End-of-life targeted auxin-mediated degradation of DAF-2 Insulin/IGF-1 receptor promotes longevity free from growth-related pathologies

Preferably, lifespan-extending therapies should work when applied late in life without causing undesired pathologies. However, identifying lifespan-extending interventions that are effective late in life and which avoid undesired secondary pathologies remains elusive. Reducing Insulin/IGF-1 signaling (IIS) increases lifespan across species, but the effects of reduced IIS interventions in extreme geriatric ages remains unknown. Using the nematode C. elegans, we engineered the conditional depletion of the DAF-2/insulin/IGF-1 transmembrane receptor using an auxin-inducible degradation (AID) system that allows for the temporal and spatial reduction in DAF-2 protein levels at time points after which interventions such as RNAi may lose efficacy. Using this system, we found that AID-mediated depletion of DAF-2 protein efficiently extends animal lifespan. Depletion of DAF-2 during early adulthood resulted in multiple adverse phenotypes, including growth retardation, germline shrinkage, egg-retention, and reducing offspring. By contrast, however, AID-mediated depletion of DAF-2 specifically in the intestine resulted in an extension of lifespan without these deleterious effects. Importantly, AID-mediated depletion of DAF-2 protein in animals past their median lifespan allowed for an extension of lifespan without affecting growth or behavioral capacity. Thus, both late-in-life targeting and tissue-specific targeting of IIS minimize the deleterious effects typically seen with interventions that reduced IIS, suggesting potential therapeutic methods by which longevity and healthspan can be increased in even geriatric populations.


Introduction 46
The goal of aging research or geroscience is to identify interventions that promote 47 health during old age (B. K. Kennedy  pathways discovered was the insulin/insulin-like growth factor (IGF)-1 signaling 52 pathway (reviewed in (Kenyon, 2010)). Reducing insulin/IGF-1 signaling (IIS) 53 increases lifespan across species (Kenyon, 2010). Mice heterozygous for the IGF-1 54 receptor, or with depleted insulin receptor in adipose tissue, are stress-resistant and 55 long-lived (Blüher et al., 2003;Holzenberger et al., 2003), for example, and several 56 single-nucleotide polymorphisms in the IIS pathway have been associated with human 57 longevity (Kenyon, 2010). Moreover, gene variants in the IGF-1 receptor have been 58 associated and functionally linked with long lifespans in human centenarians (Suh et 59 al., 2008). This suggests that a comprehensive understanding of this pathway in 60 experimental, genetically-tractable organisms has promising translational value for 61 promoting health in elderly humans. However, whether or not reducing insulin/IGF-1 62 signaling during end-of-life stages can still promote health and longevity in any 63 organism is unknown. Therefore, we turned to the model organism Caenorhabditis 64 elegans to investigate whether reducing IIS during old age was sufficient to increase 65 lifespan. 66

67
The groundbreaking discovery that a single mutation in daf-2, which is the orthologue 68 of both the insulin and IGF-1 receptors (Kimura et al., 1997), or mutations in 69 aging, such as reduced brood size, small body size, and germline shrinkage 12,15,[19][20][21][22][23] . 96 To overcome this, RNA interference of daf-2 can be applied, which increases lifespan 97 without dauer formation during development and circumvents induction of dauer-98 associated phenotypes during adulthood (Dillin et al., 2002;Ewald et al., 2015;2018;99 S. Kennedy et al., 2004). However, the increase in lifespan by RNAi of daf-2 is only 100 partial compared to strong alleles such as daf-2(e1370) (Ewald et al., 2015). 101 Furthermore, adult-specific RNAi knockdown of daf-2 quickly loses its potential to 102 increase lifespan and does not extend lifespan when started after day 6 of adulthood 103 (Dillin et al., 2002), after the reproductive period of C. elegans. Whether this is due to 104 age-related functional decline of RNAi machinery or residual DAF-2 protein levels --or 105 whether the late-life depletion of daf-2 simply does not extend lifespan --remains 106 unclear. As such, use of an alternative method to reduce DAF-2 levels, beyond RNAi 107 or daf-2 mutation may allow us to more clearly uncouple the pleiotropic effects of 108 reduced IIS during development from those that drive daf-2-mediated longevity during 109 late adulthood. 110

