Beyond brood: the potential impacts of insect growth disruptors on the long-term health and performance of honey bee colonies

Honey bees are valued pollinators of agricultural crops, and heavy losses reported by beekeepers have spurred efforts to identify causes. As social insects, threats to honey bees should be assessed by evaluating the effects of stress on the long-term health and productivity of the entire colony. Insect growth disruptors are a class of pesticides encountered by honey bees that target pathways involved in insect development, reproduction, and behavior, and they have been shown to affect critical aspects of all three in honey bees. Therefore, it is imperative that their risks to honey bees be thoroughly evaluated. This review describes the effects of insect growth disruptors on honey bees at the individual and colony levels, highlighting hazards associated with different chemistries, and addresses their potential impacts on the longevity of colonies. Finally, recommendations for the direction of future research to identify strategies to mitigate effects are prescribed.


INTRODUCTION
As managed pollinators, honey bees contribute significantly to the production of agricultural crops throughout the world, and countries including the USA have come to rely upon their services to sustain national food security (Southwick and Southwick 1992;Aizen et al. 2008;Calderone 2012). However, since 2006, beekeepers have been reporting heavy annual losses to managed populations (vanEngelsdorp et al. 2009;Kulhanek et al. 2017;Brodschneider et al. 2018). These losses have placed a financial strain on beekeepers, who must expend more resources to replace their failed colonies (DeGrandi-Hoffman et al. 2019). It is predicted that continued losses will result in food shortages due to the limited number of bee colonies in contrast with the agricultural acreage of pollination-dependent crops required to feed expanding human populations (Aizen et al. 2008;Aizen and Harder 2009). For this reason, identifying and mitigating risk scenarios to honey bees is critical.
Four major causative factors have been identified as contributors to honey bee population declines, including exposure to infectious pathogens, parasite infestations, poor quality or limited nutritional resources, and exposure to agrochemicals (Spivak et al. 2011;Goulson et al. 2015). This review will address exposure to agrochemicals, focusing on insect growth disruptors (IGDs), a pesticide class that is known to affect pathways associated with insect growth and development (Pener and Dhadialla 2012). While the effects of IGDs on immature insects are fairly well understood (Wright 1976;Dhadialla et al. 1998;Tunaz and Uygun 2004), effects on adult insects are discussed less frequently. This review will cover two broad categories of IGDs, chitin synthesis inhibitors (CSI), and insect hormone mimics (IHM), examining their effects on honey bees and discussing gaps in the current literature relating to the effects of IGDs on honey bee reproduction and colony performance. Based on the summary of the literature, mode of pesticidal action, and the current understanding of honey bee social biology, the impacts of IGDs beyond brood loss, and their potential to impact colony-level performance will be discussed. Finally, future directions for investigative work that may lead to the identification of mitigation strategies will be recommended.

