The neurodegenerative effects of selenium are inhibited by FOXO and PINK1/PTEN regulation of insulin/insulin-like growth factor signaling in Caenorhabditis elegans

Abstract

Exposures to high levels of environmental selenium have been associated with motor neuron disease in both animals and humans and high levels of selenite have been identified in the cerebrospinal fluid of patients with amyotrophic lateral sclerosis (ALS). We have shown previously that exposures to high levels of sodium selenite in the environment of Caenorhabditis elegans adult animals can induce neurodegeneration and cell loss resulting in motor deficits and death and that this is at least partially caused by a reduction in cholinergic signaling across the neuromuscular junction. Here we provide evidence that reduction in insulin/insulin-like (IIS) signaling alters response to high dose levels of environmental selenium which in turn can regulate the IIS pathway. Most specifically we show that nuclear localization and thus activation of the DAF-16/forkhead box transcription factor occurs in response to selenium exposure although this was not observed in motor neurons of the ventral cord. Yet, tissue specific expression and generalized overexpression of DAF-16 can partially rescue the neurodegenerative and behavioral deficits observed with high dose selenium exposures in not only the cholinergic, but also the GABAergic motor neurons. In addition, two modifiers of IIS signaling, PTEN (phosphatase and tensin homolog, deleted on chromosome 10) and PINK1 (PTEN-induced putative kinase 1) are required for the cellular antioxidant reduced glutathione to mitigate the selenium-induced movement deficits. Studies have suggested that environmental exposures can lead to ALS or other neurological diseases and this model of selenium-induced neurodegeneration developed in a genetically tractable organism provides a tool for examining the combined roles of genetics and environment in the neuro-pathologic disease process.

Keywords: Amyotrophic lateral sclerosis; Insulin/insulin-like signaling; Neurodegeneration; PINK1; PTEN; Selenium.

Supplemental Fig. I

Inhibition of proteasome does not prevent Se(IV)-induced damage to mitochondria. Upper panel: Animals of the strain CB5600 transiently express mitochondrial GFP in their muscles that became fragmented after Se(IV) exposure for 48 h (Fig. 1). This damage could not be prevented by inhibition of proteasomal degradation with either levamisole (Se + Lev) or MG-132 (Se + MG-132) when adult animals were exposed simultaneously to Se(IV) and either compound for 48 h. Lower panel: Black and white outline of GFP expression patterns above, nuclei are represented by asterisks for orientation. Lev = 1 mM levamisol; MG-132 = 200 µM Z-Leu-Leu-Leu-CHO (MG-132). Se = Se(IV).

Fig. 1

Mutations in the IGF pathway affect selenium sensitivity. Animals containing mutations in the daf-2 IGF pathway show an altered sensitivity to Se(IV)-induced motility effects that is enhanced by increasing temperature. (A) When grown at 20 °C, adult animals were found to be more resistant (#) or to exhibit significantly increased sensitivity (*) to Se(IV) in comparison to the WT strain by one-way ANOVA performed across each time point (p = 0.05; significance was determined by post hoc analysis using the Bonferroni–Holm method). (B) The phenotypes (sensitivity or resistance) observed at 20 °C were significantly enhanced by Se(IV)-exposure at 25 °C at some or all of the time points tested for all strains [^ significantly different at 25 ° when each strain was compared to the same strain and time exposure at 20 °C; p < 0.05 by Student's t-test (two tailed, unequal variance)]. Comparisons of the mutant strains to WT when all were grown at 25 °C, showed significant increases in resistance (#) or sensitivity (*) in the animals at some or all of the time points tested. Analysis was by one-way ANOVA performed across each time point (p = 1.4 × 10^-11; significance was determined by post hoc analysis using the Bonferroni–Holm method). When not exposed to Se(IV), all strains had =95% motility at each time point and temperature tested (data not shown), with the only exception being akt-1 which exhibited 80% motility across all time points at 25 °C only (data not shown). Each bar graph represents the average of three populations (n = 150 at 20 °C; n = 60 at 25 °C) with error bars indicating SEM. WT = N2 strain; Se = Se(IV). All mutations are reduction-of-function mutations with the exception of mg144 which is an activating mutation in the akt-1 gene.

Fig. 2

DAF-16 nuclear localization and dosage effects on motility with selenium exposure. Se(IV) induces DAF-16::GFP to translocate to the nucleus which is required for protection from the Se(IV)-induced effects on motility. (A–F) A strain (TJ356) containing a transgene (zIs356) that overexpresses wild-type DAF-16::GFP under control of the daf-16 promoter (Henderson and Johnson, 2001) was examined to determine its subcellular localization under various conditions. (A) Well-fed animals maintained at 20 °C show cytoplasmic expression of DAF-16::GFP. (B) Heat-shock-induced stress results in nuclear localization of DAF-16 (indicated by arrows) as previously described (Henderson and Johnson, 2001; Lin et al., 2001). (C) Se(IV) exposure induced DAF-16::GFP translocation after 2 h. (D) Exposure to Se(IV) for 2 h followed by heat shock increased nuclear translocation. (E and F) Longer exposures to Se(IV) (without heat shock) did not vastly increase localized GFP expression. All animals were maintained in a well-fed state at 20 °C and were exposed in the presence of food. Magnification = 100×. (G) In comparison to the WT strain, the presumptive null-mutation, daf-16(mu86) conferred significantly more sensitivity to the Se(IV)-induced effects on motility at 24 and 72 h [*, sensitivity to Se(IV), p = 8.0 × 10^-3] while animals overexpressing DAF-16 [daf-16(o/e) = zIs356] were significantly more resistant at 72 h [# resistance to Se(IV); p = 4.3 × 10^-6] by one-way ANOVA performed across each time point with significance determined by post hoc analysis using the Bonferroni–Holm method. Each graph bar represents at least three populations of 20 animals per strain (n = 60); error bars indicate SEM. Se = Se(IV).

