Direct and Indirect Gene Regulation by a Life-Extending FOXO Protein in C. elegans: Roles for GATA Factors and Lipid Gene Regulators

SUMMARY

In long-lived C. elegans insulin/IGF-1 pathway mutants, the life-extending FOXO transcription factor DAF-16 is present throughout the animal, but we find that its activity in a single tissue can delay the aging of other tissues and extend the animal’s life span. To better understand the topography of DAF-16 action among the tissues, we analyzed a collection of DAF-16-regulated genes. DAF-16 regulated most of these genes in a cell-autonomous fashion, often using tissue-specific GATA factors to direct their expression to specific tissues. DAF-16 could also act cell nonautonomously to influence gene expression. DAF-16 affected gene expression in other cells, at least in part, via the lipid-gene regulator MDT-15. DAF-16, and probably MDT-15, could act cell non-autonomously in the endoderm to ameliorate the paralysis caused by expressing Alzheimer’s Aβ protein in muscles. These findings suggest that MDT-15-dependent intercellular signals, possibly lipid signals, can help to coordinate tissue physiology, enhance proteostasis, and extend life in response to DAF-16/FOXO activity.

INTRODUCTION

Insulin and IGF-1 signaling pathways influence the rate of aging in many species, and they appear to affect human aging, as well (Barzilai et al., 2012; Kenyon, 2010). However, the mechanisms by which insulin and IGF-1 hormones coordinate the aging of individual tissues are poorly understood. In C. elegans, the lifespan extension produced by reduced insulin/IGF-1 signaling requires the FOXO transcription factor DAF-16 (Kenyon et al., 1993; Lin et al., 1997; Ogg et al., 1997). DAF-16/FOXO is an important longevity regulator, as its disruption accelerates the rate of normal aging (Garigan et al., 2002; Haithcock et al., 2005)—and increasing its activity can extend life span (Henderson and Johnson, 2001; Lee et al., 2001; Lin et al., 2001). Its function in life-span regulation may be ancient: forkhead-box transcription factors can extend life span in yeast (Postnikoff et al., 2012), and FOXO proteins can extend the life span of Drosophila (Giannakou et al., 2004; Hwangbo et al., 2004) and possibly humans (Barzilai et al., 2012).

DAF-16 extends the life span of C. elegans insulin/IGF-1-pathway mutants by affecting the expression of stress-response, metabolic, innate-immunity, signaling, germline, and other genes (Curran et al., 2009; Lee et al., 2003; McElwee et al., 2003; Murphy et al., 2003; Wang et al., 2008). Presumably, these genes hold much of the answer to the question of how life span can be extended; yet, we know little about their positions in this regulatory network. How are their activities distributed among the different tissues? In which tissues are their activities altered in long-lived mutants? Is each gene regulated directly by DAF-16, or are intermediate, possibly intercellular, factors required? These are fascinating, system-wide questions that address the function of this endocrine pathway as a whole.

DAF-16 is expressed in many tissues, raising the possibility that it directly regulates many genes in a strictly cell-autonomous fashion. Consistent with this idea, expressing daf-16(+) exclusively in the intestine, muscles or neurons of a daf-2(−) mutant switches on the sod-3 superoxide dismutase gene in that tissue alone (Libina et al., 2003). The sod-3 promoter contains consensus DAF-16/FOXO-binding elements (DBEs) (Biggs et al., 2001; Furuyama et al., 2000; Pierrou et al., 1994), which bind DAF-16 both in vitro (Furuyama et al., 2000) and in vivo (Oh et al., 2006).

DAF-16 activity can also influence cells at a distance. First, in a process we call FOXO-to-FOXO signaling, increasing daf-16 gene dosage in one tissue (neurons, intestine) upregulates DAF-16 activity in other tissues (Libina et al., 2003). Intestinal DAF-16 mediates FOXO-to-FOXO signaling, at least in part, by downregulating an intestinal insulin gene (Murphy et al., 2007). The C. elegans intestine is the animal’s entire endoderm, also functioning as the adipose tissue, liver, and pancreas. Consistent with this, overexpressing dFOXO specifically in the fat body of Drosophila reduces the expression of the insulin-like peptide gene dilp-2 in neurons and reduces insulin/IGF-1 signaling in peripheral tissues (Giannakou et al., 2004; Hwangbo et al., 2004). DAF-16 also appears to initiate a fundamentally different type of cross-tissue communication, one that does not require DAF-16 activity in responding tissues. Expressing daf-16 exclusively in the intestine of a daf-16(−); daf-2(−) double mutant extends life span by 50%–70%, and expressing daf-16 only in nonintestinal tissues extends life span by 50% (Libina et al., 2003). To a lesser extent, DAF-16 can also act exclusively in neurons (Libina et al., 2003) or skin (this study) to increase life span, as well. The finding that DAF-16 is not absolutely required in any one tissue to extend C. elegans’ life span implies that DAF-16 can extend life span by modulating expression of genes encoding downstream hormones or metabolites that act independently of daf-16 to delay aging in responding tissues. If so, then genes like sod-3, which require DAF-16 cell autonomously for their expression, might be the exception and not the rule.

