. Thus, the crosstalk between macrophages and P4-producing luteal cells appears to be the key regulatory mechanism that is required for CL regression. The underlying mechanisms and the actual roles of macrophages and their activation status during CL progression and regression remain to be defined.

Macrophages in ovarian pathologies

Due to the diverse functions and the ubiquitous distribution of immune cells, the immunological aspect of local tissue dysfunctions and pathologies has drawn increasing research interest in recent decades. Many ovarian pathological conditions have been linked to the abnormal activities of macrophages and other resident immune cells. Here, we will discuss the studies that contribute to our understanding of the interplay between macrophages and various ovarian dysfunctions and pathologies.

Ovarian aging

The mammalian ovary starts the process of aging early in life and remains functionally active only through a finite period of time. Females of advanced reproductive age tend to produce fewer oocytes that are of inferior quality. For women, the risk of producing oocytes with meiotic nondisjunction grows exponentially after the age of 35 [96], and the ovarian cycle all together ceases after menopause at the average age of 51.4 [97]. Along with the reproductive senescence, ovarian aging also leads to the diminished production of various steroid hormones, which cause menopause-associated symptoms. As delaying maternity has become an increasingly common practice in modern society, the loss of reproductive potential due to ovarian aging represents a significant concern for women who choose to devote time and energy in pursuit of their knowledge and careers.

The underlying mechanisms of aging-associated deterioration of ovarian functions are a main focus of reproductive aging-related research. Several lines of evidence suggest that the process of ovarian aging is accompanied by an elevation of basal inflammatory status: (1) transcriptomic analyses revealed an upregulation of genes associated with inflammatory responses in ovaries and follicles retrieved from reproductively old mice, compared with that from reproductively young mice [20, 98]; (2) ovaries from old mice exhibit excessive collagen deposition in the ovarian stroma that are consistent with tissue fibrosis [99]; (3) histological sections of old mouse ovaries revealed the unique presence of a group of multinucleated giant cells in the stroma and follicles, which resemble a hyperactivated form of macrophages [99, 100].

A recent report proposed the aging-associated changes of the ovarian biochemical milieu as a potential cause for the rise of basal inflammatory status in the aging ovary. It is found that hyaluronan, a high molecular weight polymer that makes up the ECM in the ovarian tissue, goes through fragmentation over the course of aging. Exposure to fragmented low molecular weight (LMW) hyaluronan leads to a shift in the secretory profile of cultured ovarian stromal cells towards type 2 (Th2) inflammatory cytokines [101].

Intriguingly, LMW hyaluronan treatment increases the expression of IL-4 and IL-13 [101], which are major cytokines that induce alternative activation in macrophages. Consistent with this dynamic, the comparison of macrophage subpopulations in young and old mouse ovaries revealed an aging-associated shift in macrophage activation status, as the proportion of M2-like macrophages increases significantly in the old ovaries, compared with that in the young [20]. In addition, there appears to be a depletion of local tissue-resident macrophages in the ovary at an advanced reproductive age, and macrophages derived from recruited monocytes become more predominant [20]. It is broadly true that tissue-resident macrophages can proliferate locally, and their renewal is independent of blood monocytes under homeostasis [102]. The replacement of tissue-resident macrophages by monocyte-derived macrophages, however, has been reported under acute and chronic inflammatory conditions [103, 104]. The shifted macrophage ontogeny in the aged ovary indicates that monocyte-derived macrophages can also contribute to the macrophage pool during low-grade, chronic inflammation, in this case, inflammaging. The aged ovary represents a drastically altered tissue environment, including significantly low levels of estrogen and a severely reduced number of follicles. Thus, newly generated ovarian macrophages from monocytes must rapidly adapt to this new tissue niche. Whether monocyte-derived ovarian macrophages are able to fully replenish the ovarian tissue-resident macrophages functionally and metabolically remains to be determined. Moreover, future studies that examine the impact of the shifting macrophage population on different aspects of ovarian functions may reveal more information about the role of macrophages in ovarian aging.