IGDS AND HONEY BEES
The term IGD refers to a broad class of chemicals that disrupt insect growth, development, and reproduction (Pener and Dhadialla 2012) by acting on the insect endocrine system to interfere with metamorphosis and developmental processes (Williams 1967;Hoffmann and Lorenz 1998;Jindra and Bittova 2020) or by acting as chitin synthesis inhibitors to impair the formation of new cuticle (Grosscurt 1978;Cutler and Scott-Dupree 2007;Matsumura 2010). Because of their selectively lethal activity towards immature insects and their low lethality to beneficial adult insects such as pollinators, IGDs are often considered lower risk relative to other more traditional chemistries such as neurotoxins (Wright 1976), and they are frequently preferred over other insecticides in agricultural settings where beneficial insects are present. One particularly high-profile example of this use pattern is in blooming almond orchards in California (Mullin et al. 2016;CalDPR 2019;Wade et al. 2019).
The almond bloom in California is the largest pollination event in the world, requiring two thirds of the managed colonies in the USA to be transported, often across thousands of miles, to orchards in the Central Valley where they remain for weeks during the blooming period. It is estimated that the almond production industry adds over $20 billion to the economy of California (Sumner et al. 2016), and it has become a critical event for commercial beekeepers, who rely upon the premium fee almond growers pay for hive rentals during almond bloom (Rucker et al. 2012). Part of the reason this premium is offered is the stress on colonies associated with pollination events (Ahn et al. 2012;Zhu et al. 2014;Cavigli et al. 2016;Simone-Finstrom et al. 2016).
During the bloom period, growers often apply various agrochemicals to control insects and plant pathogens. Tank mixtures typically contain IGDs, fungicides, and adjuvants (Mullin et al. 2016;Wade et al. 2019) and are applied at night when bees are not actively foraging. However, the residues of these agrochemicals persist in the environment. Foraging bees encountering these residues in pollen, nectar, and other relevant matrices can experience negative effects by direct exposure, but foragers may also return to the colony with the contaminated resources (Mullin et al. 2010;Traynor et al. 2016). This can result in a systemic exposure scenario in which other members of the colony such as the nurse bees, reproductive castes, and developing brood may be exposed (Berenbaum and Liao 2019). In recent years, some beekeepers have reported colony loss characterized by heavy brood loss following almond pollination (Pollinator Stewardship Council 2014). While other factors likely contribute to this phenomenon, including exposure to fungicides (Mao et al. 2017), adjuvants (Mullin et al. 2016), andpathogens (Cavigli et al. 2016;Fine et al. 2017a), IGD toxicity remains a likely explanation due to their known effects on immature insects. While the effects on developing brood may be the most obvious sign of potential IGD poisoning in a colony, beekeepers should be aware of the potential effects these chemistries may be exerting on other members of the colony, which may be more difficult to diagnose and treat ( Figure 1).

CHITIN SYNTHESIS INHIBITORS
Chitin is a critical component of insect cuticle, but it also has functions in insects that are often overlooked. For example, chitin is a major component of the peritrophic membrane (PM) lining the insect gut. The PM is critical to ensuring nutrient absorption while protecting insects from Beyond brood: the effects of IGDs on honey bees invading pathogens (Merzendorfer and Zimoch 2003). In honey bee larvae, the PM acts as a barrier to prevent infection with the American Foulbrood causative agent, Paenibacillus larvae. Once the PM is fully formed 48 h after larval eclosion (Davidson 1970), larvae have increased immunity to infection by this pathogen (Brødsgaard et al. 1998;Yue et al. 2008). Similarly, chitin is a component of cocoons that holometabolous insects, including honey bees, produce during pupation (Moussian 2019). The residual cocoon material left after adult eclosion is thought to increase homeostasis in the colony by buffering humidity (Ellis et al. 2010).
Chitin synthesis inhibitors (CSIs) are IGDs designed to target and disrupt pathways associated with the formation of insect cuticle (Matsumura 2010). Typically, this results in the mortality of immature insects due to an unsuccessful molt following oral or topical exposure. However, CSIs can also affect oviposition (Moffitt et al. 1983;Cutler et al. 2005) and embryogenesis (Henzell et al. 1979;Medina et al. 2002) in reproductive adult insects. Due to their low adult lethality, CSIs are often applied to blooming crops, though this practice is discouraged when bees are actively foraging due largely to their toxicity to honey bee brood. Currently, the CSI diflubenzuron remains one of the most commonly applied agrochemicals in California almond orchards during the weeks when honey bee colonies are foraging (CalDPR 2019; Wade et al. 2019).
While the disruption of chitin formation may not have well-established effects on adult worker honey bees, exposure to CSIs can be lethal to immature honey bees as they develop from larvae to pupae and adults (Fine et al. 2017b;Wade et al. 2019). Similarly, chitin formation is critical Figure 1 IGD exposure has been found to affect numerous aspects of adult worker and queen physiology and behavior in the following ways: A.) accelerating behavioral maturation (Jaycox et al. 1976;Jaycox et al. 1974;Robinson 1985), reducing lifespan, and impairing foraging performance of worker honey bees (Perry et al. 2015;Chang et al. 2015;Fisher et al. 2018;Meikle et al. 2019); B.) impairing learning and memory (Abramson et al. 2004) and increasing aggressive responses to stimuli (Robinson 1985); C.) HPG development and royal jelly protein production (Gupta and Chandel 1995;Chen et al. 2016); D.) vitellogenin synthesis in the fat body and uptake into other tissues (Pinto et al. 2000); E.) oogenesis and embryogenesis (Thompson et al. 2005a;Milchreit et al. 2016;Fine 2020). This figure was created using Photoshop CC 2021 (Adobe Inc., San Jose, CA). during embryogenesis when the first instar larval cuticle is formed (Moussian et al. 2005;Moussian 2019), and chitin is a major component of insect egg chorion. A thin or poorly formed egg cuticle could result in detrimental effects including the dehydration of the embryo (Jacobs et al. 2013), impaired respiration, and hatching (Hinton 1969). Maternal exposure to CSIs has been observed to have embryotoxic effects in insects (Henzell et al. 1979;Medina et al. 2002), yet surprisingly few studies have explored this as a possibility in honey bees (Milchreit et al. 2016;Chen et al. 2016;Fine et al. 2017b;Fine 2020).

