Follicle development in pigs: State of the art

Understanding the factors and pathways involved with recruitment, atresia, and selection of follicles in the pig, may provide insight into approaches to limit fertility failures. Antral follicles depend upon FSH to the 2–3 mm stage, become codependent upon LH at 4–5 mm, and rely on LH when >5 mm. Within the follicle, gonadotropin binding, steroids, growth factors, and inhibin interact to determine the fate of the follicle. Continuous recruitment appears likely for follicles, and once >1 mm, they may have a limited period for survival, before selection or atresia. If true, then the number of healthy follicles that can respond to a hormone signal for selection, could vary by size and development stage. Which follicles are selected may depend upon their age, numbers of capillaries, granulosa and thecal cells, and FSH and LH receptors. This might also suggest that factors such as management, nutrition, and stress in prior weeks, could affect different cohorts of follicles to determine which of those from the ovarian population will be selected.


| INTRODUCTION
Follicle development determines puberty, estrus, ovulation, and subsequent fertility in the pig and has been previously reviewed (Guthrie, 2005;Knox, 2019;Lucy et al., 2001;Schwarz et al., 2008;Soede et al., 2011). While much is known, the impact of hormone patterns and concentrations on follicles at various stages of the cycle remains uncertain. Further, endocrine, paracrine and autocrine factors interact to regulate follicle development to internal and external physiological cues. Activation is a term that refers to the period when primordial follicles leave the resting pool, while recruitment denotes the stage when antral follicles are responsive to FSH. A period of cyclic recruitment is most obvious when all surface follicles disappear after ovulation, and a new wave of small follicles emerge at the start of the new cycle (LaVoie, 2017). Follicles are destined for atresia, but can be rescued, which extends their survival a few days, until selection, which refers to the event where a cohort of small-medium sized follicles can be stimulated by LH to advance to the large ovulatory size.

| METHODS FOR ASSESSING FOLLICLES IN PIGS
Assessment for numbers and sizes of follicles in different stages of reproduction in the pig has been performed using various combinations of collection of the reproductive tract, surgery, and ultrasound.
Use of ultrasound for assessment of follicles in live pigs has enabled repeated measures and links with endocrinology, estrus, ovulation, and fertility. The ability to accurately count and measure follicles >1 mm can vary based on the quality of the equipment, frequency of the transducer, and distance and level of interference to the target. In female pigs, ultrasound is capable of accurately measuring follicles to 0.1-0.5 mm (Ryan et al., 1994;Soede et al., 1998) while reliably counting follicles >2 mm in size (Gonzalez-Añover et al., 2009). At estrus, ultrasound counts of follicles ≥4 mm (Soede et al., 1992) or 6-10 mm (Bolarin et al., 2009) are both highly related to ovulation rate (OR). Yet at selection, which follicles will ovulate may not be evident by size alone, as variation in follicle cell numbers, metabolic activity, and atresia are reported. Perhaps 3-D color doppler sonography with automated analysis may be useful to study the processes related to selection when distinguishing individual follicles by size volume, blood vessels and flow (Jokubkiene et al., 2006;Peres Fagundes et al., 2017).

| FOLLICLE STRUCTURE AND CLASSIFICATION
Follicles during the reproductive cycle have been described for the pig (Guthrie et al., 1995b;Knox, 2005;Schwarz et al., 2008).
Primordial follicles remain quiescent, until stimulated or released from a local inhibiting factor (Guthrie, 2005;Matsuda et al., 2012) when they advance to the secondary stage with the oocyte surrounded by a few layers of granulosa cells (GC). At the tertiary stage, follicles form a basement membrane, theca cell (TC) layers, capillaries (Shimizu et al., 2003), and an antrum, which increases in size with fluid accumulation driven by an osmotic gradient (Binelli & Murphy, 2009). Follicle size classifications in the pig vary widely, but for this review, the following class sizes (mm) will be used; Small (S, <3.0), Medium 1 (M1, 3.0-4.9), Medium 2 (M2, 5.0-6.49), and Large (L, ≥6.5). Growth of follicles to the antral stage requires approximately 80 days without FSH support, while three more weeks are required before they reach ovulatory size (Guthrie, 2005;McGee & Hsueh, 2000). Follicles first appear on the surface of the ovary midway through the prepubertal (PP) period (Guthrie & Garrett, 2001;Prunier et al., 1993), but thereafter, size variation is evident among gilts (Bolamba et al., 1994). While gilts still remain PP, follicle size is dynamic, with largest size sometimes changing by week (Driancourt, 2001;Grasso et al., 1988) in groups of 40-60 follicles (Schwarz et al., 2013). While follicles ≥6 mm were identified in PP gilts, their numbers are few and they fail to ovulate, and soon regress.
However, in mature females, selected follicle growth is considered linear, and Soede et al. (1998) reported average follicle size at weaning was 3.0 mm, reaching 6.0 mm by day 3, and 7.0 mm at estrus. These and other data indicate similar follicle sizes at the start of the follicle phase in weaned sows (Kunavongkrit et al., 1982;Lucy et al., 2001), but variation before and at time of selection has been associated with variation in the wean to estrus interval (WEI) (Bracken et al., 2006;Lopes et al., 2020). The cause of the follicle variation is uncertain, but could link to physiological response to in management or environment. Whether these types of responses determine heterogeneity in follicles during the follicle phase (Knox, 2019), and in GC and estrogen content at estrus is uncertain (Hunter et al., 1989). Because it is not clear which follicles will ovulate, different studies include counts for all follicles >4, >5, or >6 mm, and interestingly, each have been associated with eventual OR. Nevertheless, at estrus, the average size of the largest follicles is most often 6.5-8.0 mm.