PCOS

PCOS is a prevalent endocrine disorder that describes a collection of symptoms including polycystic ovary morphology, hyperandrogenism, and ovulatory dysfunction [105]. PCOS may affect up to 15% of all women of reproductive age [106] and represents a major cause of infertility in the clinic. No clear single factor has been identified as the cause of the disorder. Rather, PCOS is thought to be a result of the complex interplay between genetic predeposition and environmental elements [106]. The polycystic ovary is known to produce excessive androgens and insufficient progesterone, which perturb the feedback regulation of FSH and LH release, and leads to granulosa cell apoptosis, failure of ovulation, and polycystic ovary morphology [107]. The condition is also frequently associated with metabolic disorders. Among all PCOS patients, 50–70% also suffer from insulin resistance and >60% also show signs of obesity. Even among patients with a normal weight, 30% suffer from abdominal adiposity [108].

An emerging characteristic of PCOS that caught the attention of researchers is the development of systematic, chronic, low-grade inflammation. PCOS patients commonly show elevated serum level of pro-inflammatory cytokines including TNF-α and IL-6 [109]. In the same line, genetic polymorphisms that cause higher IL-6 expression have been associated with the development of hyperandrogenism [110]. TNF-α is a regulator that promotes the classical activation in macrophages (M1) through nuclear factor kappa B (NF-κB) signaling [10]. A study using the PCOS rat model (PCO rats) induced by estradiol valerate revealed that spleen macrophages isolated from induced rats exhibit more pro-inflammatory, M1-like expression profiles with higher production of nitric oxide and higher expression of TNF-α and IL-6 [49]. Consistent with this, blood monocytes and macrophages isolated from PCOS patients in the follicular phase of the ovarian cycle show higher potency for M1-like activation ex vivo, compared with those retrieved from healthy cycling counterparts [111].

On the other hand, the activation status of local ovarian macrophages is also dysregulated in polycystic ovaries. In the PCO rat model, induced by prolonged 5α-dihydrotestosterone (DHT) treatment, the presence of M1-like macrophages in both antral and preovulatory follicles significantly increased after 7 days of DHT exposure [21]. Intriguingly, polycystic ovaries from DHT-treated rats are shown to have an upregulated local expression of pro-inflammatory adipokine chemerin [21]. Chemerin is shown to increase monocyte migration in vitro, and its overexpression is accompanied by an upregulation in its receptor, chemokine-like receptor 1, in M1-like, but not M2-like macrophages in the ovary [21], linking elevated chemerin activity with the shifted macrophage dynamics in the polycystic ovary. Furthermore, chemerin levels are found to be elevated in the ovary, but not the serum of nonobese PCO rats [21], implicating that the hyperandrogenism-induced rise of chemerin production in the ovary may contribute to the pro-inflammatory tissue environment independent of the systematic rise of inflammatory status.

The contribution of ovarian macrophage activities to PCOS pathology still remains unclear due to the lack of convincing in vivo studies. However, in vitro experiments have provided some important clues. It has been shown that DHT treatment leads to increased apoptosis in granulosa cells only when they are cocultured with macrophages, suggesting a role of resident macrophages in mediating the granulosa cell apoptosis [21]. Figueroa et al. [49, 112] studied the effect of macrophage secretions on cultured rat ovarian tissue. They found that exposure to macrophage secretions from PCO rat donors significantly increased the release of androstenedione and decreased the release of estradiol by PCO rat ovaries compared with treatment of that from healthy control donors, which point to a possible influence of macrophages on steroid-producing cells in the polycystic ovary through paracrine signaling. However, it is worth noting that the macrophage secretions used in these ovarian tissue culture experiments were harvested from spleen macrophages, and thus may not fully reflect ovarian macrophage activities at the local level.

In summary, PCOS is characterized by hyperandrogenism, metabolic dysfunctions, and an elevated inflammatory status. The condition is associated with altered macrophage dynamics at the systematic level, as well as locally within ovarian tissues. Intriguingly, clinical data shows that metformin, a commonly used insulin-sensitizing drug, can reliably ameliorate PCOS-associated clinical outcomes [113]. Metformin treatment is shown to effectively lower the serum level of harmful adipokines [114], alleviate inflammatory state [115], and lower the expression of M1 macrophage-associated products (IL-6, TNF-α) and markers (CD11c, MCP-1) in the adipose tissue of obese rats [116]. However, whether metformin influences ovarian macrophages directly is yet to be defined.

