All animal species possess multiple insulin-related genes, and these genes appear to work in combination to regulate diapause. In contrast, in the species where IIS signaling has been shown to regulate diapause, these ligands appear to signal through a single receptor. Worms and insects have only a single insulin/IGF receptor, and in fishes IGFR1 is the key receptor that regulates diapause, as described below.

The combinatorial nature of the ligands is most apparent in C. elegans, whose genome encodes 40 ILPs. C. elegans ILPs have been studied singly and in combination to reveal a complex network of IIS agonists and antagonists that regulate diapause entry, and a slightly different set that regulates diapause exit (Pierce et al., 2001; Cornils et al., 2011; Fernandes-de-Abreu et al., 2014). ILPs that regulate diapause are expressed in sensory neurons, and their expression changes in response to environmental cues (Pierce et al., 2001; Li et al., 2003; Cornils et al., 2011). The best-studied regulators of dauer diapause include daf-28, ins-6, and ins-1. daf-28 and ins-6 both promote IIS to oppose dauer entry and promote dauer exit; however, daf-28 plays a more prominent role in dauer entry and ins-6 plays a more prominent role in dauer exit (Malone et al., 1996; Li et al., 2003; Cornils et al., 2011). By contrast, ins-1 antagonizes IIS to promote dauer entry and oppose dauer exit (Pierce et al., 2001; Cornils et al., 2011). If the human insulin protein is introduced into C. elegans it also antagonizes IIS, similar to ins-1, demonstrating the conservation between ILPs across species (Pierce et al., 2001). In addition to these three ILPs that have been studied in detail, systematic studies have shown that many other ILPs also play more subtle roles in regulating dauer entry and/or exit (Pierce et al., 2001; Fernandes-de-Abreu et al., 2014). ILP signaling converges on the sole C. elegans IR, DAF-2, which was discovered on the basis of its effect on entry to dauer diapause. Reduction-of-function mutations in daf-2 cause a dauer formation phenotype whereby larvae enter dauer diapause even in conditions that normally promote non-dauer development (Riddle, 1977). Therefore, active DAF-2/IR signaling promotes non-diapause development.

In insects, the role of IIS in regulating diapause has been studied extensively with respect to the reproductive diapause that occurs in adult mosquitos (C. pipiens) and fruit flies (D. melanogaster). In C. pipiens, levels of two ILPs, ILP-1, and ILP-5, are reduced during diapause compared with non-diapausing animals. RNAi of ILP-1 but not ILP-5 is sufficient to induce adult diapause as measured by ovarian maturation and follicle length (Sim and Denlinger, 2009). Similarly, RNAi of IR induces adult diapause by the same measures (Sim and Denlinger, 2008). Therefore, IIS signaling opposes diapause in C. pipiens.

Similarly, in D. melanogaster, loss of dilps 1-5 strongly induces reproductive diapause. Single dilp2 and dilp5, but not dilp1, mutations are sufficient to induce reproductive diapause at low penetrance, whereas overexpression of dilp2 and dilp5 from any tissue strongly suppresses reproductive diapause (Kubrak et al., 2014; Liu et al., 2016; Schiesari et al., 2016). Furthermore, hypomorphic mutations in InR or the insulin receptor substrate protein Chico cause reproductive diapause entry (Schiesari et al., 2016). Surprisingly, although dILPs functionally oppose reproductive diapause, mRNA levels of the five ILPs tested (dilp1, dilp2, dilp3, dilp5, dilp6) are all upregulated during reproductive diapause and then return to non-diapause levels after recovery from reproductive diapause (Kubrak et al., 2014; Liu et al., 2016). This upregulation may be a compensatory response to reduced IIS signaling under diapause-inducing conditions (Schiesari et al., 2016).

Another insect that undergoes reproductive diapause is the bumblebee Bombus terrestris. In this species, genes within the IIS pathway were upregulated in post-diapause animals compared to diapause. Furthermore, injection of ILPs stimulated metabolic changes consistent with non-diapause development and increased the numbers of offspring produced (Chen et al., 2021). Finally, there are many insect species where reduced expression of ILPs or IR has been reported to correlate with diapause at different stages of development, consistent with the notion that downregulation of IIS is a broadly conserved phenomenon. Some examples include egg diapause in the aphid Acyrthosiphon pisum and in Locusta migratoria (Barberà et al., 2019; Hao et al., 2019), embryonic diapause in B. mori (Gong et al., 2020), and pre-pupal diapause in leafcutting bees, Megachile rotundata (Cambron et al., 2021).

