Heat shock factor-1 intertwines insulin/IGF-1, TGF-β and cGMP signaling to control development and aging

Abstract

Background: Temperature affects virtually all cellular processes. A quick increase in temperature challenges the cells to undergo a heat shock response to maintain cellular homeostasis. Heat shock factor-1 (HSF-1) functions as a major player in this response as it activates the transcription of genes coding for molecular chaperones (also called heat shock proteins) that maintain structural integrity of proteins. However, the mechanisms by which HSF-1 adjusts fundamental cellular processes such as growth, proliferation, differentiation and aging to the ambient temperature remain largely unknown.

Results: We demonstrate here that in Caenorhabditis elegans HSF-1 represses the expression of daf-7 encoding a TGF-β (transforming growth factor-beta) ligand, to induce young larvae to enter the dauer stage, a developmentally arrested, non-feeding, highly stress-resistant, long-lived larval form triggered by crowding and starvation. Under favorable conditions, HSF-1 is inhibited by crowding pheromone-sensitive guanylate cyclase/cGMP (cyclic guanosine monophosphate) and systemic nutrient-sensing insulin/IGF-1 (insulin-like growth factor-1) signaling; loss of HSF-1 activity allows DAF-7 to promote reproductive growth. Thus, HSF-1 interconnects the insulin/IGF-1, TGF-β and cGMP neuroendocrine systems to control development and longevity in response to diverse environmental stimuli. Furthermore, HSF-1 upregulates another TGF-β pathway-interacting gene, daf-9/cytochrome P450, thereby fine-tuning the decision between normal growth and dauer formation.

Conclusion: Together, these results provide mechanistic insight into how temperature, nutrient availability and population density coordinately influence development, lifespan, behavior and stress response through HSF-1.

Figure 1

HSF-1 represses daf-7 expression, and is negatively regulated by DAF-11 and DAF-21. A, A signaling network of the C. elegans insulin/IGF-1, cGMP and TGF-β neuroendocrine systems. The model relies on previously published data. TGF-β signaling promotes reproductive growth. Under adverse environmental conditions, daf-7 becomes downregulated, resulting in the activation of the nuclear hormone receptor DAF-12 that eventually induces dauer development. Crowding pheromone-dependent cGMP signaling mediated by the receptor guanylate cyclase DAF-11 acts upstream of DAF-7 to inhibit dauer larva formation. Insulin/IGF-1 signaling hampers dauer development through the TGF-β pathway. In addition, activity of the IGF-1 receptor DAF-2, which is inhibited or activated by different insulin-like peptides, such as DAF-28, accelerates aging by inhibiting nuclear translocation of the forkhead transcription factor DAF-16. HSF-1 is negatively regulated by insulin/IGF-1 signaling, and required for longevity response triggered by DAF-16. The inhibitory effect of DAF-2/IGF-1 signaling on HSF-1 occurs through the activation of the DDL-1/2 proteins, two negative regulators of HSF-1. DAF-11 upregulates daf-28 (the dotted arrow), while DAF-3 influences DAF-2 activity through its regulation of the ins-7 agonist and ins-18 antagonist (the dotted green bar). Arrows indicate activations, bars represent inhibitory interactions. C. elegans proteins are in blue; their mammalian counterparts are in black. Dotted lines show known signaling links between these neuroendocrine systems. FoXO: Forkhead box O transcription factor; PDK1: 3-phosphoinositide-dependent kinase 1; Akt: AKT8 virus protooncogene; PKB: protein kinase B; SGK: serum- and glucocorticoid-inducible kinase; PTEN: phosphatase and tensin homolog; PI3K: phosphoinositide 3-kinase; IGF-1: insulin-like growth factor receptor-1; HSF-1: heat shock factor-1; NHR: nuclear hormone receptor; SMAD: Caenorhabditis elegans protein small (SMA) and Drosophila protein mothers against decapentaplegic (MAD); TGF-β: transforming growth factor-beta; HSP90: heat shock protein 90; GC: guanylate cyclase. B, Both DAF-11 and DAF-21 activate daf-7 expression via inhibiting HSF-1. DAF-7 abundantly accumulates in the two ASI neurons (white arrows) in wild-type L1 larvae. In contrast, daf-7 is strongly downregulated in daf-11(m47) and daf-21(p673) mutants (only a faint daf-7 expression is visible in the ASIs). HSF-1 deficiency suppresses daf-7 repression in daf-11(m47) and daf-21(p673) mutant genetic backgrounds. At 20°C, mutations in hsf-1 do not significantly alter daf-7 expression. This implies that HSF-1 has no or a weak activity at this temperature. C, Quantification (mean value) of daf-7::gfp expression in the ASIs in wild-type and mutant genetic backgrounds. *indicates p<0.0001. D, Dauer development in both daf-11(m47) and daf-21(p673) mutants requires HSF-1 activity. Mutational inactivation of hsf-1 largely protects daf-11(m47) and daf-21(p673) mutant animals from developing as dauer larvae. In each single mutant vs. double mutant comparison, p<0.0001. E, Transcriptional activity of daf-7 depends on the ambient temperature and HSF-1 activity. qRT-PCR analysis shows that daf-7 transcript levels decrease at 27°C (the left panel; p<0.001), as compared with those measured at 20°C, and this response requires HSF-1 activity (the right panel; p=0.993). In agreement with these results, hyperactivation of HSF-1 decreases daf-7 mRNA levels at 20°C (bottom; p<0.001). hsf-1(gf) represents a hyperactivating effect of an integrated hsf-1 transgene (hsf-1 cDNA). F, Epistasis model showing that DAF-11 and DAF-21 stimulate DAF-7 activity via inhibition of HSF-1. Thus, HSF-1 is an upstream component of the TGF-β cascade; it represses daf-7, thereby promoting dauer development at high temperatures. In fluorescent figures, images were captured with the same exposure time, and animals were examined at the L1 stage. N indicates number of animals tested, bars represent S.E.M. For statistics: Student’s t-test.

