Glucose availability impacts proteotoxic stress in Caenorhabditis elegans

Alterations in protein folding may lead to aggregation of misfolded proteins, which is strongly correlated with neurotoxicity and cell death. Protein aggregation has been shown as a normal consequence of aging, but it is largely associated with age-related disease, particularly neurodegenerative diseases like Huntington disease (HD). Huntington disease is caused by a CAG repeat expansion in the huntingtin gene and serves as a useful model for neurodegeneration due to its strictly genetic origin. Research in the model organism Caenorhabditis elegans suggests that glucose protects against cell stress, including proteotoxicity related to aggregation, despite the well-known, lifespan-shortening effects of glucose. We hypothesized that glucose could be beneficial by alleviating energy deficiency, a well-characterized phenomenon in HD, or by upregulating stress resistance pathways. We used C. elegans expressing polyglutamine repeats to quantify lifespan, motility, reproduction, learning, and activity of succinate dehydrogenase (SDH), with and without glucose, to identify the role of glucose in proteotoxicity and neuroprotection. Our data show HD worms on glucose plates exhibited shorter lifespans, no change in motility, learning, or SDH product formation, but had altered reproductive phenotypes similar to dietary restriction. Additionally, worms expressing toxic polyglutamine repeats were unable to learn association of food with a neutral odorant. We also observed tissue-specific differences; polyglutamine appeared to be slightly more toxic to muscle cells than neurons. Rather than increasing energy production, glucose appeared to decrease mitochondrial metabolism, as SDH formation decreases with added glucose. Future work investigating glucose-mediated neuroprotection should focus on connecting metabolism, sirtuin activation, and DAF-16 activation.

Introduction 81 lifespan, motility, reproduction, SDH product formation, and learning in model organisms expressing 82 expanded (40Q), non-toxic (24Q), and threshold (35Q) polyglutamine repeats. We found that while 83 the presence of glucose failed to improve the resistance to proteotoxic stressors for poly-Q worms, 84 through the simultaneous use of multiple different polyglutamine strains of C. elegans, some tissue-85 specific differences were shown. Glucose effects on reproduction were surprising, as we saw no 86 significant effect of glucose on the brood size or egg-laying pattern of N2 animals, contrary to that 87 seen in studies using lower concentrations of glucose [14,15]. Glucose effects on two polyglutamine 88 strains, however, mirror the reduced brood size and delayed egg-laying phenotype of worms under 89 DR conditions [20,21]. Using a butanone food-association assay, we showed HD worms to be 90 deficient in learning, even as 1-day-old adults, before other proteotoxic phenotypes are apparent.
91 Glucose failed to have a measurable effect on this learning phenotype or on SDH activity.