INSECT HORMONE MIMICS
Insect hormone mimics (IHM) are compounds that act on the insect endocrine system to disrupt the process of insect metamorphosis (Wright 1976;Hoffmann and Lorenz 1998;Jindra and Bittova 2020). Like CSIs, IHMs result in the mortality of immature insects by interfering with the hormone balance required for successful ecdysis, often blocking molting or triggering it prematurely, resulting in death (Dhadialla et al. 1998;Retnakaran et al. 2003;Jindra and Bittova 2020). The most common signaling pathways targeted by IHMs are those normally mediated by juvenile hormone (JH) and ecdysone. Despite the established role of these pathways in mediating critical aspects of adult bee physiology and behavior (Gilbert et al. 1980;Bloch et al. 2013;Pandey and Bloch 2015), IHMs exhibit low lethal toxicity to adults and are widely considered to be safer to beneficial insects (Dhadialla et al. 1998). Because of this, the use of IHMs such as the ecdysone agonist, methoxyfenozide, or the JH analogue, pyriproxyfen, in blooming, beepollinated crops like almonds is encouraged relative to other, more acutely toxic pesticides.
In immature insects, growth and development are regulated by JH and ecdysteroids (Weir 1970;Nijhout 1998;Smagghe 2009;Riddiford 2012). Generally, JH titers in the hemolymph of immature insects are higher between molting periods, but dip when the larva has reached a critical weight, triggering ecdysone levels to rise and induce a molt (Mirth et al. 2014). This interaction is more complex in the latter stages of holometabolous insect development, when JH tightly regulates the action of ecdysone and the timing of the pupal and adult molt (Hartfelder 2000;Zhou and Riddiford 2002;Dubrovsky 2005). Interfering with this process, either by administering artificial hormone analogues or compounds that antagonize their production, produces deleterious effects in immature insects. Although the sublethal effects of IHMs on developing brood are not well studied, JH and ecdysteroid levels influence ovary development and caste differentiation in immature honey bees (Hartfelder et al. 1995;Capella and Hartfelder 1998;Elekonich et al. 2003;Amdam et al. 2010), and the presence of ecdysteroids in honey bee queen ovary tissue and in embryos suggests a role in embryogenesis (Feldlaufer et al. 1986;Mello et al. 2014). Therefore, it is likely that exposure to low levels of IHMs during embryogenesis and subsequent development could influence the reproductive potential, health, and physiology of honey bee queens and workers.
In adult insects, JH regulates oogenesis as well as numerous behaviors associated with reproduction Shu et al. 1998;Webb et al. 1999), and JH titers are typically positively correlated with vitellogenin (Vg) levels (Hagedorn and Kunkel 1979;Bellés 2005). However, in honey bees, this relationship is inverse. Young honey bee workers have higher hemolymph titers of Vg and low levels of JH, but as workers age and transition to foraging tasks, JH levels rise and Vg declines (Robinson 1987;Robinson et al. 1991;Pinto et al. 2000). In mated queens, JH levels are generally low when the queen is engaged in oogenesis, but are higher in unmated, previtellogenic queens or queens that are prevented from laying by caging, and JH suppresses Vg synthesis and uptake into tissues (Pinto et al. 2000;Wegener et al. 2013). The antioxidant properties of Vg and its relationship with JH are thought to be important regulators of honey bee lifespan (Amdam et al. 2004), as evidenced by the senescence of worker bees when JH levels are high and the longer lifespan of queens, whose Vg levels remain high relative to workers for at least a year (Corona et al. 2007) or while she is actively laying eggs (Wegener et al. 2013). In worker bees, JH also regulates behavioral Beyond brood: the effects of IGDs on honey bees maturation, and increases in JH titers result in increased aggression and the onset of foraging behaviors (Robinson 1985).
Similar to JH, ecdysteroids influence aspects of adult insect behavior and physiology including ovariole development, longevity, circadian rhythm, and learning behaviors (Simon et al. 2003;Ishimoto et al. 2009;Ishimoto and Kitamoto 2010;Ables et al. 2016). In adult honey bees, ecdysteroids are primarily synthesized by the ovarian tissue and are involved in oocyte maturation and embryogenesis (Robinson et al. 1991;Amdam et al. 2010;Wegener et al. 2013). In worker honey bees, ecdysteroids are known to affect hypopharyngeal gland (HPG) development and protein synthesis (Corby-Harris et al. 2016, 2019Winkler et al. 2018), suggesting they be involved in nutrient signaling and behavioral maturation. Additionally, ecdysteroids influence brain development and aversive learning in workers (Geddes et al. 2013), further supporting their role in regulating behavior.
Because insect development is regulated by the hormone pathways targeted by IHMs, these IGDs are often highly toxic to immature insects. However, while disrupting hormonal pathways in adult insects may not be deleterious, insect endocrine signaling regulates numerous behaviors and aspects of physiology in adult insects. Without evaluation, it should not be assumed that nonlethal doses do not have profound impacts on adult honey bees, and by extension, on the entire colony.