In the annual fish A. limnaeus, levels of IGF-I are reduced before and during DII compared to escape embryos at the same stages. After recovery, IGF-I levels climb, eventually surpassing those of escape embryos (Woll and Podrabsky, 2017). IGF-II levels are similar in escape and diapause embryos before and during DII, but are higher post-DII than in escape embryos. Higher levels of IGF after diapause may be related to the catch-up growth that is seen in post-diapause embryos. In addition to the correlation of IGF levels, treatment of embryos with an IGF1R inhibitor induces diapause in a dose-dependent manner. Pre-DII and DII embryos induced by this treatment show similarities to naturally induced DII, including slowed heart rate and slowed somitogenesis. However, there were a few differences, including the presence of melanocytes and hemoglobin in circulating blood, indicating that IGF signaling is not the sole regulator of diapause (Woll and Podrabsky, 2017).

The role of downstream components of IIS in the regulation of diapause has been studied primarily in C. elegans, where the dauer diapause-regulating pathways have been worked out in detail (Fielenbach and Antebi, 2008; Murphy and Hu, 2013). Each component functions as expected given its role in insulin/IGF signaling. Specifically, when active, the orthologs of PI3K and AKT oppose dauer formation, whereas the ortholog of PTEN promotes dauer formation. The single PI3K in C. elegans is encoded by age-1, also called daf-23 (Morris et al., 1996). age-1 was first identified based on the long-lived phenotype conferred by mutants (Klass, 1983). Separately, this gene was identified in a screen for mutants that enter dauer in conditions that normally promote non-dauer development (Gottlieb and Ruvkun, 1994). Similarly, the single PTEN ortholog, daf-18 was identified on the basis of the inability of a daf-18(-) mutant to enter dauer under dauer-inducing conditions (Riddle et al., 1981). There are two AKT homologs in C. elegans, akt-1 and akt-2 that function partially redundantly to oppose dauer diapause. Loss of akt-1 produces a weak dauer-constitutive phenotype, whereas loss of both akt-1 and akt-2 produces a very strong constitutive dauer diapause phenotype (Ailion and Thomas, 2003; Oh et al., 2005).

In insects, the PI3K ortholog, Dp110 was identified based on the variation in reproductive diapause rates between two natural strains of D. melanogaster, one that enters reproductive diapause more frequently than the other. Dp110 appears to be one locus responsible for this difference (Williams et al., 2006). Although the molecular changes in Dp110 that lead to this reproductive diapause phenotype were not clear, reducing the dosage of Dp110 by half increases the percentage of flies adopting reproductive diapause, whereas misexpressing Dp110 in neurons decreases the percentage of flies adopting reproductive diapause (Williams et al., 2006).

One common output of IIS is regulation of the FOXO transcription factors. IIS negatively regulates FOXO, such that when IIS is active, FOXO proteins are phosphorylated by AKT, which triggers nuclear exit and therefore inactivity (Manning and Cantley, 2007). Therefore, low levels of IIS are required for FOXO nuclear localization and activity. Across species, FOXO transcription factors promote diapause by directly and indirectly controlling the expression of numerous genes that are involved in diapause. Furthermore, if situations are contrived whereby animals within diapause have reduced FOXO activity, those animals lose key characteristics of diapause, suggesting that FOXO promotes both the decision to enter diapause and the physiological and structural changes that accompany diapause, as explained below.

The specific role of FOXO proteins in diapause has not been studied in fishes, but there is a great deal of information from invertebrate species, which typically have a single FOXO ortholog. In C. elegans, the FOXO protein is encoded by daf-16. Like most of the genes in the IIS pathway, daf-16 was also discovered on the basis of its dauer formation phenotype. Whereas loss-of-function mutations in daf-2 or other IIS components cause dauer formation even in environments favorable for growth, loss-of-function mutations in daf-16 cause the opposite phenotype: inability to enter dauer diapause even in dauer-inducing conditions. Furthermore, when double mutants are created, the daf-16 dauer-defective phenotype is observed (Murphy and Hu, 2013). The complete suppression of the daf-2(-) dauer-constitutive phenotype by loss of daf-16 indicates that the major output of IIS with respect to regulation of diapause is daf-16/FOXO.

While daf-16(-) larvae do not readily enter dauer diapause, it is possible to obtain daf-16(-) dauer larvae by making use of dauer-constitutive mutations outside of the IIS pathway, or by the use of very potent dauer-inducing cues. However, these daf-16(0) dauer larvae display defects in some aspects of the remodeling that occurs as part of dauer morphogenesis, demonstrating that daf-16/FOXO is not only responsible for the decision to enter dauer but for some of the aspects of diapause itself. For this reason, daf-16(-) dauer larvae have been referred to as “partial dauers” (Vowels and Thomas, 1992; Larsen et al., 1995; Ogg et al., 1997). For example, daf-16 mutants lack the particular collagens in their cuticle that are normally enriched in dauer larvae, and perhaps for this reason they are more sensitive to environmental assaults (Gottlieb and Ruvkun, 1994; Nika et al., 2016; Wirick et al., 2021).