| ENDOCRINE CONTROL OF FOLLICLE DEVELOPMENT
Once follicles reach the antral stage, GC must synthesize FSH receptors (FSHR) and bind circulating FSH. Number of FSHR and LHR on each cell are in the thousands for the pig (Gebarowska et al., 1997) and can change 2-10-fold, depending upon size and maturity. Liu et al. (2000) reported FSHR are greatest in 2 mm, decline by 4 mm, and are undetectable in 6-8 mm follicles. However, Cárdenas and Pope (2002) noted FSHR were similar among the largest follicles during the late luteal to mid-follicle phase, and would suggest FSHR could be present in selected M1-M2 follicles into the mid-follicle phase. In contrast, LHR increase in 2-6 mm follicles, before a decline in the 8 mm size. Overall, dependence on FSH for S follicles is clear, with co-dependence on FSH/LH for M1 indicated, and reliance on LH for selected M2-L follicles.
Follicle recruitment appears as a continuous process in pigs and may allow for a predetermined number of follicles to be selected at any time physiological conditions allow. The numbers and sizes of follicles counted and reported may not be indicative of potential ovulation, as differences at selection between healthy and atretic follicles may be minimal (≤1.0 mm). While most follicles <1 mm are healthy, atresia occurs in the 1-5 mm size (Dufour et al., 1985). After ovulation on d 0, all surface follicles disappear, but by day 2, a new wave of healthy follicles classified as estrogen active emerge. But by day 7, many are atretic and no longer estrogenic (Guthrie et al., 1995a). Later in the luteal phase, as recruitment progresses, atresia remains at approximately 35%, but steroidogenesis is low in both healthy and atretic follicles. Guthrie and Cooper (1996) reported that in the pig, wave-type dominant follicle development, with suppression of subordinate follicles, occurs only during the follicle phase, and is not evident based on changes in follicle size, numbers, and atresia in the luteal phase. While FSH does not exhibit wave like patterns, some females do display unique patterns, suggestive of feedback control, but these would be masked when averaged among different females (Knox et al., 2003). Further, if recruitment is continuous, then on any given day, mixed populations of growing and regressing follicles would make waves difficult to discern. At selection, approximately 65% of S-M follicles are healthy, but soon after, nonselected follicles quickly begin to degenerate. An illustration of the anatomy of a follicle, and changes in antrum size, cells and blood vessels with increasing development is shown (Figure 1 (Table 1). As healthy follicles develop in size from S to L, GC and TC numbers increase, but the number of cell layers and their thickness does not, and appears to range from 7 to 12 for each (Fricke et al., 1996). The oocyte has been reported to increase only slightly in size as the follicle grows from the S to the M2 size, but with little change thereafter (Chiou et al., 2004;Hunter, 2000). Further, the quality and viability of oocytes has been linked to the density and numbers of layers (3 to >5) of the surrounding cumulus cells and with an increase in steroid concentration within the FF (Costermans et al., 2020a).
F I G U R E 1 Labeled structures (arrows) of healthy follicles as they grow from small (2-3 mm), to medium (4-6 mm) and large ovulatory size (7-8 mm) in the pig. The figure depicts an increase in the diameter of the follicle antrum and oocyte, an increase in granulosa and theca cells, and blood vessels F I G U R E 2 Example of antral follicle development during lactation in cohorts (1-9) of small (1-3 mm) follicles, and the timeline for their growth to medium 1 (4-5 mm), medium 2 (5-6 mm) size, before atresia or selection for ovulation (7-8 mm) relative to weaning. The figure is based on a model of continuous recruitment and a 6-7-day follicle lifespan once they emerge on the surface of the ovary at 1-2 mm. Healthy follicles (open circles) grow in size (mm) by day, until atresia (vertical hatched circles) and their regression (dashed line) or selection as healthy, ovulatory sized follicles (solid black circles). For simplicity, cohorts are shown in groups of three identical follicles, but variation in size and health would be expected as they develop. Heterogeneity in follicle size and health on any day could be explained by the different cohorts present FSH has been associated with higher OR in PP and cyclic gilts Knox et al., 2003), while other studies have not detected similar relationships (Hunter et al., 1996;Mariscal et al., 1998;Wise et al., 2001). Discrepancies among studies could be related to variation among individuals and time period for analysis.
But other factors also impact FSH and LaVoie (2017)  shown and can rescue follicles, which would also match the response of pigs to exogenous eCG (Guthrie, 2005). This hormone can induce angiogenic factors and growth of 4-5 mm follicles (Shimizu et al., 2003) and delay apoptosis in S-M follicles (Liu et al., 2003).
That selectable follicles are initially dependent on FSH and soon after LH, can be interpreted from studies where purified FSH increases numbers of S-M follicles within 48 h, but these disappear soon after without advancing to ovulatory size unless adequate LH is provided (Breen & Knox, 2012;Guthrie, 2005).