EFFECTS OF IGDS ON HONEY BEE REPRODUCTION
By far, the most well-characterized effects of IGDs on honey bees relate to reproduction (Figure 2), specifically, their detrimental effects on brood rearing (last reviewed by Tasei 2001). CSIs are highly toxic to honey bee brood in laboratory (Fine et al. 2017b;Wade et al. 2019) and colony-level experiments (Barker and Waller 1978;Emmett and Archer 1980;Thompson et al. 2005a;Fine et al. 2017b), so much so that the CSI triflumuron was once suggested as a means to control Varroa mite infestation by completely disrupting the brood cycle of a colony . In this study, free flying colonies fed a solution of 150 ppm triflumuron reared no brood following the treatment, but 22 days later, the colony appeared to have recovered. Interestingly, in caged colonies, triflumuron did not disrupt queen egg laying, but when placed in foster colonies, the eggs were removed by workers, suggesting they were not viable. Similar findings have been reported for the CSIs diflubenzuron and novaluron, which can both disrupt colony-level brood production at high enough doses (Barker and Waller 1978; Thompson et al. 2005a;Fine et al. 2017b). The deleterious effects on honey bee brood caused by exposure to IHMs have been documented in both laboratory (Zufelato et al. 2000) and field experiments (Chen et al. 2016), and sublethal effects on bees that survive larval exposure including wing deformities and reduced lifespan have been described (Fernandez et al. 2012;Fourrier et al. 2015). In addition to directly affecting brood, the modes of action of IHMs suggest that they are likely to affect aspects of adult physiology and behavior that contribute to brood rearing capacity. For example, JH and ecdysteroids may influence Vg synthesis and uptake in adult queen and worker bees (Barchuk et al. 2002;Wegener et al. 2009Wegener et al. , 2013, which could, in turn, influence the production of royal jelly and brood food by the HPGs and oocyte maturation (Amdam et al. 2003;Seehuus et al. 2007). Indeed, the JH analogue, pyriproxyfen, inhibits Vg synthesis in worker bees (Pinto et al. 2000) and decreases royal jelly production at the colony level (Chen et al. 2016). In contrast, chronic consumption of an ecdysone agonist, methoxyfenozide, has been shown to affect HPG size in 2-week-old worker bees, resulting in larger glands relative to similarly aged controls (Fine 2020). In the field, methoxyfenozide exposure resulted in decreased forager activity, which could be a consequence of delayed or altered behavioral maturation (Meikle et al. 2019).
While the majority of experiments do not specifically assess the effects of IGDs on honey bee queen performance, a small number of studies have addressed the possibility. Barker and Waller (1978) observed that colonies treated with diflubenzuron had more eggs present relative to control following a 40-day exposure period, which was attributed to a transovarial effect on embryo viability, resulting in a high number of inviable eggs remaining in the colony, or a compensatory increase in egg production to account for the loss of brood due to the diflubenzuron treatment. By closely monitoring and manipulating queens in full-sized colonies, Chen et al. (2016) found a negative effect on the rate of first instar eclosion from eggs laid under exposure to JH analogue pyriproxyfen. Using a targeted, cross-fostering approach, Milchreit et al. (2016) demonstrated the transovarial toxicity of a JH agonist to honey bees by showing that maternal exposure to fenoxycarb negatively affects first instar eclosion and survival to adult eclosion. Thompson et al. (2005) reported an effect of the same JH agonist on mating success and egg production, though the viability of the eggs produced during this experiment was not specifically assessed. In the same study, sperm counts of mature drones reared under exposure to CSIs and IHMs were assessed. No significant differences were detected among treatments, though the availability of drones from colonies treated with the CSI diflubenzuron and IHM fenoxycarb was negatively affected by the treatment related heavy brood loss. The CSI novaluron has also been found to disrupt brood production in honey bee colonies and studies on honey bees (Fine et al. 2017b) and alfalfa leaf-cutter bees (Hodgson et al. 2011;Pitts-Singer and Barbour 2017) suggest it may have a dramatic effect on embryogenesis. In a semi-field-level study performed on nucleus colonies, it was demonstrated that exposure to 18.6 ppm novaluron for a period of 7 days resulted in no apparent decreases in queen egg laying rates, but no evidence of hatching could be seen in the colonies for up to 21 days following the treatment (Fine et al. 2017b). In a recent experiment involving the use of specialized cages to quantitatively track queen egg laying, no effect on the quantity of eggs produced was documented when queens were exposed through worker diet to novaluron, diflubenzuron, or methoxyfenozide at 10 ppm, Figure 2 The potential effects of IGDs on intra-colony dynamics. Inside a healthy colony (left), brood rearing activities are sustained by a productive queen and a population of young worker bees acting as nurses and queen attendants. Inside an IGD-exposed colony (right), these activities can be compromised due to A.) a decreased number of available and responsive queen attendants due to accelerated behavioral maturation (Perry et al. 2015), B.) fewer viable eggs produced by queens (Milchreit et al. 2016;Chen et al. 2016) and fewer surviving larvae (Zufelato et al. 2000;Thompson et al. 2005;Chen et al. 2016;Fine et al. 2017b), and C.) decreased provisioning of larvae due to D.) fewer nurse bees present in the colony and/or decreased vitellogenin synthesis (Pinto et al. 2000;Perry et al. 2015;Chen et al. 2016). This figure was created using Photoshop CC 2021 (Adobe Inc., San Jose, CA).
Beyond brood: the effects of IGDs on honey bees though a decrease in egg eclosion rate was observed for all three treatments (Fine 2020).