111
To this end, we used an auxin-inducible degradation (AID) system to induce the 112 depletion of the degron-tagged DAF-2 protein with temporal precision (Zhang et al., 113 2015). The Arabidopsis thaliana IAA17 degron is a 68-amino acid motif that is 114 specifically recognized by the transport inhibitor response 1 (TIR1) protein only in the 115 presence of the plant hormone auxin (indole-3-acetic acid) (Dharmasiri et al., 2005). 116 Although cytoplasmic, nuclear, and membrane-binding domain proteins tagged with 117 degron have been recently shown to be targeted and degraded in C. elegans (Beer et 118 al., 2019;Zhang et al., 2015), to our knowledge, the AID system has not been used 119 previously to degrade transmembrane proteins, such as the DAF-2 insulin/IGF-1 receptor. We find that using AID effectively degrades DAF-2 protein and promotes 121 dauer formation when applied early in development. Dauer-associated phenotypes are 122 present in adults when AID of DAF-2 is applied late in development. Some of these 123 adulthood dauer-traits are induced by the loss of daf-2 in neurons, but others seem to 124 be caused by the systemic loss of daf-2. More importantly, the post-developmental, 125 conditional degradation of DAF-2 protein extends lifespan without introducing dauer-126 like phenotypes. Remarkably, we demonstrate that when half of the population has 127 died at day 25 of adulthood, AID of DAF-2 in these remaining and aged animals is 128 sufficient to promote longevity. Our work suggests that therapeutics applied at even

159
Dietary changes modulate endogenous DAF-2/Insulin/IGF-1 receptor abundance 160 Next, we monitored endogenous DAF-2 protein levels under different environmental 161 conditions, such as temperature and diet. Previously, Kimura and colleagues used 162 DAF-2 antibody immunostaining of whole animals and reported that mutant DAF-163 2(e1370) protein is present at 15°C but barely detected at 25°C, whereas mutant DAF-164 2(e1370) protein in a daf-16 null background or wild-type DAF-2 protein persists at 165 both 15°C and 25°C (Kimura and Riddle, 2011). By contrast, upon 24 hours of 166 starvation, the DAF-2 receptor is no longer detectable by using immunofluorescence 167 in fixed C. elegans (Kimura and Riddle, 2011). Since the FOXO transcription factor 168 daf-16 is the transcriptional output of daf-2 signaling Gems et al., 169 1998), these results suggest that DAF-2 protein levels may be autoregulated by insulin/IGF-1 signaling and might be influenced by temperature and food availability. 171 We first asked whether our DAF-2::degron::3xFLAG tag allows quantification of 172 endogenous DAF-2 levels. We observed comparable wild-type DAF-173 2::degron::3xFLAG levels across temperatures (15-28°C; Figure 1E), indicating that 174 temperature does not influence DAF-2 levels in wild type. Intriguingly, however, we 175 found that using FLAG-HRP antibodies to monitor protein levels, levels of DAF-2 176 protein almost completely disappeared after 36 to 48 hours of starvation ( Figures 1F,  177 1G). In keeping with this result, well-fed animals, for which we added 1% glucose into 178 the bacterial diet (OP50), increased the DAF-2 protein levels ( Figures 1F, 1G). These 179 effects were also influenced by the specific strain of E. coli used in each experiment: 180 when C. elegans were fed HT1115 (L4440), DAF-2 levels did not decrease after 24 or 181 48 hours of starvation, suggesting that the nutritional composition of the animal's diet 182 prior to starvation influences DAF-2 stability ( Figure 1H). Therefore, we conclude that 183 different food sources and dietary cues control not only the secretion of insulin-like 184 peptides to regulate DAF-2 activity (Pierce et al., 2001), but also DAF-2 receptor 185 abundance directly. 186 187