The specific gene expression changes that occur in the absence of daf-16/FOXO have been studied primarily in non-diapause contexts: adult worms with activated DAF-16 due to a mutation in IR are compared to IR mutants that lack daf-16. As in other species, daf-16 promotes expression of genes involved in stress resistance and metabolism (Tepper et al., 2013). A recent study examined the changes in gene expression of dauer larvae with and without daf-16. In this context, genes related to collagens and signaling were differentially expressed (Wirick et al., 2021).

In mosquitos, levels of FOXO are increased in the fat body of diapausing females, compared to non-diapause animals (Sim and Denlinger, 2013b). Fat storage is an important component of diapause and appears to be dependent on FOXO because knockdown of FOXO with dsRNA in diapausing females causes a reduction in lipid stores (Sim and Denlinger, 2008). These females also showed reduced survival over time in diapause. Reduced survival may be due to insufficient fat storage, but because addition of an oxidoreductase (Mn(III)TBAP) partially rescues survival, oxidative stress is also implicated (Sim and Denlinger, 2008). FOXO promotes resistance to oxidative stress across species. Genes bound by FOXO during diapause were identified by ChIP-seq. These target genes are involved in stress-resistance, metabolism, longevity, cell-cycle, development, and circadian rhythm (Sim et al., 2015). RNAi knockdown of three of these targets reduced glycogen and lipid levels during diapause, consistent with the idea that FOXO regulation of target genes is important for diapause characteristics (Olademehin et al., 2020). In Drosophila, a FOXO-responsive luciferase reporter is upregulated during reproductive diapause. Abdomens isolated from cold-reared flies in reproductive diapause showed six-fold more activity of this reporter than non-diapausing flies reared under standard lab conditions (Schiesari et al., 2016).

In Locusta migratoria egg diapause, if FOXO is knocked down in the mother by RNAi under diapause-inducing conditions (short photoperiod), the percentage of eggs that enter diapause is reduced, indicating that FOXO promotes entry into diapause (Hao et al., 2019). Consistent with the activity of FOXO being important for diapause, the known FOXO target, MnSod/sod2, was found to be upregulated under diapause-inducing conditions (Hao et al., 2019).

In the cotton bollworm Helicoverpa armigera pupal diapause, levels of total FOXO are high but phospho-FOXO levels are low, indicating that FOXO is in its active state. Interestingly, this regulation occurs by a different mechanism than the typical IIS pathway (Zhang X.-S. et al., 2017). Although ILP levels are low, these ILP levels do not appear to regulate FOXO activity via Akt, because phospho-AKT levels are high in diapause. Instead, high levels of ROS lead to phosphorylation of Akt as well as high levels of protein arginine methyltransferase 1 (PRMT1). PRMT1 methylates FOXO, thus abrogating phosphorylation by Akt. In this way, FOXO remains active to promote diapause at the same time as active Akt can phosphorylate other targets (Zhang X.-S. et al., 2017).

Class II NHRs play important roles in the regulation of diapause across phyla. The NHRs implicated in diapause are phylogenetically related and include the vitamin D receptor in annual fishes, the ecdysone receptor (EcR) in insects, and DAF-12 in C. elegans. In most of these cases, low levels of ligand are associated with diapause whereas ligand-bound NHR leads to non-diapause development (Figure 2 and Table 3).

In C. elegans, daf-12 was discovered on the basis of the dauer formation phenotypes observed in daf-12 mutants. Notably, both dauer-defective and dauer-constitutive alleles of daf-12 were isolated (Antebi et al., 1998). Dauer defective alleles are mutants that failed to enter diapause even in diapause-inducing conditions. Dauer-constitutive alleles enter diapause even in conditions favorable for growth. This complexity stems in part from the two activities of NHRs as activators and repressors. Null alleles remove both activator and repressor functions, whereas other alleles can affect one function more than the other. Null alleles do not enter diapause under any known condition, indicating that daf-12 activity is required for diapause. By contrast, all of the dauer-constitutive alleles disrupt the ligand binding domain, suggesting that it is the unliganded version of DAF-12 that is required for dauer formation (Antebi et al., 2000). Consistent with this idea, the DIN1S co-repressor was found to bind unliganded DAF-12 and to be required for dauer formation (Ludewig et al., 2004).

As mentioned above, most Class II NHRs function as heterodimers with RXR proteins. However, C. elegans lacks a clear RXR ortholog despite the expansion of NHR-encoding genes in C. elegans: 284, compared to 48 NHRs in humans (Antebi, 2015). NHRs without clear orthology to RXR have been suggested to play a similar role in some cases, however, whether DAF-12 functions as a homo- or heterodimer is still unknown (Antebi, 2015). DAF-12 does possess a conserved dimerization domain, and the DNA response element to which DAF-12 binds is consistent with either homo- or heterodimerization (Antebi et al., 2000; Shostak et al., 2004; Hochbaum et al., 2011).