EFFECTS OF IGDS ON WORKER BEHAVIOR AND PHYSIOLOGY
Despite the low lethal toxicity of IGDs to adult honey bees, sublethal effects have been reported, including impaired memory and learning (Abramson et al. 2004) and decreased weight gain and HPG development (Gupta and Chandel 1995). In addition, exposing foragers to certain IGDs at tank mix concentrations reportedly decreases longevity (Fisher et al. 2018). One of the most well-known effects of IGDs is the accelerated behavioral maturation caused by exposure to JH mimics like methoprene, pyriproxyfen, and fenoxycarb (Jaycox et al. 1974;Robinson 1985Robinson , 1987. This phenomenon, known as precocious foraging, in which workers leave the colony to forage at an earlier age, ultimately shortens worker lifespan (Jaycox et al. 1974) and reduces the population of young worker bees inside the colony that support colony-level reproduction (Perry et al. 2015). Bees that precociously transition to forager tasks are not as efficient and do not forage for as long as their naturally aged counterparts (Perry et al. 2015;Chang et al. 2015).
Conversely to JH, the role of ecdysteroids and ecdysteroid agonists is less well-understood, though there is a growing body of evidence suggesting that like JH, ecdysteroids influence behaviors in adult bees (Pandey and Bloch 2015). The ecdysone agonist methoxyfenozide is one of the most commonly used IGDs in blooming almonds. Despite its selective activity for lepidopteran pests (Dhadialla et al. 1998), methoxyfenozide has been shown to influence colony thermoregulation and foraging activity (Meikle et al. 2019). At a higher dose than Meikle et al., increased HPG size has been reported (Fine 2020), similar to what has been documented in workers following treatment with 20-hydroxyecdysone (Corby-Harris et al. 2016).