Follicle development is regulated by inhibin feedback on FSH.
Suppression of FSH for short periods before or into the early follicle phase, using steroid free FF containing inhibin, reduces numbers of small and medium sized follicles soon after FSH suppression (Bracken et al., 2006;Guthrie, 2005). But subsequent effects on estrus or OR do not occur when performed in the early follicle phase, as hormones and follicles readjust quickly. However, if FSH is reduced too late into the follicle phase, then follicle rescue cannot occur, resulting in delays in estrus and reduction in OR. After the start of the follicle phase, estrogen (E2) increases in circulation (Flowers et al., 1991), and will suppresses LH to limit the numbers of medium follicles selected (Guthrie et al., 1987). Overall, the data suggest that at selection, healthy S-M follicles bind circulating FSH/LH, produce E2 and inhibin, which suppresses gonadotropins in circulation, to prevent less advanced follicles from competing. Sato (2015) indicated that at T A B L E 1 Changes in relative characteristics of healthy follicles by size and days during the follicle phase in the pig for ovulation. The primary factor involved in angiogenesis around the follicle is vascular endothelial growth factor (VEGF), and in cattle, fibroblast growth factor, insulin-like growth factor, and angiopoietin are involved as well (Berisha et al., 2016). In hens, it has been reported that before selection, VEGF from follicles is regulated by autocrine and paracrine growth factors, but after selection, gonadotropins assume control (Kim et al., 2016). Hunter et al. (2004) indicated that GC are the major source of VEGF, which increases and accumulates in FF. TCs express receptors for VEGF, and a gradient forms to stimulate vessel development to the follicle. Vascularity can be dense just outside the basement membrane, and disperse in the theca interna and externa layers (Barboni et al., 2004). Gonadotropin stimulation in PP gilts induced selection of some follicles for ovulation, and atresia in others. For those follicles selected, angiogenesis was active and increasing, while for those undergoing atresia, blood vessel development was low (Jiang et al., 2004).  (2017) reported GC not only produce but also bind E2 through its receptors. Estrogen synthesis from S-M follicles occurs at low levels (not estrogen active), during the early luteal phase, but is still greater compared to production by these same size follicles in the mid-late luteal phase (Guthrie & Cooper, 1996). With progression of days during the follicle phase, E2 production increases with follicle size (Liu et al., 2000), but not for S follicles which are estrogen inactive, and only for a limited proportion of medium follicles, but does include all large follicles (Guthrie et al., 1993). Linking estrogen active follicles with health is evident in the follicle phase and correlates with LHR (Liu et al., 2000), but does apply very well in the luteal phase. It is also interesting to note that FF concentration of steroids (E2) may not reflect levels in circulation in the early follicle phase (Costermans et al., 2019), in comparison to other steroids in the luteal phase (Naskar et al., 2016) and E2 in later follicle phase. It is not clear what might influence the results, but could suggest a different mechanism regulating synthesis and release of E2 into circulation. Another regulator of follicles includes IGF which is synthesized by GC through the FSH pathway to affect steroidogenesis and cell proliferation (LaVoie, 2017). While IGF is present in all size follicles, its bioavailability in free form, increases with follicle size during the follicle phase as a result of reduced production of IGF binding protein 2 (IGFBP) by GC and TC (Guthrie, 2005).