To address these questions more systematically, in vivo, we surveyed longevity genes that are either up- or downregulated by DAF-16 in long-lived daf-2(−) mutants to ask (1) in which tissues they are expressed; (2) whether DAF-16 regulates them in a strictly cell-autonomous fashion, or remotely, at a distance; and (3) what mechanisms control their tissue-specific expression.

RESULTS

DAF-16 Can Affect Tissue Aging at a Distance

daf-2(−) mutants expressing daf-16 only in the intestine are long-lived (Libina et al., 2003) (Table S3 and Figure S6A), but is this increased longevity correlated with a more youthful appearance of individual tissues? Aging C. elegans muscles resemble those of human sarcopenia patients, in which muscle filaments fragment and break. We found that expressing daf-16 exclusively in the intestine of daf-16(−); daf-2(−) mutants reduced this muscle deterioration (Figures 1A and 6A) and improved body movement (Figure S6B). Thus, intestinal DAF-16 can act at a distance to delay the aging of daf-16(−) muscles.

Figure 1. DAF-16 Acts Cell Nonautonomously to Regulate Muscle Aging and Autonomously to Regulate mtl-1 and dod-17 Expression.

(A) Intestinal DAF-16 protects daf-16(−) muscles from age-dependent deterioration. Top: Extent of sarcomere degeneration on day 10 of adulthood (micrographs scored/animals analyzed) (see Supplemental Experimental Procedures) (Mann-Whitney-Wilcoxon test of all data points; n.s., not significant). Bottom: Percentages of animals with extensive degeneration (Class C, as defined by Herndon et al. [2002]) (Student’s t test of independent experiments). A gap in the sarcomere of a day 10 Class C animal is shown; 250X magnification. Scale bar: 50 µm. Note that intestinal daf-16(+) does not fully restore muscle quality to that seen in nontransgenic daf-16(+); daf-2(−) mutants, just as it does not extend life span to the extent seen in daf-16(+); daf-2(−) animals.

(B) Left panels: A mtl-1::rfp translational reporter is expressed mainly in the pharynx (ph) and intestine (i) of wild-type, and is upregulated in the intestines of daf-2(−) mutants in a daf-16-dependent manner. Right panels: intestinal GFP-tagged DAF-16 upregulates mtl-1::rfp only in the intestine. Muscle or neuronal daf-16 does not affect mtl-1::rfp expression. The body is outlined. daf-2(e1370) and daf-16(mu86) mutations were used. Young adults, 100X magnification. Scale bar: 130 µm.

(C) The ges-1 promoter, which drives gfp::daf-16 expression, is expressed in a mosaic fashion in the intestine (middle panel). DAF-16 turns on mtl-1::rfp expression in the same cells (left panel). Young adults, 100X magnification. Scale bar: 130 µm.

(D) Top panels: A nuclear-localized Pdod-17::rfp transcriptional reporter is expressed mainly in the intestine (i) of wild-type and is downregulated in the intestine of daf-2(−) mutants and upregulated in daf-16(−); daf-2(−) mutants. Bottom panels: expression of daf-16 in the intestine, but not muscles or neurons, suppresses Pdod-17::rfp expression. Inset: daf-16 expression (green) is inversely correlated with dod-17 expression (red). Note that Pdod-17::rfp expression is higher in an intestinal cell that does not express gfp::daf-16 (arrowhead). The body and pharynx (ph) are outlined. Young adults, 250X magnification. Scale bar: 50 µm. Representative images from 2 or more experiments are shown for all figures.

Figure 6. Effects of Intestinal DAF-16 and MDT-15 on Muscle Aging and Age-Related Disease.

(A) mdt-15 RNAi reduced the ability of intestinal DAF-16 to protect daf-16(−) muscles during aging. Top: average sarcomere degeneration indexes of RNAi-treated animals on day 12 of adulthood. Each point represents one animal (three to four images were taken for each animal). Higher number (y axis) represents more severe sarcomere degeneration. Mean degeneration index is indicated by the gray bar (Mann-Whitney-Wilcoxon test; n.s., not significant).