C. elegans Strains
94 Caenorhabditis elegans strains were acquired from the Caenorhabditis Genetics Center at the 95 University of Minnesota. C. elegans strains used in this project were: the N2 Bristol (wild-type), 96 AM101 (rmls110), AM138 (rmls130), AM140 (rmls132), and AM141 (rmls133) strains. The 97 polyglutamine strains (AMXXX) express a construct that contains a polyglutamine repeat of varying 98 length tagged by yellow fluorescent protein (YFP), further differentiated by the promotor-driven 99 tissue localization. The AM101 strain has a length of 40Q and the construct is expressed pan-100 neuronally. The AM138 strain has a length of 24Q, expressed in the body-wall muscle cells. The 101 AM140 strain has a length of 35Q and is expressed in the body-wall muscle cells; interestingly, the 102 AM140 strain's polyglutamine fragment is on the verge of the toxic threshold and only older worms 103 show aggregation. The AM141 strain has a length of 40Q and is expressed in the body-wall muscle 127 The movement of N2 and polyQ worms was measured on plates with (G) or without (NGM) glucose 145 Every day thereafter, the worm was transferred onto a new plate of the same condition. The progeny 146 on the old plates were counted after a hatching period of 24 hours. Only viable progeny were scored 147 for this experiment; unhatched eggs were excluded. This continued until no progeny were found on 148 the plate for 2 consecutive days. Data are from 3 separate trials, with n=10 worms per strain for each 172 The ability of N2 and polyQ worms to associate a normally neutral odorant (butanone) to food (E. 173 coli) was measured after exposure to the odorant in the presence of food. Parent worms were 174 transferred to new NGM or glucose plates spotted with 500µL of OP50 E. coli and allowed to 175 reproduce for 8-16 hours, then removed so that progeny were age synchronized. The learning assay 176 was performed according to previously described methods [22].
177 Chemotaxis index was scored by the number of worms at each location and calculated as: CI = 178 (butanone -EtOH)/(total -origin). Statistical significance was analyzed with JMP software version 179 12 using a two-way ANOVA. 188 As expected, the presence of a toxic polyglutamine repeat did shorten the lifespan of C. elegans.
189 Although the lifespan of the pan-neuronal 40Q worms on NGM was not significantly shortened 190 compared to N2 (Figures 1A,B; log-rank test, p = 0.0137), there was a strong trend toward this. The 191 lifespan of the body-wall muscle 40Q strain on NGM plates was significantly shortened ( Figure 1C; 192 log-rank test, p < 0.0001). The 24Q worms had an increased lifespan on NGM plates compared to 193 N2/wildtype ( Figure 1D; log-rank test, p = 0.003), while the 35Q worms exhibited a reduced lifespan 220 Motility 221 The spontaneous movement of wild-type and polyglutamine strains were measured to test the 222 hypothesis that glucose affects polyQ phenotypes based on increased energy availability. Toxic 223 polyQ worms should be less motile than wild-type due to energy deficiency, and glucose enrichment 224 could increase the energy available and provide protection. This assay was designed to observe 225 energy-output and relate output to energy fluctuations; therefore, worms were left undisturbed on the 226 plates, contrary to measurements of the ability to move. This protocol assumed that the amount of 227 spontaneous movement (without experimenter intervention) reflects the abundance of energy 228 available to spend on movement.
230 The N2 strain was more motile than the 40Q, 24Q, and 35Q body-wall muscle strains ( 248 The wild-type strain of C. elegans lays many eggs over a relatively short period of time (~300 over 7 249 days, Figure 3A). The addition of a toxic polyglutamine repeat, regardless of localization, decreases 250 the brood size ( Figures 3B,C). Contrary to expectations, the addition of glucose decreased the brood 251 size in both 40Q strains (Figure 3). However, glucose also increased the reproductive lifespan of the 252 worms, as worms on glucose plates continued to lay eggs for 1-2 days longer than worms on NGM 267 SDH product formation seemed to vary widely between strains and conditions, though there was a 268 detectable overall significant difference between groups (Figure 4; two-way ANOVA, F 9,20 = 2.79, p 269 = 0.026). The strain effect was the only significant effect in the model (F 4 = 3.53, p = 0.024), 321 was not different in the pan-neuronal 40Q strain compared to the N2 strain (Figure 4). One 322 explanation for this could be that increased ROS production negatively regulated or reduced activity 323 of SDH before product formation was measured [31].
324 Increased calcium influxes may also increase acetylcholine signaling at the neuromuscular junction, 325 leading to increased motility in the pan-neuronal 40Q strain (Figure 2; p = 0.0001). Interestingly, this 326 effect seems to be time-dependent; increased motility in this strain appears to be restricted to early in 327 life, although this was not tested statistically. Aberrant calcium signaling is unlikely to affect the 328 motility of the body-wall muscle 40Q strain because expression of the polyQ is restricted to muscles.
329 Neurons of the body-wall muscle 40Q strain should fire normally, and thus the incoming signal at the 330 neuromuscular junction should resemble that of wild-type. Dysfunction in the body-wall muscle 40Q 331 strain should originate in the muscle and this is supported by the motility of the body-wall muscle 332 40Q strain, which was decreased compared to the pan-neuronal 40Q and N2 strains (Figure 2; p < 333 0.0001 for both comparisons). The increase in motility in the pan-neuronal 40Q strain and the 334 decrease in motility in the body-wall muscle 40Q strain suggest that proteotoxicity manifests 335 differently in neurons than in muscles.
336 The tissue-specific effects of polyglutamine extended to other phenotypes: lifespan of the body-wall 337 muscle 40Q strain was slightly shorter than that of the pan-neuronal 40Q strain (Figure 1), though 338 this effect was not statistically significant. Motility in all three muscular polyglutamine strains was 339 decreased when compared to the pan-neuronal strain (Figure 2; p < 0.0001 for all comparisons).
340 Brood size was slightly decreased in body-wall muscle 40Q compared to pan-neuronal 40Q (Figure   341 3). SDH product formation was slightly decreased in polyQ muscles compared to wild-type, but not 342 neurons (Figure 4). Considering that both the pan-neuronal and body-wall muscle 40Q strains 343 express a repeat of the same length, these data suggest that muscles are slightly more vulnerable to