Ovarian cancer

Ovarian cancer is a common gynecological malignancy with a high mortality rate, and it is responsible for an estimated 152 000 deaths around the globe annually [117]. Although ovarian cancer can originate from epithelial cells (of the distal fallopian tube, endometrium, or ovarian surface), stromal cells, or germ cells, epithelial ovarian cancers (EOC) account for >90% of all malignant tumors due to a lack of effective early detection methods [117], which represents a serious health threat for women.

Similar to PCOS, the development and progression of ovarian cancer also results from a complex interplay of genetic predepositions and environmental cues. The local biochemical milieu of the tissue is one of the crucial factors that influence the progression of ovarian tumors. Many signaling molecules that regulate cancer cell proliferation and metastasis, including cytokines, enzymes, and growth factors, are known products of macrophages.

Studies of ovarian tumor-associated macrophages (TAMs) revealed a link between TAM activation status and tumor progression. M1-like macrophages are generally thought to be inhibitive to tumor progression. However, some research shows that TNF-α secreted by M1-like macrophages promotes the invasive activities of ovarian cancer cell lines through NF-κB signaling [118]. In contrast, M2-like macrophages are widely accepted as tumorigenic and are commonly associated with poor prognosis of the cancer. In general, ovarian cancer TAMs are more predominantly M2-like macrophages. Ovarian cancer patients with higher M1/M2 TAM ratio tend to show longer overall survival and progression-free survival compared with the others [119–121]. TAMs primarily influence the regulation of tumor progression through paracrine regulation. TAM secretions are known to repress immune responses and promote angiogenesis and cell proliferation. For example, MMP9 and heparin-binding EGF produced by M2-like macrophages can promote tumor cell proliferation and spheroid formation [122–124]. Secretions from TAM and EOC cell cocultures increase endothelial cell migration in vitro [125].

Ovarian cancer cells are known to actively recruit macrophages by secreting chemoattractants like MCP-1. According to a study in mice, macrophages that are initially recruited to the tumor predominantly express M1-specific markers including C-C chemokine receptor type 2 (CCR2), nitric oxide synthase 2 (NOS2), and Interferon Alpha And Beta Receptor Subunit 1 (IFNAR1). These markers are gradually lost as the tumor progresses, whereas the expression of M2-specific markers (CD206, CD163, and CX3CR1) drastically increase [126], suggesting a polarization towards the M2 phenotype and accumulation of M2-like macrophages during tumor progression. Consistent with these findings, clinical data suggests that the progression of ovarian tumors is generally accompanied by a gradual decrease of M1/M2 TAM ratio [120]. In addition to cytokines that stimulate alternative macrophage activation, molecular studies showed that EOC tumor cells can also release exosomes with microRNAs, such as miR-940 and miR-223, which promote alternative activation of TAMs [127–129].

In summary, the interplay between macrophages and ovarian cancer cells is evidently crucial for the progression of the ailment. Interventions that inhibit TAM secretory activities or shift TAM activation status may represent valid strategies for ovarian cancer treatment in the future.

Emerging technologies for the study of ovarian macrophages

Two major factors that constrain the study of ovarian macrophages are the relative scarcity of ovarian macrophages in common model animals and the heterogeneity of the macrophage population. Though macrophages are abundantly present in ovarian tissue, in common rodent models, the ovary is a relatively small organ that harbors a limited number of cells. In addition, like macrophages in other organs, ovarian macrophages exhibit significant heterogeneities in their phenotypes [30], activation status, and changing behaviors during each ovarian cycle as the endocrine and paracrine environment change. These characteristics make it challenging to stably maintain these leukocytes in vitro, as their identities and functions alter rapidly once they are isolated from the tissue niche. On the other hand, traditional analytical methods such as bulk RNA sequencing may leave out important information of differential gene expression among different macrophage subpopulations. Intriguingly, the advance of high-throughput sequencing technologies has brought about several new tools in the past 5–10 years that show great promise for the study of ovarian macrophages at the single-cell level.