In insects, EcR signaling has been studied in depth in D. melanogaster and other insects, primarily for its role in molting. Indeed, ecdysteroids are often referred to as molting hormones. In insects, the EcR functions as a heterodimer with the RXR ortholog, Ultraspiracle (USP) (Oro et al., 1990; Koelle et al., 1991; Yao et al., 1992; Thomas et al., 1993). There are three isoforms of EcR: A, B1, and B2, that share DNA-binding and ligand-binding domains but differ at the N-terminus. These isoforms differ in expression pattern and biochemical activity; however, they all function to activate transcription of target genes when bound to ligand, and to repress transcription when ligand-free (Talbot et al., 1993). In some contexts, the unliganded form of EcR is specifically required (Mansilla et al., 2016).

With respect to diapause, EcR signaling plays a regulatory role in many insects, a few of which will be highlighted here. These include reproductive diapause in D. melanogaster and embryonic diapause in B. mori. For a more comprehensive view of EcR signaling regulating diapause across insect species (see Denlinger et al., 2012). Unlike in C. elegans, functional consequences of reduced EcR signaling have not been described in the context of insect diapause, and instead more work has been done manipulating hormone levels, as described in section “Nuclear Hormone Receptor Signaling Ligands and Biosynthetic Enzymes and the Regulation of Diapause.”

In the annual fish A. limnaeus, a role for the vitamin D receptor (VDR) in the regulation of diapause II (DII) was recently described (Romney et al., 2018). There are two VDR-encoding genes in A. limnaeus: VDR-A and VDR-B. These receptors also bind to target DNA as homodimers or heterodimers, where heterodimerization with RXR is the predominant mode (Carlberg et al., 1993; Carlberg and Campbell, 2013). In A. limnaeus embryos the expression of VDR-A and RXR does not change between pre-diapause and pre-escape embryos. However, the expression of VDR-B is increased in pre-diapause embryos relative to those on the escape trajectory (Romney et al., 2018).

As described in the prior section, in the examples given, active NHR signaling promotes non-diapause development. The control of this signaling appears to occur primarily at the level of production of the active ligand. In insects, the bulk of the studies on the role of EcR signaling in diapause have focused on the ligand and the enzymes necessary to produce the active form. The most active EcR ligand is 20-hydroxyecdysone (20E). In Drosophila, 20E is synthesized from cholesterol by a series of enzymatic reactions. These reactions begin with cholesterol or plant-derived sterols taken up from the environment, since insects cannot synthesize cholesterol themselves (Niwa and Niwa, 2014). The first step in the pathways is catalyzed by Rieske oxygenase Neverland (Nvd) which converts cholesterol to 7-dehydrocholesterol. Next is a “black box,” between 7-dehydrocholesterol and 5β-ketodiol, where the precise reactions are not known. However, the short-chain dehydrogenase Non-molting glossy/Shroud (Nm-g/Sro) and the cytochrome P450 enzymes Spook and Spookier (Spo, Spok) are thought to catalyze reactions within the black box. After that, the rest of the biosynthetic pathway is known, and each step is catalyzed by a series of cytochrome P450 enzymes, including Phantom (Phm), Disembodied (Dib), Shadow (Sad), and Shade (Shd), in that order (Niwa and Niwa, 2014).

To regulate molting, ecdysone is synthesized in the prothoracic glands and then ecdysone is released into the hemolymph. Ecdysone is then converted into 20E in target tissues around the body (Petryk et al., 2003). However, ecdysone can also be synthesized in the ovaries. In D. melanogaster, the strongest evidence for a role for ecdysone signaling in reproductive diapause comes from experiments where injection of diapausing animals with 20E caused recovery, as assessed by the resumption of vitellogenesis (Richard et al., 2001a). Additionally, levels of 20E are low in reproductive diapause animals compared with non-diapause animals. When flies in reproductive diapause are stimulated to recover, the 20E levels rise rapidly (Richard et al., 1998).

Similarly, in B. mori, high levels of 20E are found in non-diapause embryos, but not diapause embryos. Instead, diapausing embryos contain high levels of 20E precursors or inactive derivatives. Furthermore, treatment of diapausing embryos with 20E can induce the resumption of development and recovery from diapause (Makka et al., 2002; Sonobe and Yamada, 2004). Therefore, regulation of the production of 20E is necessary to control diapause entry. Because the prothoracic gland has not yet developed in these embryos, the regulation of 20E synthesis differs from the canonical pathway. 20E in embryos is derived from two sources. First, inactive, phosphorylated forms of 20E and its precursors are bound to yolk proteins and maternally packaged into the embryo. Second, 20E is synthesized de novo. In both cases, the regulation of enzymes determines whether active 20E is produced. The ecdysteroid-phosphate phosphatase (EPPase) is required to activate the phosphorylated molecules, whereas the enzyme 20-hydroxylase (E20OHase, also called Shade in some species) is required for de novo synthesis. The expression of these enzymes is upregulated over time in non-diapause embryos whereas expression remains low during diapause (Sonobe and Yamada, 2004; Maeda et al., 2008; Gu et al., 2021).