(B and C) Intestinal DAF-16 attenuated the toxicity of muscle-expressed Aβ protein. In (B), Kaplan-Meier survival analysis of transgenic animals in which human Aβ(1–42) is expressed in body-wall muscles (see Table S3). Log-rank test, ***p < 0.001. n.s., not significant. In (C), age-dependent Aβ aggregation-induced paralysis of the animals shown in Figure 6B (see Table S3). Log-rank test, ***p < 0.001. n.s., not significant.

(D) daf-16 or mdt-15 RNAi abolished the ability of intestinal DAF-16 to ameliorate toxic Aβ aggregation-induced paralysis. RNAi was initiated at the L4 stage, and RNAi-treated animals were scored for paralysis at room temperature (see Table S3). Log-rank test, ***p < 0.001. n.s., not significant.

(E) Model for daf-16-dependent gene regulation. First, DAF-16 can act directly on its target genes in a cell-autonomous fashion. Second, DAF-16 activity in one tissue (e.g., intestine, muscle) can stimulate downstream signaling pathways that act on daf-16(−) cells to influence gene expression, aging, and protein-aggregation toxicity (FOXO-to-FOXO(−) signaling). Analysis of MDT-15 suggests that downstream lipid signals may play a role in this signaling. Third, DAF-16 action in one tissue can affect DAF-16 activity elsewhere, for example, by feedback regulation of insulin genes (FOXO-to-FOXO signaling).

Analysis of DAF-16-Regulated Genes In Vivo

To better understand the DAF-16 regulon at the tissue level, we studied, in vivo, a diverse collection of DAF-16-regulated genes that we had identified in gene-expression arrays comparing long-lived daf-2(−) mutants to daf-16(−); daf-2(−) mutants (Murphy et al., 2003). Our set included stress-resistance, chaperone, signaling, innate immunity, and metabolic genes whose RNAi knockdown influenced life span, as well as additional genes for which RNAi analysis had not been performed.

We analyzed the expression of each gene by using ~1–3 kb upstream DNA to drive expression of red fluorescent protein (RFP), or, in a few cases, green fluorescent protein (GFP). We first examined each reporter for its response to daf-2 RNAi, which stimulates DAF-16/FOXO’s transcriptional activity. Twenty of the forty-four new genes we tested exhibited the predicted (up or down) response to daf-2 inhibition (Table S1). The 24 negatives could include microarray false positives, or genes influenced by regulatory elements downstream of the promoter. Consistent with this, the mtl-1 transcriptional fusion exhibited little or no response to daf-2 RNAi, but a translational fusion driven by the same promoter responded very strongly (Figures 1B and S1C).

Genes Upregulated in Long-Lived daf-2(−) Mutants

Fifteen of the twenty daf-2-sensitive reporters were upregulated under daf-2(−) conditions (Table S1). One, lys-7, encodes an innate-immunity lysozyme. Several, like sod-3, were stress-resistance genes: mtl-1 encodes a metallothionein protein that confers resistance to heavy metals. hsp-12.6 and hsp-16.2 encode small heat-shock proteins that contribute to the isotonic stress resistance of insulin/IGF-1-pathway mutants (Lamitina and Strange, 2005), as do four additional proteins, encoded by hgo-1 (homogentisate 1,2-dioxygenase), tps-1 and tps-2 (trehalose-6-phosphate synthases), and tre-4 (trehalase) (Lamitina and Strange, 2005). In addition, several genes encode metabolic enzymes, including gpd-2 (glyceraldehyde-3-phosphate dehydrogenase), nnt-1 (nicotinamide nucleotide transhydrogenase), dod-11 (sorbitol dehydrogenase), and dod-8 (17 beta-hydroxysteroid dehydrogenase). F09F7.7 and ZK384.3 encode proteins that are similar to human α-ketoglutarate-dependent dioxygenase and gastricsin, respectively. Finally, sma-10 encodes a TGF-β-pathway member.

daf-2 and daf-16 act exclusively during adulthood to influence aging (Dillin et al., 2002), so we examined expression during early adulthood. In the wild-type, 13 of the 15 genes were expressed in more than one tissue, and, with few exceptions (dod-8, tps-2, and nnt-1), their expression increased in each of those tissues under daf-2(−) conditions (Table S1). Interestingly, all but one of these upregulated genes was expressed and upregulated in the intestine. Moreover, three genes whose functions contribute to the longevity of daf-2(−) mutants (lys-7, mtl-1, and hsp-16.2) (Murphy et al., 2003; Walker and Lithgow, 2003) were expressed and upregulated only in the intestine.