Single-cell sequencing is a sequencing technique that allows the examination of genomic or transcriptomic features at the single-cell resolution. So far, the most common application of the technique is transcriptomic analysis by single-cell RNA sequencing (scRNA-seq). The general workflow of the scRNA-seq methods involve the isolation of cells, barcoding of the transcripts from each isolated cell, compilation of the barcoded transcripts, library preparation and amplification, and data analysis [130]. Popular methods for scRNA-seq include Drop-seq [131], CEL-seq [132], massively parallel single-cell sequencing (MARS-seq) [133], etc. These methods can be used in conjunction with fluorescence-activated cell sorting for the analysis of relatively rare cells in selected tissues. In the context of studying ovarian macrophages, these powerful tools will enable the capture of transcriptomic heterogeneity of ovarian macrophage populations and the examination of the dynamics of macrophage subpopulations in ovarian tissue under healthy and diseased conditions. It may also provide a close view of ovarian macrophage activities in different stages of the ovarian cycle.

Several pioneering studies have applied scRNA-seq to the transcriptomic analysis of ovaries and follicles [134, 135]. Wang et al. [134] constructed and compared the single-cell transcriptomic profiles of young and old cynomolgus monkey ovaries in a study of primate ovarian aging. Among the 2183 somatic cells captured, ovarian macrophages can be clearly identified with CD68 and CD14 expression, demonstrating the potential of scRNA-seq technique in aiding ovarian macrophage research. Likewise, using a similar approach, CD68-expressing ovarian macrophages were also observed in the human adult ovary, indicating a conserved phenotype of ovarian macrophages. Moreover, scRNA-seq revealed the expression of C1q in the ovarian macrophage cluster. C1q is a key recognition molecule of the complement system in innate immune responses, and has been shown to promote alternative activation of macrophages towards the M2 phenotype to facilitate tissue repair. The data suggests that ovarian macrophages may act as a cellular source of C1q in the human ovary and highlight the unknown functions of C1q in ovarian health. Further studies on ovarian macrophages would benefit greatly from scRNA-seq on flow cytometry sorted ovarian macrophages, which is likely to reveal their molecular signatures and pave a path to their functions.

Another novel technique for single-cell resolution expression analysis is Slide-seq. This technique utilizes a monolayer matrix of uniquely barcoded beads that are fixed to a slide to capture transcripts directly from frozen tissue sections [136]. In addition to single-cell level transcriptomic data, the association between the barcodes and spatial coordinates in the matrix makes it possible to map the transcriptomic information back onto the original tissue section. In this way, cell identity and transcriptomic profiles can be examined in the context of the tissue structure. Rodriques et al. [136] demonstrated, with a case study of brain traumatic injury response, that microglia/macrophage activities can be captured and visualized using this technique. Slide-seq may be particularly useful for the inspection of macrophage localization and transcriptional activities in different substructures of ovarian tissue. Its application may help to provide direct evidence for the involvement of macrophage-derived products in processes like follicle growth and ovulation.