Other species with larval and pupal diapause show similar effects whereby increased levels of 20E have been shown to correlate with and/or actively promote non-diapause development. Since EcR signaling is necessary for progression through larval molts and for metamorphosis, it makes sense that removing 20E would allow for developmental arrest. Somewhat surprisingly, however, in some insect species, the role for EcR signaling is reversed, and high levels of 20E are required to maintain diapause. For example, this mode of diapause regulation was shown for the obligate diapause at the end of embryogenesis in the gypsy moth, Lymantria dispar (Denlinger et al., 2012; Schiesari and O’Connor, 2013).

In C. elegans the first two ligands that activate DAF-12 signaling to be discovered were two 3-Keto-4-Cholestenoic Acids termed Δ⁴-dafachronic acid and Δ⁷-dafachronic acid (Δ⁴-DA and Δ⁷-DA) (Motola et al., 2006). The term “dafachronic” comes from the two major roles of DAF-12: the regulation of diapause and the regulation of developmental timing (see section “Developmental Changes Accompanying Diapause in C. elegans”). More recent work identified additional DAs, the most prominent being Δ^1,7-DA. This study failed to find Δ⁴-DA, suggesting that ligand may be less relevant in endogenous contexts (Mahanti et al., 2014).

Fewer steps in the synthesis of DAs are known in detail in comparison to the synthesis of 20E. However, some key enzymes show similar requirements across species. First, the Rieske oxygenase DAF-36 is required for the synthesis of Δ⁷-DA. Similar to the role of Nvd in ecdysone synthesis, this enzyme catalyzes the first reaction within this biosynthetic pathway: conversion of cholesterol to 7-dehydrocholesterol (Rottiers et al., 2006; Wollam et al., 2011; Yoshiyama-Yanagawa et al., 2011). Indeed, these enzymes are functionally conserved as addition of daf-36 to D. melanogaster with a mutation in nvd rescues the lethal phenotype (Yoshiyama-Yanagawa et al., 2011). The short chain dehydrogenase DHS-16 works further downstream in this pathway. The enzymes involved in the production of other dafachronic acids are unknown. The hydroxysteroid dehydrogenase HSD-1 was proposed to be involved in the production of Δ⁴-DA, but subsequent work raises doubts about this notion. At the end of the pathway, the cytochrome P450 enzyme DAF-9 is required to catalyze the final step in biosynthesis of both Δ⁴-DA and Δ⁷-DA (Antebi, 2015). The expression of daf-9 is highly regulated during dauer formation, where mildly stressful conditions increase daf-9 expression, but more severe, dauer-inducing conditions drastically reduce daf-9 expression, leading to a decrease in DA production and entry into dauer diapause (Gerisch et al., 2001; Gerisch and Antebi, 2004; Schaedel et al., 2012).

In vertebrates, biosynthesis of the active VDR ligand 1,25(OH)₂D₃ is similar in many respects to the biosynthesis of the invertebrate NHR ligands. Biosynthesis in vertebrates also begins with the conversion of cholesterol to 7-hydroxycholesterol, in this case by the action of the dehydrocholesterol reductase enzyme, DHCR7. Unlike in invertebrates, the next step of the biosynthetic pathway requires UVB light, which enables the conversion of 7-hydroxycholesterol to pre-vitamin D3. Once pre-vitamin D3 is produced, the remainder of the biosynthetic steps are catalyzed by a series of cytochrome P450 enzymes, similar to the invertebrate pathway (Tuckey et al., 2018; Saponaro et al., 2020).

Within the annual fish A. limnaeus, expression of the active VDR ligand 1,25(OH)₂D₃ is reduced in pre-diapause embryos compared to pre-escape embryos. Consistent with this reduction, the expression of several biosynthetic enzymes necessary to produce 1,25(OH)₂D₃ are also reduced (Romney et al., 2018). This difference in expression is functionally relevant because incubating embryos reared under diapause-inducing conditions with 1,25(OH)₂D₃ at picomolar concentrations strongly reduced the number of embryos that underwent diapause (Romney et al., 2018). Remarkably, performing a similar experiment with C. elegans Δ⁴-dafachronic acid also increased the number of embryos following the non-diapause, escape developmental trajectory. This effect was specific because incubating embryos with ligands for the pregnane X, farnesoid X, and liver X receptors, which are closely related to VDR, was ineffective at preventing diapause (Romney et al., 2018). Furthermore, treating A. limnaeus embryos grown in non-diapause conditions with dafadine A causes embryos to enter DII. Dafadine A is a compound that inhibits the activity of the cytochrome P450 enzyme DAF-9 and its mammalian counterpart CYP27A1 (vitamin D3 25-hydroxylase) (Luciani et al., 2011). The diapause-inducing effect of dafadine A in A. limnaeus can be countered by addition of 1,25(OH)₂D₃, consistent with the interpretation that dafadine A promotes diapause by interfering with the production of 1,25(OH)₂D₃ (Romney et al., 2018). Taken together, these results demonstrate the conserved nature of NHR signaling and its effect on diapause in A. limnaeus and C. elegans.