Genes Downregulated in Long-lived daf-2(−) Mutants

DAF-16 also downregulates genes in daf-2(−) mutants, and RNAi knockdown of certain downregulated genes lengthens wild-type life span (Murphy et al., 2003). We analyzed: pept-1, which encodes a predicted dipeptide transporter; vit-5, which encodes a putative lipid transporter related to vertebrate vitellogenins and mammalian ApoB-100, a core LDL particle constituent; his-24, which encodes a C. elegans H1 linker histone; ZC416.6, similar to human leukotriene A-4 hydrolase; and one gene with unknown function, dod-17. Remarkably, all five genes were expressed only in the intestine (or intestine plus pharynx) (Table S1).

DAF-16 Regulates Many Genes Cell Autonomously

Next, we chose nine genes with strong daf-2(−) induction ratios for DAF-16 cell-autonomy studies. We introduced transgenic reporters for these genes into daf-16(−); daf-2(−) mutants in which daf-16 was expressed in only one tissue: intestine, neurons, or muscles. We confirmed the tissue specificity of daf-16 expression by using a functional GFP::DAF-16 protein fusion (Libina et al., 2003). Surprisingly, as with sod-3 (Libina et al., 2003); used as a control), six of these nine genes were regulated in a strictly cell-autonomous fashion (lys-7, mtl-1, hgo-1, gpd-2, nnt-1, and dod-17). The demonstration of cell autonomy was particularly striking in the intestine, because the ges-1 intestinal promoter we used did not fire evenly in all cells. We observed a close correlation between GFP::DAF-16 and reporter expression among individual intestinal cells with each of these six genes (Figure 1C, mtl-1; Figure 1D, dod-17; Figure S1B, lys-7), plus sod-3 (Figure S1A).

DAF-16 Binds Multiple DNA Sequences to Upregulate Gene Expression

Does DAF-16 bind directly to genes that it regulates cell autonomously? All but one of the 20 new DAF-16-regulated transgenes we analyzed contain at least one copy of the canonical DAF-16-binding element, consistent with this possibility. Previously, Schuster et al. (2010) demonstrated DAF-16 binding to the DBE-containing genes hsp-12.6, hgo-1, tps-1, gpd-2, and F09F7.7 in chromatin profiling experiments. We tested for DAF-16’s binding to our eleven most highly regulated genes, plus sod-3, in chromatin immunoprecipitation (ChIP) experiments. In our experiments, DAF-16 bound to sod-3, dod-8, hsp-12.6, and tps-1 preferentially in daf-2(−) mutants in both of two experiments, and to mtl-1, nnt-1, hgo-1, dod-11, and dod-17 in one of two experiments (see Figure S4D for details). We also examined the online modENCODE database for DAF-16 binding profiles for these same genes (http://intermine.modencode.org/release-30/report.do?id=64000352). Even in wild-type, in which DAF-16 is only partially activated, DAF-16 bound to genomic regions of ten of these twelve genes (sod-3, dod-8, mtl-1, gpd-2, nnt-1, hgo-1, tps-1, tps-2, dod-11, and hsp-12.6) (Figure S4D).

The one daf-2/daf-16-responsive reporter that lacked a DBE contained a small, 0.5 kb, dod-8 promoter sequence (Figures 2A and S2). This reporter was noteworthy, as, to our knowledge, DAF-16 has not been shown to bind any non-DBE sites in C. elegans. One potential DAF-16-binding site in this fragment was the so-called DAE (DAF-16-associated element, CTTATCA), as this sequence is overrepresented in promoter regions of DAF-16-regulated genes (Murphy et al., 2003). The DAE was significant in vivo, as deleting it prevented a daf-2 mutation from upregulating Pdod-8::gfp expression (Figure 2A). In gel-shift assays, DAF-16 bound to an oligonucleotide that contained six tandem wild-type copies (but not six mutant copies) of this site (Figure 2B). Thus, DAF-16 can bind the DAE.

Figure 2. DAF-16 and GATA Factors Bind to dod-8 Promoter Sequences In Vitro and Regulate dod-8 In Vivo.