Auxin-induced degradation of degron-tagged Insulin/IGF-1 receptor 188
In C. elegans, cytosolic degron-tagged proteins are almost completely degraded after 189 30 minutes of auxin treatment (Zhang et al., 2015). However, the degradation of 190 transmembrane proteins using AID in vivo has not been previously reported. We 191 hypothesized that C. elegans might exhibit similar kinetics of degradation of a 192 transmembrane protein following auxin treatment. In keeping with that hypothesis, after 193 30 minutes of 1 mM auxin treatment, we observed a dramatic decrease in 194 transmembrane DAF-2 protein abundance ( Figures 1I, 1J). Levels of DAF-2 were only 195 slightly further reduced by continued auxin treatment, as indicated at 4 h and 24 h time 196 points ( Figures 1I, 1J). After 24 hours of 1 mM auxin treatment, we observed only a 197 40% total decrease in DAF-2 protein abundance, rather than a complete loss (Figures 198 1C,1D). Similar kinetics in the degradation of DAF-2::degron::3xFLAG levels were 199 confirmed using additional FLAG and degron antibodies (DataSource File 2). Taken 200 together, these results suggest that our AID system allows for the partial, rapid  Even more surprising was that dauer formation was also observed at 15°C and 20°C 243 with complete penetrance (Figures 2G, 2H). We found that dauer formation was 244 temporally related to developmental speed at a given temperature: At 15°C, it took six 245 days; at 20°C, it took four days; and at 25°C, it took three days to form dauers ( Figure  246 2H). We verified that all auxin-induced DAF-2::degron dauers showed dauer-specific 247 characteristics, such as SDS resistance, cessation of feeding, constricted pharynxes, 248 and dauer-specific alae ( Figure 2H, Supplementary Figures 2C-E), suggesting a 249 complete dauer remodeling and transformation. Thus, AID of DAF-2 promotes 250 complete dauer formation in a mechanism that is independent of temperature, but is 251 dependent on DAF-2 protein abundance. 252 253

Dauer commitment at mid-larval stage one upon DAF-2 degradation 254
Wild-type animals enter the pre-dauer L2d stage, where they keep monitoring their 255 environment before completely committing to dauer formation (Hu, 2007;Karp, 2018). 256 However, previous temperature-shifting experiments (from 15°C to 25°C) with daf-2 257 mutants suggested a dauer decision time-window from L1 to L2 stage before the L2d 258 stage (Swanson and Riddle, 1981). The AID allows for precise temporal degradation 259 of DAF-2. We pinpointed the dauer entry decision to mid-L1 by shifting synchronized 260 The FOXO transcription factor DAF-16 is required for dauer formation in daf-2 mutants. 270 We crossed DAF-2::degron with DAF-16::degron (Hobert, 2019) and found that daf-16 271 was required for dauer formation and developmental speed alterations after DAF-272 2::degron depletion (Supplementary Figure 3D). Previously reports suggest that many are not daf-16 dependent (Patel et al., 2008). Surprisingly, we observed no embryonic 278 lethality or L1 arrest in the progeny of animals placed animals on 1 mM auxin as L4s. 279 Similar results were seen using either the DAF-2::degron or DAF-2::degron; germline 280 TIR1 strains. Taken together, these results suggest that, although there is still DAF-281 2::degron::3xFLAG detected with the FLAG-HRP antibody after 24 hours of auxin 282 treatment, the persistent pool of DAF-2 receptor is insufficient to provide the wild-type 283 functions. Higher concentrations of auxin treatment lead to toxicity in both wild type 284 and DAF-2::degron animals (Supplementary Figure 3E). Inactivation of DAF-2 by the 285 AID is 100% penetrant for dauer formation at any temperature. Still, the absence of 286 embryonic lethality or L1 arrest at 1 mM auxin suggests that DAF-2::degron functionally 287 is more similar to a non-conditional and severe loss-of-function mutation than a null 288 allele. 289 290