Caenorhabditis elegans geneticists have worked out the major regulators of dauer formation in detail. In addition to the IIS and NHR/DAF-12 pathways described in the previous sections, TGFβ signaling is also involved. In brief, TGFβ and IIS pathways operate largely in parallel and both pathways converge to regulate DAF-12 (Figure 3A; Baugh and Hu, 2020; Fielenbach and Antebi, 2008).

Favorable environmental cues lead to the production of both DAF-7/TGFβ and multiple insulin-like peptides in sensory neurons (Ren et al., 1996; Schackwitz et al., 1996; Cornils et al., 2011; Murphy and Hu, 2013; Ritter et al., 2013; Fernandes-de-Abreu et al., 2014; Zheng et al., 2018). These signals are released and then received in multiple tissues where DAF-7/TGFβ blocks activity of the downstream DAF-3/SMAD-DAF-5/Ski complex, and insulin signaling blocks the activity of the downstream DAF-16/FOXO transcription factor (Kimura et al., 1997; Lin et al., 1997; Ogg et al., 1997; Patterson et al., 1997; Pierce et al., 2001; da Graca et al., 2004). Both IIS and TGFβ pathways regulate the expression of daf-9/CYP, which encodes the last enzyme required to produce dafachronic acid, the ligand for the DAF-12 nuclear hormone receptor (Gerisch et al., 2001; Gerisch and Antebi, 2004; Mak and Ruvkun, 2004; Motola et al., 2006). As described above, ligand-bound DAF-12 promotes continuous (non-dauer diapause) development whereas ligand-free DAF-12 is required for dauer formation (Antebi et al., 1998, 2000; Ludewig et al., 2004). IIS appears to also act in parallel to daf-12 because when both daf-2/IR and daf-12/NHR signaling are compromised, larvae arrest development but do not enter dauer. Overexpression of daf-9 in a daf-2 mutant background produces a similar arrested phenotype that is different from dauer diapause (Larsen et al., 1995; Gems et al., 1998; Gerisch et al., 2001; Gerisch and Antebi, 2004).

In insects, another major regulator of diapause is juvenile hormone (JH). JH is a sesquiterpenoid that is produced in a group of endocrine organs called the corpora allata (CA) in response to the hormone allatotropin (Flatt et al., 2005; Bendena et al., 2020). JH is then secreted and travels through the hemolymph to act on the ovaries, stimulating ovarian development and vitellogenesis. Since reproductive diapause occurs prior to vitellogenesis, it is not surprising that JH is a central regulator of reproductive diapause across insect species (Denlinger et al., 2012).

Juvenile hormone and Insulin/IGF signaling exhibit an interesting and complex relationship in C. pipiens (Figure 3B). During diapause in C. pipiens, JH levels are low, and addition of synthetic JH can induce exit from diapause (Spielman, 1974). Addition of a JH analog is effective at stimulating diapause exit even in IIS mutants, indicating that IIS regulates diapause formation upstream of JH (Sim and Denlinger, 2008, 2009). However, both IIS and JH signaling appear to converge on the same target: negative regulation of FOXO. Negative regulation of FOXO by IIS is expected, but the role of JH is novel. Addition of JH to diapausing animals reduces FOXO levels in the fat body (Sim and Denlinger, 2013b). FOXO appears to be the most direct regulator of diapause genes, including genes that regulate stress-resistance, metabolism, longevity, and other key characteristics of diapause (Sim et al., 2015).

Finally, there is a potential connection to EcR signaling in C. pipiens as well. This connection was proposed by Sim and Denlinger and is based on work done in another mosquito species, Aedes aegypti (A. aegypti) (Sim and Denlinger, 2013b). In this species, a blood meal enables reproduction by stimulating vitellogenesis. Vitellogenesis requires EcR signaling and the EcR/USP heterodimer directly binds to the promoter of genes encoding yolk proteins in the fat body. JH expression promotes expression of the NHR bFTZ-F1, which in turn recruits the FISC p160/SRC coactivator. These proteins form a complex with EcR and USP at the promoters of target genes and are required for full EcR activation (Zhu et al., 2006). If the same process occurs in C. pipiens in the context of diapause, that would forge a connection between the signaling pathways described here.