(A) The 504 bp dod-8 promoter fragment drives gfp (cytoplasmic) expression in the intestine (i), hypodermis (h), body-wall muscles (m), and neurons (n) of wild-type and is upregulated in most tissues of daf-2(−) mutants. Deletion of the DAE did not have marked effects on Pdod-8::gfp expression in wild-type but significantly attenuated intestinal expression and abolished hypodermal expression in daf-2(−) mutants (right, bottom panel). Young adults, 250X magnification. Scale bar: 50 µm.

(B) Bacterially expressed GST-tagged DAF-16 gel-shifted an oligonucleotide containing three DBEs or, to a lesser extent, six DAEs. DBE and DAE point mutations abolished the binding. Shifted oligos are highlighted. Mock: purified GST. Representative autoradiograph from two or more experiments is shown for all figures.

(C) DAF-16 gel-shifted the 504 bp dod-8 promoter fragment independently of the DAE. Left block: DAF-16 binding, which could be super-shifted with a DAF-16 antibody, could be inhibited with cold competitor, a dod-8 promoter fragment that lacked the DAE. Right block: DAF-16 still bound to the dod-8 promoter fragment following DAE deletion.

(D) Knockdown of daf-16 or elt-2/GATA significantly attenuated Pdod-8::gfp induction in the intestine of daf-2(−) mutants. Young adults were photographed (100X) using a low exposure to avoid signal saturation. Scale bar: 130 µm. Lower panel: The GFP signal in the anterior quarter of the intestine (“i,” as indicated) was quantified. Bars, mean value ± SD. daf-2(+) background: RNAi control, n = 36 (animals); daf-16(RNAi), n = 36, p = 2.19E–07 (Student’s t test, versus control); elt-2(RNAi), n = 36, p = 2.32E-08. daf-2(−) background: RNAi control, n = 24; daf-16(RNAi), n = 20, p = 2.69E–17; elt-2(RNAi), n = 24, p = 3.21E–10.

(E) Both ELT-2 and ELT-3 gel-shifted DAE/GATA site-containing DNA. Binding could be competed away by a wild-type dod-8 promoter fragment, but not a mutant DAE/GATA(−)-promoter fragment. Likewise, neither ELT-2 nor ELT-3 could gel-shift the dod-8 promoter fragment lacking the DAE/GATA site. Mock: copurified proteins produced by the empty vector pET-28.

(F) Both ELT-2 and DAF-16 gel-shifted wild-type but not mutant DAE/GATA sequences. Mock: purified GST.

(G) Model for coregulation of dod-8 by DAF-16 and GATA factors. Note that DAF-16 can bind additional, unidentified promoter sites, as well.

Unexpectedly, we found that DAF-16 was able to bind to the 0.5 kb dod-8 promoter fragment lacking the DAE (but not to control plasmid DNA) (Figures 2C and S3B). Thus, we looked for additional DAF-16-binding sites. Certain suboptimal DBE-like sites with a conserved core sequence “AAACAA” have been observed in upstream sequences of several FOXO-regulated genes (Santo et al., 2006; Tran et al., 2002). Some of these are functionally significant (Tran et al., 2002). The dod-8 promoter contains different suboptimal DBEs (Figure S3A). We found that DAF-16 bound to dod-8 promoter-derived oligonucleotides that contained these motifs, but not to random oligonucleotides (Figure S3A). Removing these sites in the transgene prevented hypodermal Pdod-8::gfp expression in vivo and attenuated its induction in the intestine and muscles of daf-2(−) mutants (Figure S3C). Thus, these sites appear to play an important role in vivo.

Unexpectedly, DAF-16 still bound to the dod-8 promoter in the absence of the DAE site and all three noncanonical binding sites (Figure S3B). Thus, additional site(s) may contribute to daf-16-dependent dod-8 expression in vivo, especially since, compared with the canonical DBE, DAF-16 appeared to have much lower affinity for the DAE and other noncanonical sites (Figures 2B and S3A).

GATA Factors Bind to the DAE and Regulate the Expression of Pdod-8::gfp

Since DAF-16 bound many sites in the dod-8 promoter fragment, we wondered why the DAE was so important in vivo. The DAE is the reverse complement of the mammalian GATA factor binding site (Plumb et al., 1989). Thus, the DAE site in dod-8 might also be recognized by GATA factors. C. elegans has at least fourteen GATA-factor genes (Kormish et al., 2010; Maduro and Rothman, 2002). Using RNAi, we knocked down ten characterized GATA factors (elt-1, elt-2, elt-3, elt-5(egl-18), elt-6, elt-7, end-1, end-3, and med-1/med-2). Knocking down elt-2, which is expressed in the intestine, specifically decreased intestinal expression of Pdod-8::gfp (Figure 2D). In contrast, knocking down elt-3, which is expressed in the hypodermis but not the intestine (Gilleard et al., 1999; Tonsaker et al., 2012) (data not shown), specifically decreased hypodermal expression (Figure S4C and Table 1).