| CONTROL OF FOLLICLE DEVELOPMENT
Different pathways for regulating FSH/LH provides a mechanism for adaptive responses to physiological conditions to control reproduction.
FSH and LH synthesis and release are regulated by neural and feedback control of the hypothalamic pituitary axis (HPX). LH in circulation is determined by pulses of GnRH and ovarian feedback of E2 and progesterone to stimulate or inhibit the HPX (Barb et al., 2001;Britt et al., 1991). FSH is also controlled by GnRH, but also by inhibin and perhaps E2 feedback at the pituitary (McNeilly et al., 2003). While GnRH neurons do not possess E2 receptors, other neurons do, and are involved in regulating GnRH (Lents, 2019;Scott et al., 2018). After ovulation and disappearance of all surface follicles, the absence of inhibin allows for the second surge of FSH, which precedes the appearance of a new wave of small follicles on day 2. FSH remains elevated with patterns unique in some animals during the luteal phase, and might suggest inhibin regulation, but when values are averaged among females, patterns and relationships may not be revealed (Knox et al., 2003). Nevertheless, at the start of the follicle phase, FSH remains elevated for 24-36 h before inhibin increases and FSH declines, and follicle size and estrogen increase (Knox et al., 2003;Noguchi et al., 2010). Due to pulsatile release of LH, changes are often subtle or difficult to detect (Esbenshade et al., 1982;Guthrie & Bolt, 1990;Guthrie, 2005), but when detected, can be classified as low frequency-high amplitude resulting in low concentration during the luteal phase, and changing to high frequency-low amplitude pulses, with a slight increase in mean concentration during the follicle phase (Flowers et al., 1991;Soede et al., 2011;van Leeuwen et al., 2015).
LH is inhibited by negative E2 feedback in the PP gilt, by negative P4 feedback in the cyclic gilt, and by endogenous opioids suppressing GnRH pulses in lactating and nursing sows (Soede et al., 2011). But when these inhibitors are removed, rapid changes in LH pulses and stimulation at the ovary, quickly select the most sensitive follicles for development.

| METHODS FOR ASSESSING FOLLICLE DEVELOPMENT IN PIGS
Approaches to study factors affecting follicle development in pigs have used weaning, gonadotropins, or progestogen to synchronize the start a follicle phase in groups of females. The synthetic progestogen altrenogest (ALT) simulates an extended luteal phase, prevents the decline of S follicles, reduces LH, and prevents ovulatory follicle growth (Guthrie & Bolt, 1990). Using this approach, in the mid-follicle phase, healthy medium follicles were shown to be consistent in inhibin production, while varying greatly in their estrogen activity (Guthrie et al., 1993). Perhaps related to heterogeneity, studies have noted increased fertility in pigs synchronized with ALT, with the effects associated with an increase in follicle size before or at weaning (Soede et al., 2007a;van Leeuwen et al., 2010), more preovulatory follicles (Am-In & Kirkwood, 2019), or increased follicle size at estrus (Lopes et al., 2017;van Leeuwen et al., 2010). Perhaps variation in follicle development is related to whether females can generate high frequency LH pulses, as demonstrated by van Leeuwen et al. (2015), which could also be influenced by E2 active follicles and feedback to the HPX. In this context, information on the developmental status of the follicles in addition to counts and size measures appears critical to understanding the extent of variation in the follicles before selection.
Administration of the gonadotropin, eCG, with predominant FSH activity, quickly rescues or selects FSH responsive follicles to grow. In contrast, hCG, with predominant LH activity, stimulates LH responsive follicles only to advance to ovulation. Providing eCG/hCG in a single or sequenced injection, stimulates follicles having the appropriate receptors over a period of days before clearance. Use of eCG/hCG is effective for induction of estrus and ovulation but also can display dose-dependent effects on OR and cysts, without improving estrus (Breen et al., 2006). If differences in follicle size before or during lead to heterogeneity at estrus, perhaps the model could include cysts. For example, cystic follicles (>12.9 mm) are not evident at the start of the follicle phase, but can be induced by improper dose, bioactivity, or timing of eCG. Cysts often only develop in a few follicles and become evident midway into the follicle phase (Breen et al., 2006;Manjarín et al., 2015;Ziecik et al., 2017). These studies illustrate heterogeneity in follicle response to eCG and suggest other factors besides follicle size in the cystic response.