Similar to the situation in C. pipiens, JH levels are reduced in D. melanogaster in reproductive diapause. Furthermore, addition of JH to diapausing animals induces recovery from reproductive diapause (Saunders et al., 1990). However, EcR was later proposed to be more directly involved in the regulation of reproductive diapause than JH because levels of ecdysteroids rose more rapidly after stimulation to recover from reproductive diapause, and addition of ecdysteroids to diapausing animals was more effective at terminating reproductive diapause than addition of JH (Richard et al., 1998, 2001b). Unlike in C. pipiens, the IIS pathway has been proposed to act downstream of JH and EcR signaling because loss of the insulin receptor substrate protein Chico induces reproductive diapause in an ovary-autonomous manner, suggesting that external signals such as JH and 20E do not affect reproductive diapause in this mutant context. chico mutant ovaries did not develop upon transplantation to a wild-type host, whereas wild-type ovaries did develop when transplanted to a chico mutant host. Furthermore, chico mutants expressed normal levels of JH and ecdysteroids (Richard et al., 2005). Consistent with the notion that IIS is downstream of JH, the addition of a JH inhibitor did affect the expression of dilp1, however, the functional consequences of this observation are unclear (Liu et al., 2016).

Diapause in B. mori is controlled by a neuropeptide called diapause hormone (DH). This mode of regulation is well studied, though uncommon in other insect species (Denlinger et al., 2012). B. mori diapause is transgenerationally programmed (see section “Diapause in Insects: Drosophila melanogaster, Culex pipiens, and Bombyx mori”). Mothers reared in diapause-inducing conditions during their own embryonic development will express DH in the subesophageal ganglion during the pupal stage. DH is released and then received in the ovary via the DH receptor, a G-protein coupled receptor. This stimulus prompts the developing ovaries to induce diapause (Xu et al., 1995b; Yamashita et al., 2001; Homma et al., 2006; Denlinger et al., 2012; Sato et al., 2014). A connection between DH and EcR signaling has been proposed, but experiments testing this connection have yielded few conclusions, such that the relationship between these pathways is still unclear (Xu et al., 1995a; Shiomi et al., 2015). However, MEK/ERK signaling is needed to terminate diapause, and this process appears to occur upstream of EcR signaling (Fujiwara et al., 2006a, b; Schiesari and O’Connor, 2013). Additionally, the IIS pathway has been implicated in diapause in B. mori, but it is unclear how this function fits with DH or EcR pathways (Zheng et al., 2016; Gu et al., 2019).

When L1 larvae hatch from embryos, their germline consists of two cells. In response to Notch signaling, these cells proliferate and expand the germline stem cell population over the first several larval stages. Meiosis does not begin until after dauer would have occurred, during the mid-L3 stage (Hubbard and Schedl, 2019).

During L2d, germ cell proliferation slows as larvae prepare to enter dauer, and proliferation ceases during dauer diapause (Narbonne and Roy, 2006). Reducing the rate of germ cell proliferation in L2d requires several genes within the AMPK and IIS pathways, including genes encoding both subunits of the energy-sensing AMP-activated Protein Kinase (AMPK), called aak-1 and aak-2 in worms, the gene encoding the AMPK-activating kinase PAR-4/LKB1, and the gene encoding the PTEN phosphatase DAF-18 (Figure 4). Although these genes all affect related pathways, they are each required independently. Furthermore, these pathways function downstream of or in parallel to the Notch pathway that promotes germ cell division in favorable environments (Narbonne and Roy, 2006, 2009; Tenen and Greenwald, 2019).

In addition to its importance in slowing germ cell proliferation in L2d larvae, AMPK is required during and after dauer to regulate small RNA pathways that in turn maintain proper chromatin marks and therefore affect gene expression. Dauer and post-dauer larvae that lack aak-1 display aberrant gene expression and post-dauer adults are sterile. Remarkably, AMPK activity is required in somatic cells for these functions in germ cells (Kadekar and Roy, 2019).

daf-18/PTEN is unique among the genes listed above in that it is required in the somatic gonad for quiescence in both the germ line and the somatic gonad (Tenen and Greenwald, 2019). Furthermore, daf-18 is required to maintain developmental arrest during dauer in the somatic gonad progenitor cells, because cell fate markers that indicate developmental decisions that normally occur in the L3 or L4 stage are observed in dauer larvae that lack daf-18 (Tenen and Greenwald, 2019).

In parallel to AMPK and IIS signaling, the DAF-12 NHR and its co-repressor DIN-1S also inhibit proliferation of germline and somatic gonad cells in the L2d stage (Figure 4). In contrast to aak-1 and daf-18 that function non-cell-autonomously in somatic cells, din-1s is required autonomously in germ cells to prevent germ cell proliferation and in somatic gonad cells to prevent their proliferation (Colella et al., 2016).

Developmental pathways appear to be not only inactivated but actively re-set during dauer diapause. Some of the best evidence for active re-setting of developmental pathways comes from work examining the vulval precursor cells (VPCs). VPCs are born during the L1 stage and remain multipotent and unspecified until the L3 stage. During the L3 stage these cells respond to a combination of EGFR/Ras and Notch signaling pathways to specify VPCs to one of two vulval cell fates (1° and 2°). The absence of signaling leads to a default non-vulval cell fate (3°) (Sternberg, 2005). In wild-type larvae, 1° cell fate markers are expressed in the presumptive 1° VPC during the molt into dauer diapause, but this specification is erased during dauer. After recovery from dauer, larvae enter the post-dauer L3 (PDL3) stage that is the developmental equivalent of the L3 stage that occurs during the non-diapause developmental trajectory. VPC specification is again initiated in PDL3 (Karp and Greenwald, 2013).