Table 1.

ELT-2 and ELT-3 Are Required for Expression of Some DAF-16-Regulated Genes

Gene Reporter	Expression under daf-2(−) Conditions	Promoter Used for Analysis				Expression upon RNAi
		Length (kb)	DBE	DAE	GATA	daf-16(−)	elt-2(−)	elt-3(−)
sod-3	Up, intestine, hypodermis, muscles, neurons	1.1	4 + 1 sub	0	4	Down	Down (intestine, modest)	No Change
dod-8(*)	Up, intestine, hypodermis, muscles, neurons	0.5	0*	1	1	Down	Down (intestine)	Down (hypodermis)
lys-7	Up, intestine	3.1	7 + 4 sub	1	10	Down	Down (intestine)	No Change
mtl-1	Up, intestine	3.6	5 + 15 sub	2	12	Down	Down (intestine)	No Change
gpd-2	Up, intestine	1.1	1	0	2	Down	Down (intestine, modest)	No Change
nnt-1	Up, intestine, hypodermis	1.9	10 + 6 sub	0	8	Down	Down (intestine)	Down (hypodermis)
hgo-1	Up, intestine, hypodermis	2.9	4 + 9 sub	1	12	Down	Down (intestine)	Down (hypodermis)
tps-1	Up, intestine, muscles	2.8	6 + 2 sub	1	10	Down	No Change	No Change
tps-2	Up, hypodermis, muscles	2.9	5 + 8 sub	1	12	Down	No Change	Down (hypodermis)
dod-11	Up, intestine, hypodermis, muscles	5.3	6 + 14 sub	2	10	Down	Down (intestine, modest)	No Change
hsp-12.6	Up, intestine, hypodermis, muscles	2.7	9 + 13 sub	0	11	Down	Down (intestine)	No Change
dod-17	Down, intestine	2.6	4 + 7 sub	0	10	Up	No Change	No Change

The table lists the 12 genes analyzed for ELT-2 and ELT-3 influence. These genes were chosen because their expression was most strongly changed by daf-2 RNAi. DBE, RTAAAYA, R = A/G, Y = C/T; DAE, CTTATCA; GATA, WGATAR, W = A/T; sub, suboptimal DBE (e.g., TAAAACAA and TTGTTTGT [Santo et al., 2006; Tran et al., 2002]). RNAi experiments were initiated at the L4 stage, 25°C, and RNAi-treated animals were analyzed as young adults. At least two independent RNAi experiments were performed, and at least ten transgenic animals were analyzed for each reporter.

^*There are two DBEs ~0.66 kb upstream of the dod-8 translational start site, not included in the 0.5 kb promoter fragment we analyzed.

Both ELT-2 and ELT-3 bound to the wild-type dod-8 promoter in vitro, but not to a DAE-mutant dod-8 promoter (Figure 2E) or DAE-mutant oligonucleotide (Figure 2F). As expected, the dod-8 promoter could bind GATA factors and DAF-16 at the same time (Figures S4A and 2G). In vivo, knocking down either daf-16 or elt-2 did not further reduce expression of the Pdod-8::gfp reporter that lacked the DAE (Figure S4B). Together, these results suggested that ELT-2 and ELT-3 recognize the DAE/GATA site in vivo, thereby promoting expression of dod-8 in daf-2(−) mutants.

GATA Factors and DAF-16 Coregulate a Subset of DAF-16 Target Genes

We tested the 12 reporters that exhibited the greatest change under daf-2(−) conditions for elt-2 and elt-3 dependency. Knockdown of the intestinal GATA-factor gene elt-2 affected 9 of these 12 reporters (sod-3, dod-8, lys-7, mtl-1, gpd-2, nnt-1, hgo-1, dod-11, and hsp-12.6), and only in the intestine (Table 1). Likewise, knockdown of the hypodermal factor elt-3 affected dod-8, nnt-1, hgo-1, and tps-2 reporters, and only in the hypodermis. Using qPCR, we found that RNAi inhibition of elt-2, but not elt-3, resulted in significant attenuation of a subset of intestine-expressed DAF-16 target genes in daf-2(−) mutants, including sod-3, lys-7, mtl-1, nnt-1, and hgo-1 (Figure 3A). Notably, all of the gene reporters that responded to GATA-factor knockdowns contained at least one DAE/GATA site in their promoters. Together, these results suggested that DAF-16 and tissue-specific GATA factors collaborate to establish tissue-specific expression of multiple downstream target genes.