Fig. 3

Selenium exposure and the effects of tissue-specific expression of DAF-16 on motility. Overexpression of DAF-16 in nerves, intestines, or muscles of adult animals resulted in an increased resistance to the Se(IV)-induced motility defects. (A) Animals expressing the DAF-16 tissue-specific constructs were more resistant (#) to the Se(IV)-induced effects on motility when compared to animals with the same genetic background (daf-16;daf-2) and that expressed the same reporter plasmid, pRF4. This plasmid contained within the yzIs71 transgene induces animals to roll in place (roller phenotype) rather than move forward or backward in a normal sinusoidal wave pattern, a phenotype that results from the expression of a dominant mutation (su1006) in the rol-6 gene (Mello et al., 1991). No significant difference (NS) was observed between the daf-16;daf-2 roller and non-roller strains at all temperatures, between the daf-16 and the daf-16;daf-2 roller strains at all temperatures, nor the daf-2 roller strain and the daf-16;daf-2 intestine expressing strain at 24 h only (as determined by one-way ANOVA). Significance is defined as p = 0.01 as determined by one-way ANOVA followed by post hoc analysis using the Bonferroni–Holm method. All non-Se(IV)-treated strains had a greater than 90% survival at all times tested (data not shown). Each bar graph represents the average of six populations of 20 animals per strain (n = 120); error bars ± SEM. daf-16 = daf-16(mu86); daf-2 = daf-2(e1370); daf-16;daf-2;yzIs71 = OM249; daf-16;yzIs71 = OM285; daf-2;yzIs71 = OM148. (B) Upper panel: Animals of the strain CB5600 transiently express mitochondrial GFP in their muscles and are shown here after 48 h of either mock- (dH₂O) or Se(IV)-(Se)exposure. The GFP pattern is tubular indicating healthy mitochondria in the water control (dH₂O). Se(IV)-exposure results in fragmentation of the GFP expression indicating damage to the muscle mitochondria (Se). Lower panel: Black and white outline of GFP expression patterns above, nuclei are represented by asterisks for orientation. Se = Se(IV).

Fig. 4

Selenium exposure and the dosage effects of DAF-16 on neurodegeneration. The level of DAF-16 expression affects the observed levels of neurodegeneration in the ventral cord. (A) The DOP-3::RFP construct (vsIs33) expresses in both cholinergic (small arrows) and GABAergic (large arrows) motor neurons, but more weakly expresses in the former than the latter (Chase et al., 2004). In animals with WT levels of DAF-16 [daf16(+) = strain OM261], the normal fusiform shape of the motor neuron cell bodies observed in the control animals [daf-16(+), dH₂O] is altered in animals exposed to Se(IV) [daf-16(+), Se]. Axonal blebbing of the ventral cord (line) is also observed with Se(IV) exposure, but not in the control animals; as was previously reported (Estevez et al., 2012). With loss of DAF-16 [daf-16(-) = strain OM325], neuronal damage is increased. Over-expression of DAF-16 [daf-16(o/e) = strain OM324] reduced the blebbing in the cord as well as cell body rounding. Cell loss was still observed in the DAF-16 over-expressing animals. This had been shown previously to occur in WT animals exposed to Se(IV) (Estevez et al., 2012). (B) Se(IV) does not induce DAF-16::GFP (zIs356) localization to the nucleus in either the cholinergic (small arrows) or GABAergic motor neurons (large arrows) of adult animals (strain = OM324), but was observed in muscle cells (M). Nuclear localization of DAF-16 in the hypodermal cells (H) of animals not exposed to Se(IV) (dH₂O) is most likely due to stress induced on the animals which were alive, but paralyzed while being imaged, as previously observed (Lin et al., 2001). Magnification = 400×. Se = Se(IV). The genotype of each strain is listed in Section 2. Animals shown here are representative of 30–50 animals examined under.