The re-setting of VPC specification during dauer is more apparent in mutants that cause VPC specification prior to dauer. There are two categories of these mutants. First, loss of the heterochronic gene lin-28 causes precocious VPC specification during the L2 stage (Ambros and Horvitz, 1984). Larvae lacking lin-28 that develop through the dauer trajectory show VPC specification during the pre-dauer L2d stage. This specification is erased during dauer, and then re-initiates in VPCs or their descendants in post-dauer larvae (Euling and Ambros, 1996; Karp and Greenwald, 2013). Second, mutants in which EGFR/Ras or LIN-12/Notch signaling have been activated also show VPC specification during the pre-dauer L2d stage. Again, this specification is lost during dauer (Karp and Greenwald, 2013). Therefore, there is an active mechanism that erases any VPC specification that has occurred prior to dauer formation and re-establishes multipotent VPC fate.

Both DAF-12/NHR signaling and DAF-16/FOXO have been implicated in VPC development during dauer (Figure 4). Mutations in daf-12/NHR, din-1s, which encodes a corepressor that binds unliganded DAF-12, or daf-9, which encodes an enzyme required for synthesis of the DAF-12 ligand, all result in VPC division and vulval development during dauer (Karp and Greenwald, 2013; Colella et al., 2016). However, this phenotype is suppressed when daf-12 or daf-9 mutants are exposed to dauer-inducing conditions (dauer pheromone), suggesting that these genes are not the most proximal regulators of VPC quiescence. By contrast, loss of daf-16 results in VPC division as well as adoption of 1° and 2° fate markers, even in the presence of dauer pheromone (Karp and Greenwald, 2013). Therefore, during dauer, daf-16/FOXO blocks EGFR/Ras and LIN-12/Notch signaling to prevent precocious specification.

One of the developmental pathways most profoundly affected by diapause is the “heterochronic” pathway that controls stage-specific cell fate, particularly in the stem-cell-like seam cells within the lateral hypodermis. Seam cells divide in a particular pattern and sequence at each larval stage before terminally differentiating at adulthood. The gene network that controls this pattern of stage-specific cell divisions and differentiation is termed the “heterochronic” pathway (Ambros and Horvitz, 1984). These heterochronic genes function as a molecular timer whereby transcription factors and RNA-binding proteins that specify early cell fate are expressed at each larval stage. These early-promoting factors are then downregulated by microRNAs in order to allow progression to the next cell fate (Rougvie and Moss, 2013). A molecular timer would seem to be less useful in an interrupted developmental trajectory, and perhaps for that reason the heterochronic pathway is altered in larvae whose development is interrupted by dauer diapause (Liu and Ambros, 1991). For example, microRNA activity is modulated in L2d and dauer larvae, so that a different set of microRNAs becomes more important in the dauer diapause context (Karp and Ambros, 2012; Ilbay and Ambros, 2019).

DAF-12/NHR plays an important role in coordinating the decision to enter dauer with seam cell fate (Figure 4). In diapause-inducing conditions, there is little of the DAF-12 ligand produced and DAF-12 binds to its co-repressor, DIN-1S to promote dauer formation. DAF-12 also directly regulates the transcription of heterochronic microRNAs within the let-7 family (Bethke et al., 2009). These microRNAs are upregulated in order to promote L3 cell fate. In the lengthened L2d stage, the repressor form of DAF-12 keeps levels of these microRNAs low and L3 fate is not adopted prematurely (Bethke et al., 2009; Hammell et al., 2009). The let-7 family microRNAs also directly target daf-12, forming a feedback loop that helps to coordinate the dauer formation decision with the regulation of seam cell development (Hammell et al., 2009).

As described in section “Effect of Diapause on Vulval Precursor Development in C. elegans,” VPC fate appears to be re-set during dauer. A similar type of re-setting appears to occur in lateral hypodermal cells, including the seam cells. This type of re-setting was initially inferred based on the observation that the seam cell lineage defects that occur in certain heterochronic mutants were corrected in larvae that developed through the diapause trajectory (Liu and Ambros, 1991). More recently, a role for daf-16/FOXO was discovered in these cells. daf-16/FOXO is required in dauer larvae to prevent the precocious expression of collagens that are normally enriched in adults. daf-16/FOXO appears to act partially via known heterochronic genes and partially via a novel mechanism. Specifically, daf-16/FOXO opposes expression of the lin-41/TRIM71 RNA-binding protein during dauer to block adult cell fate. However, lin-41/TRIM71 does not act via its canonical target, the LIN-29 transcription factor, indicating a novel mechanism at play during dauer (Wirick et al., 2021). Thus, daf-16/FOXO coordinates the diapause developmental trajectory with multiple developmental events (Figure 4).