Enhanced lifespan extension by AID of DAF-2 in adult animals 291
Given the strong phenotypic effects of DAF-2 AID on animal development, we next 292 explored whether DAF-2 degradation by AID could affect the function of adult animals. 293 Previous studies indicate that reducing IIS either by daf-2 RNAi knockdown or in 294 genetic mutants increases lifespan at any temperature (15-25°C) 295 2015; Gems et al., 1998). We hypothesized that AID-dependent degradation of DAF-296 2 would have similar effects on the lifespan of animals. We found that auxin 297 supplementation of DAF-2::degron animals, starting from L4, resulted in a 70-135% 298 lifespan extension ( Figure 3A; Supplementary Table 1). DAF-2 degradation using 1 299 mM auxin surpassed the longevity of commonly used daf-2(e1368) and daf-2(e1370) 300 mutants ( Figure 3A, Supplementary Table 1). By contrast, auxin-treatment at 0.1 mM 301 or 1 mM concentration had little or no effect on wild-type lifespan ( Figure 3A; 302 Supplementary Table 1). These results suggest that auxin-induced degradation of daf-303 2 is a powerful tool to promote longevity.  AID of DAF-2::degron starting from L4 reduced brood size at both 15°C and 25°C 336 ( Figure 3E). This suggests that lower body and brood size manifest as non-conditional 337 traits, in keeping with insulin/IGF-1's role as an essential gene for these functions. In 338 summary, these results suggest that some dauer-associated phenotypes or 339 pathologies can be induced during adulthood independent of temperature and that 340 passing through L2d is not required for dauer-associated phenotypes in adult animals. The pleiotropic effects of DAF-2 have been ascribed to tissue-specific effects of DAF-345 2 function. DAF-2 protein levels are predominantly found in the nervous system and 346 intestine, and to a lesser extent in the hypodermis (Kimura and Riddle, 2011), while 347 daf-2 mRNA expression has also been detected in the germline (Han et al., 2017;348 Lopez et al., 2013). Importantly, dauer formation in daf-2(e1370) can be restored by 349 expressing wild-type DAF-2 only in neurons (Wolkow et al., 2000). We hypothesized 350 that select tissues might drive these dauer-associated phenotypes. To test this, we 351 expressed TIR1 specifically in muscles, neurons, and intestine driven by the myo-3, 352 rab-3, and vha-6 promoters, respectively (Materials and Methods, Supplementary 353 Table 2). TIR1 expressed from any of these three tissue-specific promoters did not Previously, transgenic expression of wild-type copies of DAF-2 in neurons or intestine 367 was shown to partially suppress the longevity of daf-2(e1370) mutants at 25°C 368 (Wolkow et al., 2000). Taking advantage of our unique AID system, we wanted to ask whether the degradation of DAF-2 in a single tissue would be sufficient to induce 370 longevity. We found that either neuronal or intestinal depletion of DAF-2 alone was 371 sufficient to extend lifespan, although not to the extent as when DAF-2 is degraded in 372 all tissues ( Figure 4A; Supplementary Table 1). Therefore, we asked whether tissue-373 specific DAF-2 degradation was also sufficient for stress resistance seen in daf-2 374  Table 1). This suggests 398 that skn-1 may function independently from dauer-associated reprogramming 399 pathways at 15°C. Furthermore, the increased longevity seen in DAF-2::degron 400 animals may result from a differential transcriptional program at higher temperatures 401 compared to lower temperatures. 402 403

Late-life application of AID of DAF-2 increases lifespan 404
Finally, we asked whether it would be possible to promote longevity in extreme geriatric 405 animals by depleting DAF-2 by AID. Previous studies using RNAi indicated that 406 reduced daf-2 expression extended lifespan when started at day 6 of adulthood but not 407 later (Dillin et al., 2002), raising the question of whether daf-2-longevity induction is 408 possible beyond the reproductive period (day 1-8 of adulthood). To address this, we 409 maintained DAF-2::degron animals on control plates and shifted them to 1 mM auxin 410 containing plates at day 0 (L4) up to day-20 of adulthood ( Figure 5A). We found that 411 shifting the animals past the reproductive period at day 10 and day 12 still led to an   It is well-established that DAF-2/insulin/IGF-1 receptor signaling connects nutrient 440 levels to growth and development (Murphy, 2013). This is attributed to insulin-like 441 peptides binding DAF-2 and activating a downstream phosphorylation kinase cascade 442 that adapts metabolism (Murphy, 2013). Surprisingly, we find that DAF-2 receptor results in dauer traits at higher temperatures, one might target intestinal DAF-2 for 515 degradation to uncouple from any dauer traits. Yet, DAF-2 is essential for growth. We 516 find the best time point for DAF-2 inhibition is rather late in life to by-pass these 517 undesired side-effects to promote longevity.
We find that as late as day-25 of adulthood, when about 50% of the population has 520 died, AID of DAF-2 is sufficient to increase lifespan. The only application that was able 521 to also increase the lifespan that late was transferring old C. elegans to culturing plates In summary, we have demonstrated that late-life interventions can increase lifespan. 577 We have established that auxin-induced degradation is suitable for targeting 578 transmembrane receptors for non-invasive manipulations during developmental and 579 longevity in-vivo studies. Using the AID of DAF-2, we reconciled a longstanding 580 question by providing evidence that dauer-traits are not a spill-over of reprogrammed 581 physiology from developing L2d pre-dauers. Instead, the essential growth-related 582 functions of DAF-2 are causing deficits when applied during development or growth 583 phases. We have shown that tissue-specific or interventions beyond reproduction or 584 growth extend lifespan without pathology or deficits. Degradation of DAF-2/insulin/IGF-585 1 receptor might not be an artificial intervention since DAF-2/insulin/IGF-1 receptor 586 abundance is read-out to adapt metabolism to the environment and food status. E3 587 ligases might play an important role in regulating DAF-2/insulin/IGF-1 receptor levels. s.d. Two-sided t-test was used for statistical analysis. *: p < 0.05, **: p < 0.01, 606 ***: p < 0.001.