Figure 2

The upstream regulatory region of daf-7 contains a conserved HSF-1 binding site that is responsive in vivo. A, The structure of daf-7; purple boxes represent exonic sequences; grey boxes indicate upstream and downstream regulatory sequences. The position of this conserved HSF-1 binding site is indicated by the red arrow. B, This site is highly conserved in the daf-7 genomic environment of Caenorhabditis species. Red letters indicate strictly conserved sequences within the consensus HSF-1 binding site. C, Scheme of a transcriptional fusion gfp-labeled daf-7 reporter (pdaf-7::gfp) driven by a 3.8 kb long upstream regulatory sequence. A mutated version of the reporter (p_mutdaf-7::gfp) lacking several nucleotides from the potential HSF-1 binding site (grey letters in the next upper panel) is also shown. D, The daf-11(m47) mutation strongly represses the expression of the wild-type reporter (the repression occurs with full penetrance). The promoter-mutated construct, however, is not responsive to the daf-11(m47) mutation: m47 fails to downregulate p_mutdaf-7::gfp. 82% of these transgenic daf-11(m47) animals showed strong (wild-type levels) daf-7 expression (N=264). This suggests an in vivo functionality for this particular HSF-1-binding element.

Figure 3

HSF-1 acts downstream of DAF-7 to regulate lifespan. HSF-1 is required for the long-lived phenotype of mutant animals defective for TGF-β signaling. HSF-1 deficiency suppresses longevity in daf-7(e1372) mutants. Kaplan-Meyer survival curves were generated by the SPSS software. For comparing daf-7(−) single mutants vs. daf-7(−); hsf-1(−) double mutants or the wild type; p<0.0001 (see the Methods).

Figure 4

HSF-1 also acts downstream of DAF-7 to regulate development. A, Inactivation of HSF-1 enhances, while its hyperactivation decreases, dauer development in daf-7(−) null mutant genetic backgrounds. Bars represent S.E.M. For each single mutants vs. double mutants or RNAi combination, p<0.001, except for the daf-7(m62) mutant background, where p<0.01 (Student’s t-test). For each genotype, at least 150 animals were tested. B, The structure of daf-9 gene encoding two isoforms. Boxes represent exons, connecting lines indicates introns. A conserved HSF-1 binding site (the red arrow) can be found in the second intron of the longer daf-9 transcript. C, This regulatory element is conserved in the daf-9 genomic environment of other Caenorhabditis species. Highly conserved nucleotides are in red. D, HSF-1 upregulates daf-9. Fluorescent images showing daf-9::gfp expression in a wild-type, a daf-11(m47) single mutant and a daf-11(m47); hsf-1(sy441) double mutant L2 larva at 20°C. HSF-1 deficiency suppresses hyperactivation of daf-9 in animals defective for DAF-11. Fluorescent images were captured with the same exposure time. 91% of the daf-11(−); hsf-1(−) double mutant animals displayed weak (wild-type levels) daf-9 expression (N=166). p<0.0001; Student’s t-test. Enhanced expression of a daf-9::gfp reporter at 25°C, as compared with that obtained at 20°C. HSF-1 is required for higher temperature-induced ectopic expression of daf-9. E, Our epistasis model showing regulatory interactions among HSF-1, DAF-7 and DAF-9. HSF-1 both promotes (through repressing daf-7) and inhibits (through upregulating daf-9) dauer formation. Thus, it acts both upstream and downstream of DAF-7 to modulate development.