| FACTORS AFFECTING FOLLICLES IN PIGS
Questions remain as to how various internal or external cues influence follicles and fertility in pigs. Some effects appear early in development as lighter birthweight gilts show lower numbers of S-M follicles before puberty (Almeida et al., 2017). In some studies, genotypes with greater OR have greater FSH before puberty and during the cycle Knox et al., 2003). Greater prolificacy has also been associated with females having more estrogen active M2 follicles persisting later into the follicle phase (Yen et al., 2005), or with sows developing ovulatory follicles over a longer period of the follicle phase (Driancourt & Terqui, 1996). This may be most relevant in young hyper-prolific sows nursing large litters, that have adapted by altering their endocrine and follicle physiology (Kemp & Soede, 2012). But in the subpopulation of females unable to adapt, failures in follicle development translates into delays in estrus and breeding (Lucy et al., 2001).
Boar exposure is important for expression of estrus, and can advance puberty in gilts and reduce WEIs in sows. However, an immediate induction of a follicle phase is obvious in only a small population of gilts (Patterson et al., 2002) and primiparous sows (Langendijk et al., 2006). Some studies show immediate effects on LH in a few females (Kingsbury & Rawlings, 1993;van de Wiel & Booman, 1993), but the response does not associate with a follicle phase and estrus. However, in susceptible females, boar stimuli can improve follicle size (Langendijk et al., 2000), and ovulation (Ulguim et al., 2018), and while the pathway remains unclear, evaluation for the effects of intensity, duration, and interval could be revealing. The effects of feed intake on follicles in pigs has been reviewed (Quesnel, 2009), with most of the attention focused on primiparous sows, due to their predisposition for negative energy imbalance during lactation and delayed estrus postweaning. Nutrient deficit reduces GnRH pulses, FSH/LH, and number, health and size of follicles (Kauffold et al., 2008;Prunier & Quesnel, 2000). Hazeleger et al. (2005) reviewed the effects of nutrition during and after lactation and reported reduced follicle size at weaning often observed, with effects linked to reduced LH during lactation and after weaning, and IGF within the follicle. Several studies support these observations with feed effects impacting the distribution of follicles, IGF-1, or steroids in sows (Costermans et al., 2020b;Han et al., 2020), gilts (Chen et al., 2012) and with effects on LH (Zhou et al., 2010). The data suggest effects at the HPX alter FSH/LH, and with direct effects in the follicles through IGF-1 in the weeks before and up to selection.
Seasonal delay in estrus is common in commercial farms around the world (Peltoniemi & Virolainen, 2006), but linking the effects of temperature and photoperiod to changes in follicles and estrus is less clear (Canaday et al., 2013;Williams et al., 2013). Data suggesting HPX involvement are inconsistent with respect to differential response to E2 on FSH/LH or estrus (Cox et al., 1987;Kennaway et al., 2015).
Nevertheless, effects of season or heat stress have been noted on follicles and estrus (Arend et al., 2019;Lopes et al., 2014) or on progesterone production in vivo (Bertoldo et al., 2011) or in vitro (Sirotkin, 2010). Photoperiod effects on follicles are undefined and perhaps may result through effects on the HPX (Peltoniemi & Virolainen, 2006), although Kermabon et al. (1995) did not detect changes in FSH/LH by season. But fertility effects of follicles may be mediated through melatonin (Arend et al., 2019) which declines with follicle size in the pig (Shi et al., 2009) and can increase GC synthesis of E2 (Lv et al., 2019). Stress effects on fertility in pigs and on follicles, may depend upon the type and duration of stress, stage of reproduction, and susceptibility of the female. Effects of stress on follicles, suggest ACTH and cortisol involvement in HPX function, leading to reduced LH, delayed follicle development, and problems in estrus, ovulation and cysts (Einarsson et al., 2008). But cortisol is present in normal adaptive situations, and direct links to problems with follicles are not clear (Turner & Tilbrook, 2006). Direct effects of cortisol on follicle steroidogenesis and cysts are also possible (Madej et al., 2009;Ryan et al., 1990). However, negative stressors occurring at daily intervals during the follicle phase, while temporarily increasing cortisol, had no detectable effects on follicle development, estrus or ovulation (Soede et al., 2007b;Turner et al., 2005). Perhaps attention may need to be focused on extending the duration and frequency of the stress into a chronic response to observe effects on follicles and fertility.

| CONCLUSIONS ON FOLLICLE DEVELOPMENT
The available information suggests that reduced follicle health in populations that are less or too advanced at the start of a follicle phase, can be linked to problems in estrus, ovulation, and OR. The

CONFLICT OF INTEREST
The author declares no conflict of interest.

DATA AVAILABILITY STATEMENT
The information listed is from referenced library sources or our own creation. The data that support the findings of this study are openly available in [repository name] at [DOI].

PEER REVIEW
The peer review history for this article is available at https://publons. com/publon/10.1002/mrd.23576