. M E T A B O L I C C O S T S O F DETOXIFYING IGDS
Even if IGDs are specific to their target pests (Dhadialla et al. 1998), detoxification itself may be costly if it diverts energy away from other metabolic processes or reproduction (Kliot and Ghanim 2012). Numerous examples show that populations selected for resistance to humanmade and plant defense chemicals have reduced fitness (reviewed in (Després et al. 2007;Kliot and Ghanim 2012), but see (Ffrench-Constant and Bass 2017)). Honey bees detoxifying plant-and human-made chemicals do seem to incur metabolic costs. Larval and adult bees exposed to the plant defense compound nicotine and the neonicotinoid imidacloprid show increased expression of genes and proteins related to energy metabolism (Derecka et al. 2013;du Rand et al. 2017), implying a degree of energetic stress. We are unaware of similar studies using IGDs as a chemical stressor, and so it is unclear whether detoxification of this class of pesticide induces metabolic costs. Because honey bees have a limited detoxification repertoire (Claudianos et al. 2006), the same enzymes that detoxify plant defense compounds and neonicotinoids in a metabolically costly manner may work to detoxify IGDs and indirectly reduce bee health.

EFFECTS OF IGDS ON COLONIES
Of the effects of IGDs on brood, workers, and queens, brood loss is the most widely discussed consequence of IGD poisoning. Certainly, brood loss can drastically affect the longevity and performance of a colony, particularly if it is sustained and if the colony is subjected to additional stressors. However, changes to adult behavior like precocious foraging caused by IGD exposure may be a contributor to colony declines over time, even more so than heavy brood loss. Thompson et al. (2005) attempted to study this phenomenon by exposing colonies to fenoxycarb and other IGDs in sucrose at maximum tank mix concentrations for 1 week in the summer and tracking colony performance through the winter to the subsequent spring. Following treatment, declines in brood production were observed, adult populations were low after winter, and the colony build-up in the spring was delayed relative to control and other IGD treatments. The authors hypothesized that this slow build-up was due to the heavy brood loss and potentially a heavy loss of adult populations caused by precocious foraging. Using a modeling approach, it was estimated that precocious foraging caused by exposure to JH analogues like fenoxycarb can have a more pronounced detrimental effect on colony longevity than brood loss (Thompson et al. 2007).
Precocious foraging occurs in response to multiple different categories of stress (Bordier et al. 2017), including nutritional deficits (Free 1961;Schulz et al. 1998;Janmaat and Winston 2000), disease (Higes et al. 2008;Woyciechowski and Moroń 2009;Goblirsch et al. 2013;Benaets et al. 2017), and exposure to agrochemicals (Thompson 2003;Colin et al. 2019). However, evidence of colony-level effects of stress-induced precocious foraging can take months (Thompson et al. 2005;Higes et al. 2008;Khoury et al. 2011;Perry et al. 2015), as the colony may initially be able to buffer the effects of population loss through brood rearing. For this reason, long-term monitoring of colonies after exposure to IGDs may be necessary to detect effects due to behavioral shifts.