Fig. 5

The effects of mutations on downstream target genes of DAF-16 transcription. C. elegans encodes for two superoxide dismutase 2 (SOD2) genes which are regulated by DAF-16 and have opposite effects on the Se(IV)-induced movement deficits. (A) Animals with a mutation in the sod-2 gene are initially more sensitive to the effects of Se(IV) than WT animals (at 24 h: p = 3.8 × 10^-3 by one-way ANOVA), but at later time points are no different than WT animals (p = 0.5) while animals with a sod-3 mutation are consistently and significantly more resistant to Se(IV)’s effect on movement than WT (p = 0.03 by one-way ANOVA across all time points tested). The sod-2;sod-3 double mutant animals are phenotypically similar to the sod-2 single mutant animals, the animals are initially more sensitive than WT to Se(IV) at 24 h (p = 4.3 × 10^-4 by one-way ANOVA), but are no different at the later time points (p > 0.1). Animals were grown at 25 °C, but the trends and significances were the same for animals grown at 20 °C (data not shown). (B) Animals of the strain OM261 which expresses DOP-3::RFP and over-expresses a SOD-3::GFP translational fusion under control of its native promoter (wuIs56) were exposed to either water (dH₂O) or Se(IV) (Se) for 24 h. The SOD-3::GFP was observed to co-localize with DOP-3::RFP to the ventral cord and the motor neurons (large arrows = GABAergic; small = cholinergic), but was unable to protect either from the Se(IV)-induced neurodegenerative effects. This SOD-3::GFP expression is discontinuous in the Se(IV)-exposed motor neurons (Se, SOD-3::GFP, right panels) in comparison to the water exposed ones (H₂O, SOD-3::GFP, right panel) suggesting that mitochondrial fragmentation has occurred as was observed in the Se(IV)-exposed muscle (Fig. 3B). In the Se(IV) exposed animals, muscles are observed to exhibit increased expression of DOP-3::RFP that did not colocalize with the SOD-3::GFP expression (Se, white outline) and was not observed in the water controls. Smaller panels on the right focus in on the motor neurons that are boxed in the larger image immediately to their left. Magnification = 400×. Se = Se(IV).

Fig. 6

The cellular antioxidant glutathione and its effect on modulators and a downstream regulator of IGF signaling when exposed to selenium. Animals containing mutations in the genes encoding the C. elegans orthologs of PTEN (phosphatase and tensin homolog, daf-18) and PINK1 (PTEN-induced putative kinase, pink-1), modulators of the IGFR signaling are not rescued from the Se(IV)-induced motility defects by exposure to the cellular antioxidant glutathione (GSH), but the Se(IV)-sensitivity observed with the daf-16(mu86) mutation was rescued. All the strains were sensitive to Se(IV), but increased sensitivity to Se(IV) (*, p = 0.05) was not observed until 24 h of exposure in the animals with a mutation in daf-16, and at 72 h in the daf-16 and daf-18(e1375 and ok480) mutant animals when compared to wild-type (WT). Analysis was by one-way ANOVA performed across each time point [significance (indicated above bar graphs) was determined by post hoc analysis using the Bonferroni–Holm method; NS = not significant]. Glutathione was able to significantly rescue the Se(IV)-induced deficits in WT animals after 72 h of continuous exposure (significance is indicated above each pair of bars and was determined by Student's t-test). When not exposed to Se(IV), all strains had =95% motility at each time point and temperature tested (data not shown). The addition of GSH [without Se(IV)] did not alter motility in comparison to the control (water only exposed) animals (data not shown). Each bar graph represents the average of three populations (n = 60) with error bars indicating SEM. Se = Se(IV).

Fig. 7

Summary model of the effects of sodium selenite in C. elegans. Adult animals with mutations in components of the DAF-2/IGFR signaling cascade responded to high dose Se(IV) within their environment similar to that observed when exposed to other stressors (Honda and Honda, 2002), i.e. the loss- or reduction-of-function daf-16 alleles had an opposite phenotype from that of many of the upstream components that inhibit its nuclear translocation. Yet, although the pattern was predictable, the specific sensitivity or resistance phenotypes were not. A model depicting the effects of Se(IV) on the IGFR signaling cascade including modulatory branches is shown and summarizes the data presented here as well as previously by us (Morgan et al., 2010), and elsewhere (Luo et al., 2013; Akundi et al., 2012; Ponugoti et al., 2012; Kim et al., 2011; Loh et al., 2009; Mei et al., 2009; Budovskaya et al., 2008). The Se(IV) (selenite) responsive phenotypes, listed in parentheses as R (resistant) or S (sensitive), are based on the phenotypes of the reduction-of-function mutants. *R, resistance phenotype of reduction-of-function mutant is assumed since the mg144 gain-of-function mutation is sensitive (Fig. 1); Dotted lines represent pathways that are assumed based on data observed in other organisms, but not confirmed in C. elegans. IGFR = insulin/insulin growth factor-like receptor; PI3K = phosphatidylinositide 3-kinase; FOXO = forkhead box transcription factor; PTEN = phosphatase and tensin homolog; GSH = reduced glutathione; GSSG = oxidized glutathione; Ref. 1 = Ponugoti et al. (2012); Ref. 2 = Kim et al. (2011), Loh et al. (2009); Ref. 3 = Mei et al. (2009); Ref. 4 = Luo et al. (2013); Ref. 5 = Budovskaya et al. (2008); Ref. 6 = Akundi et al. (2012).