Although advances of next-generation sequencing technology and flow cytometry will no doubt capture a wealth of information on transcriptional profiles and phenotypes of ovarian macrophages, they cannot reveal the dynamic roles and behaviors of macrophages in their native residential niche. This information is essential for the field to gain a comprehensive understanding of functions of ovarian macrophages, and it requires optical imaging techniques. Intravital microscopy and two-photon fluorescence microscopy are two of the most used imaging approaches that allow the tracking of immune cells, including macrophages in vivo and further investigation of their migration and interactions with other cells in tissues. Moreover, the development of genetic and chemical labeling techniques has greatly promoted the application of microscopy to investigate dynamic functions of macrophages in various contexts [137, 138]. For instance, Zhang et al. [139] applied intravital two-photon imaging for the first time to visualize peritoneal macrophages using Bhlhe40^GFP reporter mice as large peritoneal macrophages highly express Bhlhe40 and therefore result a strong green fluorescent protein (GFP) signal in this mouse model. This landmark study demonstrated that cellular adhesion and macrophage-induced coagulation cooperatively promote immune defenses in the peritoneal cavity. In another study, lung alveolar macrophages (AMs) were labeled chemically by direct administration of PKH26 dye aggregates to the lung. Together with multichannel spinning-disk confocal intravital microscopy, Neupane et al. [140] revealed that AMs constantly patrol and clean the alveolar spaces. More importantly, this migration pattern of AMs appears to be essential for bacterial clearance in the lung. Recent published reviews on in vivo optical imaging to study functions of macrophages and genetical reporter mouse models that allow us to track macrophages can provide a more in-depth discussion of this topic [141, 142]. As demonstrated by the successful applications in the mentioned studies, in vivo imaging techniques provide great tools for the study of local macrophage behaviors in ovarian tissues in real time and may help uncover previously unknown activities of macrophages in the diseased ovary.

Conclusion

Ovarian macrophages are an essential component of ovarian physiology and pathologies (Figure 2). The heterogeneity and plasticity of these leukocytes allow them to follow the rhythm of this uniquely dynamic organ. Due to the increasing awareness of their heterogeneity and plasticity, the importance of studying macrophages within context is becoming increasingly apparent. The emerging new techniques and methods have made it possible for future research to delineate macrophage populations and further examine their functions in the ovary at a higher resolution.

Figure 2.

Roles of macrophages in ovarian health. Ovarian macrophages are an integral part of many biological processes that take place in the ovary. In vitro coculture and in vivo depletion experiments have associated macrophage activities with follicular growth, where macrophages influence granulosa cell proliferation, follicle vasculature, and steroid hormone production. In atretic follicles, macrophages are thought to remove apoptotic cells and facilitate degradation of the structures. During ovulation, macrophages are recruited to the preovulatory follicles and are involved in the further recruitment of leukocytes and likely the breakdown of the ECM. In CLs, macrophages accumulate in the late luteal phase and have been associated with the CL vasculature organization, steroid production, and luteolysis. Perturbed dynamics of macrophage populations has been shown to accompany various pathological conditions in the ovary. Shifting of the macrophage activation status and ontogeny have been associated with ovarian aging and PCOS, as well as certain types of ovarian cancers.

Like macrophages that residing in other organs, both embryonic and bone marrow-derived macrophages contribute to the ovarian macrophage pool. It is apparent that the balance of ovarian macrophage activation needs to be precisely regulated for homeostasis and functions in the ovary. The perturbation of this balance is a recurring theme in many pathological conditions of the ovary. What tissue-derived signaling regulates the heterogeneity of ovarian macrophages? How to maintain a healthy balance of M1/M2 macrophage activities? Could this balance be restored in the dysfunctional ovaries under different pathological conditions? Will restoration of this balance ameliorate ovarian dysfunctions? These questions are emerging in the field and will be interesting directions to explore in future research. Although much remains to be discovered about ovarian macrophages, it is clear that macrophages are intertwined in the complexity of ovarian functionality. An in-depth understanding of ovarian macrophages will lend itself towards the development of innovative clinical solutions. Novel therapies associated with ovarian macrophage studies could help improve the current therapies for the maladies that interfere with normal fertility, steroidogenesis, and ovarian repair mechanisms.

Grant support: This work was supported by the 2020 Global Consortium for Reproductive Longevity & Equality Postdoctoral Fellowship Award through Buck Institute to ZZ; the Reproductive Scientist Development Program (5K12HD000849), Reproductive Health COBRE (P20GM121298-01), the Junior Scholar Award Global Consortium for Reproductive Longevity & Equality awarded to LMB. LH is supported in part by a COBRE grant (1P20GM103625-09) from the National Institutes of Health with the National Institute of General Medical Sciences through the Center for Microbial Pathogenesis and Host Inflammatory Responses at the University of Arkansas for Medical Sciences.