Similar to the potentially large-scale effects caused by effects on adult workers, effects on honey bee queens may be far more consequential to the health and performance of the colony over time than effects on honey bee brood. Following exposure to an IGD, brood rearing should recover once a colony has been cleared of residual IGDs in colony matrices. However, effects on honey bee queens could be much longer lasting. Unlike worker bees that typically survive for roughly 4 weeks during the active spring, summer, and fall months (Winston 1991), honey bee queens can live for multiple years, during which time they are solely responsible for producing fertilized eggs inside a honey bee colony (Page and Peng 2001). In commercial operations, beekeepers typically requeen annually as a prophylactic measure or when brood pattern becomes noticeably irregular (Lee et al. 2019). However, the indicators of queen failure could easily be missed, particularly in large-scale operations where beekeepers cannot thoroughly inspect brood frames in every colony, and often the indications come too late, when colony populations decline precipitously due to a shortage of emerging adult bees to replace the aging forager population. One further complication is the recent finding that brood pattern is not always a reliable indicator of queen quality (Lee et al. 2019), highlighting the need for more research to investigate the factors that contribute to queen performance and identify more reliable diagnostic metrics.
Queen failure, as defined by the loss of the queen or impaired queen performance, has been identified by beekeepers as one of the major factors contributing to colony loss (vanEngelsdorp et al. 2013). Negative effects on queen survival, health, and performance have been reported for queens exposed to agrochemicals, both during development (DeGrandi-Hoffman et al. 2013;Williams et al. 2015;Walsh et al. 2020) and as adults (Thompson et al. 2005;Milchreit et al. 2016;Chen et al. 2016;Fine et al. 2017b;Fine 2020). Despite this, the effects of maternal agrochemical exposure in honey bees are seldom examined or considered when evaluating risk scenarios to honey bee colonies. Part of the reason for this could be an underlying assumption that honey bee queens are insulated from the effects of agrochemical exposure on colonies (Purdy 2015). This assumption stems from the concept of a "shared stomach," (Sponsler and Johnson 2017;Schmickl and Karsai 2017) which describes the movement of resources within a colony. This movement involves forager collection of resources, the processing and storing of resources by hive bees, consumption by nurse bees and older larvae, and the conversion of these nutrients into the glandular secretions that make up the queen and larvae's diet (Purdy 2015). At each step during the handling and processing, some degradation of the agrochemical residues is expected to occur through metabolism or environmental transformation. Indeed, there is evidence that even foragers transporting nectar back to the colony in their crop may begin the process of metabolizing xenobiotic chemicals (Vannette et al. 2015). Residues of agrochemicals detected in royal jelly are typically much lower than levels in worker diet (Johnson and Percel 2013;DeGrandi-Hoffman et al. 2015;Böhme et al. 2018), suggesting that nurse bees are capable of efficiently metabolizing agrochemicals in a manner that protects young larvae and queens from exposure to high levels of many agrochemicals. However, even pesticides that do not persist at detectable levels in diet may pose risks to developing brood and reproductive castes through persistence in wax (Wu et al. 2011) or by Beyond brood: the effects of IGDs on honey bees altering the physiology and immune function of their worker caretakers (DeGrandi-Hoffman et al. 2015;Zaluski et al. 2020). As reviewed here, field-and semi-field-level experiments have demonstrated that feeding colonies IGDs can result in brood loss, suggesting that they are not always adequately protected from IGD exposure by the "shared stomach" at the doses studied. Honey bee queens are similarly provisioned by workers with glandular secretions (Crailsheim 1992), and indeed, effects of IGDs on queens have been documented (Thompson et al. 2005;Milchreit et al. 2016;Chen et al. 2016). Given the importance of honey bee queen quality to the longevity and productivity of colonies, it is imperative that the potential effects of IGDs on queens be more thoroughly investigated.