Figure 5

HSF-1 interconnects insulin/IGF-1 and TGF-β signaling. A, Inactivation of HSF-1 causes upregulation of daf-7 in daf-2(e1370) mutant dauer larvae (this interaction is fully penetrant; N=462). In the hsf-1(sy441) mutant background, a strong ectopic expression of daf-7 is evident in neurons related to the ventral nerve cord (GFP-positive cells at the ventral side of the body). Up: differential interference contrast images, down: the corresponding fluorescent images. B, Activity of the IGF-1 receptor DAF-2 causes daf-7 upregulation via inhibiting HSF-1. The daf-2 mutation e1370 leads to downregulation of daf-7 at the L1 stage, as compared with the wild-type background. Note that decrease in daf-7 expression in daf-2(−) mutants is less robust than in daf-11(−) mutants. The hsf-1(sy441) mutation suppresses downregulation of daf-7 in daf-2(e1370) mutant L1 larvae. 91% of double mutant animals exhibited strong (wild-type levels) daf-7 expression (N=220). In panels A and B, the corresponding fluorescent images were captured with the same exposure time. C, Quantification of daf-7::gfp expression intensity in the ASIs. * indicates: p<0.0001; Student’s t-test. D, Epistatic relationships among DAF-2/IGF-1, HSF-1 and DAF-7/TGF-β. Arrows indicate activations, bars represent inhibitions.

Figure 6

Hyperactivation of HSF-1 enhances dauer development in daf-2(e1370) mutant background. This interaction indicates that DAF-2 inhibits HSF-1, which in turn represses daf-7. Thus, HSF-1 links the insulin/IGF-1 and TGFβ pathways to control the developmental decision between normal reproductive growth vs. dauer larva formation. This epistasis model can explain how starvation induces, while nutrient availability suppresses, dauer development. Bars represent S.E.M. At both temperatures examined, daf-2(−) vs. daf-2(−); hsf-1(gf) comparison: p<0.001, student t-test.

Figure 7

HSF-2 antagonizes HSF-1 to influence development. A, The structure of hsf2. Blue boxes and interconnecting lines represent exonic and intronic sequences, respectively. The red line represents the extent of tm4607, which is a deletion removing upstream regulatory sequences and the first exon of hsf2. B, tm4607 promotes dauer development in daf11(−) mutants. Thus, HSF-2 acts downstream of, or in parallel to, DAF-11 to influence dauer formation, and antagonizes HSF1 in this function. hsf2(tm6407) single mutant animals are superficially wild-type; they exhibit neither a dauer formation constitutive phenotype at 25-27°C nor a long-lived phenotype. p<0.0001, Student’s t-test. C, Inactivation of hsf2 enhances dauer development in unc3(−) mutant (p<0.001), but not in unc31(−) mutant background (p=0.46), Student’s t-test. D, Our current model showing how HSF-1 interconnects insulin/IGF-1, cGMP and TGF-β signaling to control development and longevity. Arrows indicate activations, bars represents inhibitory regulations. Downstream of HSF-1, regulatory inputs affecting longevity are indicated by green coloring, and regulatory inputs on development are indicated by black arrows and bars. IGF-1: insulin/IGF-1 signaling, TGF-β: TGF-β signaling, cGMP: cGMP signaling. Grey arrows and the question mark indicate that the epistatic position of hsf2 in this signaling network is uncertain. hsf2 may act upstream of either unc31 or daf11. Regulatory inputs shown here do not necessarily represent direct interactions. For example, DAF-3 modulates DAF-2 activity (the green dotted bar) through its regulation of the INS-7 (agonist) and INS-18 (antagonist) insulin-like peptides encoding genes [5,22]. In panels B and C, bars represent S.E.M.