Figure 3. The ELT-2 GATA Factor Regulates DAF-16 Target Genes and Extends Life Span.

(A) elt-2 knockdown affected the expression of a subset of DAF-16-regulated genes in daf-2(−) mutants. RNAi-sensitive rrf-3(−) mutants were used for RT-qPCR analysis. Bars, mean value ± SD, four biological replicates, technical triplicates. Gene expression was normalized to the gene nhr-23 (Supplemental Experimental Procedures). Student’s t test, *p < 0.05; **p < 0.01; ***p < 0.001.

(B) Knockdown of daf-16 or elt-2 shortens the life span of daf-2(−) mutants. Representative data from at least six independent RNAi experiments are shown (see Table S2). Log-rank test, ***p < 0.001.

(C) Increasing elt-2 gene dosage increases life span. Life span was increased (7%–30%) in two independent lines (see Table S2). Log-rank test, ***p < 0.001.

The strong influence that the intestinal ELT-2 GATA factor had on the expression of DAF-16 targets suggested that elt-2 might be required for the long life spans of daf-2(−) mutants. elt-2 is essential for development of the intestine (Fukushige et al., 1998; Kormish et al., 2010). We found that adult-only RNAi inhibition of elt-2, but not other GATA factors (Table S2), shortened the life span of wild-type by ~10%–20%, but consistently produced a stronger, 30%–45% shortening of life span, and reduced heat-stress resistance in daf-2(−) mutants (Figure 3B and Table S2). elt-2 inhibition also shortened the life span of calorically restricted eat-2 mutants and germline-less glp-1 mutants substantially (by 40%–50% and by 25%, respectively), but it did not preferentially shorten the long life spans of respiration mutants (Table S2). Finally, increasing elt-2 gene dosage (Fukushige et al., 1999) increased the life span of wild-type by 7%–28% (Figure 3C and Table S2). Thus, activity of the intestinal GATA-factor ELT-2 during adulthood has an important influence on the life span of C. elegans.

elt-3 played an important role in the up-regulation of four DAF-16-controlled genes in the hypodermis (skin). Consistent with this, we found that hypodermal-only daf-16 expression was able to increase the life span of a daf-16(−); daf-2(−) mutant by 16% and 32% in two experiments (Figure S7). However, when we removed elt-3 in daf-2(−) mutants, we did not observe a decrease in life span (Figures S5A–S5C). This apparent paradox suggests that DAF-16 can regulate important hypodermal life-span genes independently of ELT-3.

DAF-16 Regulates Some Target Genes Cell Nonautonomously

Three of the nine genes we analyzed for cell autonomy, the sorbitol dehydrogenase gene dod-11, the small heat-shock protein gene hsp-12.6, and the steroid dehydrogenase gene dod-8, were regulated cell nonautonomously by DAF-16.

dod-11

In daf-2(−) mutants expressing daf-16(+) only in the intestine, expression of Pdod-11::rfp was induced in the hypodermis and muscles in multiple independent lines (Figures 4A and 4B). (Again, we confirmed the tissue specificity of GFP::DAF-16 using fluorescence microscopy.) Thus, DAF-16 causes intestinal cells to make a signal that can activate dod-11 independently of daf-16 in other tissues. Within the intestine itself, GFP::DAF-16-positive cells generally expressed Pdod-11::rfp, but we observed exceptions (Figure 4A). Thus, DAF-16 might regulate dod-11 in a partially cell-nonautonomous fashion within the intestine, as well. The intestine was not the only tissue capable of affecting dod-11 expression elsewhere in the animal: animals expressing daf-16 only in muscles exhibited dod-11 induction in the muscles, intestine, and hypodermis (Figure 4A).

Figure 4. DAF-16 Regulates dod-11 and hsp-12.6 Cell Nonautonomously.