FUTURE DIRECTIONS
Recently, it has been suggested that endocrine disruption by agrochemicals should be considered a critical aspect of risk assessment for pollinators (Christen et al. 2018). The current research pertaining specifically to IGDs supports this, highlighting the need for more detailed assessments of how IGDs impact honey bee reproduction and behavior. Honey bees regularly encounter IGDs in their foraging environments (Mullin et al. 2010;Traynor et al. 2016), yet, for many chemistries, effects on critical aspects of their physiology and behavior and the long-term effects of these shifts on colony productivity and longevity remain poorly understood. One criticism often leveled at toxicological studies involving honey bees is the use of doses or exposure methods that are not realistically comparable to field-level exposure scenarios (Carreck and Ratnieks 2014). To ensure accuracy in assessments, researchers should strive to base their dose and exposure methods on residue data gathered from field assessments. For IGDs, residues in the part per billion range have been detected in various colony matrices including bee bread and wax (Mullin et al. 2010;Traynor et al. 2016;Tong et al. 2018), though some degradation of the parent compounds is expected to have already occurred at the time of sampling in the referenced studies (Coats 1991). Less is known regarding their presence in matrices like royal jelly, but studies demonstrating effects on queens and larvae following hive feedings suggest that residues can persist.
Based on the literature reviewed here, the most pressing areas to be explored by future studies are those that have the potential to impact honey bee queens and drones and worker behavior, longevity, and performance. For queens, these efforts should involve targeted studies examining how maternal exposure to IGDs affects the abundance and quality of offspring. These studies should be focused on determining what scenarios and levels of exposure are likely to affect queens, what the anticipated effects will be, and their expected duration. This information will help to inform recommendations such as when to consider requeening or whether to allow the colony to recover without intervention.
For workers, much is already known regarding the physiological and behavioral effects of IGDs, but few studies have addressed the long-term consequences of these effects using field-relevant exposure scenarios. Ultimately, efforts to understand how stressors like IGDs affect long-term colony performance should be performed over multiple worker generations. As social insects, perturbations to a colony population can have consequences beyond the initial effects on individuals (Thompson et al. 2007;Perry et al. 2015). These effects are most apparent when colonies lose large proportions of nurse bees to precocious foraging, leading to reduced brood rearing, and a subsequent generation that must somehow compensate for the loss with a smaller population.
Research efforts in this area should focus on identifying how IGDs affect honey bee reproduction, physiology, and behavior and attempt to correlate these observations to effects seen in broader longitudinal studies tracking colony performance throughout the entire year under conditions experienced in commercial operations. With these efforts combined, strategies to avoid or ameliorate any potential effects of IGDs could be more readily identified.

ACKNOWLEDGEMENTS
Thank you to Dr. Gloria Degrandi-Hoffman and Dr. Paul Pratt for their helpful reviews of this manuscript. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the US Department of Agriculture. USDA is an equal opportunity provider and employer.
Code availability Not applicable.

AUTHOR CONTRIBUTION
JDF conceived the review, and JDF and VC-H participated in the review and interpretation of the literature, critical analysis, and the writing process.

DECLARATIONS
Ethics approval Not applicable.
Consent to participate Not applicable.

Consent for publication Not applicable
Conflict of interest The authors declare no competing interests.

OPEN ACCESS
This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/.
En plus des effets sur le couvain: les effets potentiels des perturbateurs de croissance des insectes sur la santé et la performance à long terme des colonies d'abeilles.
Nicht nur Effekte auf die Brut: Potentielle Auswirkungen von Insektenwachstumsrhemmern auf d i e L a n g z e i t g e s u n d h e i t u n d L e i s t u n g v o n Honigbienenvölkern.