607
(E) Immunoblot analysis of DAF-2::degron animals showed no decrease of DAF-2 608 levels at high temperatures. Animals were raised at 15°C and put as L4 for 24 609 hours at the corresponding temperature. shifted as L4 to plates containing DMSO or 1 mM auxin at 20°C.

753
(B) 1 mM auxin treatment leads to the "dauer-associated" phenotype egg retention 754 in DAF-2::degron. Animals were raised at 15°C and shifted as L4 to 25°C or 755 kept at 15°C. Two days later, at 25°C and three days later at 15°C, the animals 756 were checked for egg retention. The experiment was performed 3 independent 757 times. Error bar represents s.d. ****: p < 0.0001 758 (C) Gonads were shrunk after 1 mM auxin treatment in DAF-2::degron animals. 759 Animals were raised at 15°C and shifted as L4 to 25°C or kept at 15°C. Two 760 days later at 25°C and three days later at 15°C, the animals were checked for 761 gonad size. The experiment was performed 3 independent times. ****: p < 762 0.0001 763 (D) 1 mM auxin treatment decreased the body size of DAF-2::degron animals at 764 15°C and 25°C. Animals were raised at 15°C and shifted to 1 mM auxin or 765 DMSO plates at the L4 stage. The experiment was performed 3 independent 766 times. ****: p < 0.0001 767 (E) 1 mM auxin treatment of DAF-2::degron resulted in a smaller brood size at 15°C 768 and 25°C. Animals were shifted to 1 mM auxin or DMSO plates at the L4 stage 769 and either kept at 15°C or moved to 25°C. The experiment was performed 3 770 independent times. ****: p < 0.0001 771 (F) Tissue-specific depletion of DAF-2 in neurons caused egg retention phenotype. 772 Animals were raised at 15°C and shifted from L4 to 25°C. Two days later, the 773 animals were checked for egg retention. The experiment was performed 3 774 independent times. Error bar represents s.d. ****: p < 0.0001 775 (G) Gonads were shrunk after neuronal depletion of DAF-2. Animals were raised at 776 15°C and shifted from L4 to 25°C. Two days later, the animals were checked 777 for gonad size. The experiment was performed 3 independent times. ****: p < 778 0.0001 779 For (A-G), see Data Source File 1 for raw data and Supplementary Table 1  For (A-D), see Supplementary Table 1 and Data Source File 1 for raw data, statistics, 850 and additional independent trials. 851 852

Strains 855
All strains were maintained on NGM plates and OP50 Escherichia coli at 15°C as 856 described. The strains and primers used in this study can be found in Supplementary 857  antibodies were used to detect the proteins by enhanced chemiluminescence (Bio-Rad 928 #1705061). Quantification of protein levels was determined using ImageJ software and 929 normalized to loading control (Tubulin). Statistical analysis was performed by using 930 either a two-tailed or one-tailed t-test. All western blots and quantifications can be 931 found in Data Source File 1 and 2. Adult C. elegans were placed on 1 mM auxin plates (for "DAF-2::degron") or DMSO 963 plates (for "DAF-2::degron", daf-2(e1368) and daf-2(e1370)) and shifted to 25°C.