(A and B) (dod-11) In (A), top panels: Pdod-11::rfp is expressed in most tissues of wild-type and is upregulated mainly in the intestine (i), hypodermis (h,) and muscles (m) of daf-2(−) mutants in a daf-16-dependent manner. Bottom panel (left): intestinal DAF-16 upregulates Pdod-11::rfp in the intestine as well as in hypodermis and muscles (compare with daf-16[−]; daf-2[−]). Inset: overlay of intestinal GFP::DAF-16 (green) and Pdod-11::rfp (red). Note that one intestinal cell (arrowhead) does not express gfp::daf-16 but does express Pdod-11::rfp. Bottom panel (middle): muscle DAF-16 upregulates Pdod-11::rfp in muscles as well as in the intestine and hypodermis. Young adults, 250X magnification. Scale bar: 50 µm. In (B), left three panels: Intestinal cells expressing gfp::daf-16 and the resulting Pdod-11::rfp expression in intestinal (i), hypodermal (h), and muscle (m) cells are shown. Young adults, 400X magnification. Scale bar: 32 µm. Right two panels: Higher magnification of the same animal, viewed using differential interference contrast (DIC, top) or fluorescence microscopy (RFP, bottom). Pdod-11::rfp is expressed in the hypodermis (h) and muscles (m) when DAF-16 is activated in the intestine of daf-2(−) mutants. 1000X magnification. Scale bar: 13 µm.

(C and D) (hsp-12.6) In (C), left column: Phsp-12.6::rfp is expressed at low levels in wild-type and is upregulated in the same tissues (“i,” intestine; “h,” hypodermis; “m,” muscles) of daf-2(−) mutants in a daf-16-dependent manner. Right column: intestinal DAF-16 upregulates Phsp-12.6::rfp in the intestine as well as hypodermis (top). Note that not all daf-16-expressing cells (green) express Phsp-12.6::rfp (red) (inset: RFP only). Muscle DAF-16 upregulates Phsp-12.6::rfp in muscles as well as hypodermis (middle). Neuronal DAF-16 does not affect hsp-12.6 expression (bottom). Rectangle: coinjection marker Podr-1::rfp expression in head neurons. Young adults, 250X magnification. Scale bar: 50 µm. In (D), DAF-16 activity in muscles upregulates Phsp-12.6::rfp in muscle cells (m, oblong) (left column) as well as hypodermal cells (h) (right column) is shown. Young adults, 400X magnification. Scale bar: 32 µm.

hsp-12.6

Intestine-expressed daf-16 induced hsp-12.6 expression in the intestine as well as in daf-16(−) tissues, such as the hypodermis (Figure 4C). In addition, muscle-expressed daf-16 was able to induce the hsp-12.6 reporter in both muscles and hypodermis (Figures 4C and 4D).

dod-8

In two independent transgenic lines, intestinal DAF-16 activity strongly attenuated dod-8 reporter expression in the hypodermis and muscles (Figure S2). Together, these findings indicate that DAF-16 action in any of several tissues can influence gene expression independently of daf-16, either positively or negatively, elsewhere in the animal. These findings provide molecular correlates for DAF-16’s ability to affect the aging of tissues in which it is not expressed.

The Lipid-Gene Regulator MDT-15 and Longevity

How does DAF-16 influence gene expression at a distance? To address this question, we used RNAi to test ~250 DAF-16-regulated genes (twice, in two independent experiments) for their effects on dod-11 expression in daf-2(−) mutants, and in daf-2(−) mutants expressing daf-16(+) only in the intestine. One RNAi clone, for mdt-15, sharply decreased dod-11 expression in both strains (Figures 5A and 5B). mdt-15 transcriptional reporters (Taubert et al., 2006) (obtained from the Genome BC C. elegans Gene Expression Consortium), as well as RNA in situ hybridizations (The Nematode Expression Pattern Database), displayed mdt-15 expression in the intestine and some head neurons but not in the muscles or hypodermis. We observed a similar tissue distribution of mdt-15 reporter expression in daf-2(RNAi) strains. Thus, mdt-15 may help to mediate DAF-16’s action at a distance.

Figure 5. MDT-15 Is Required for dod-11 Expression and for Longevity.

(A and B) RNAi of either daf-16 or mdt-15 attenuated Pdod-11::rfp expression in the intestine (i) and hypodermis (h) of (A) daf-2(−) mutants and (B) daf-16(−); daf-2(−) mutants expressing daf-16 in the intestine. Intestine-daf-16(+): muIs199. In (B), shown are Pdod-11::rfp expression (red, top panels) overlaid with intestinal GFP::DAF-16 (green, bottom panels). Relative levels of Pdod-11::rfp expression in RNAi-treated animals are shown on the right (Kolmogorov-Smirnov test; n.s., not significant). Note that mdt-15 RNAi significantly reduced both the number and brightness of Pdod-11::rfp foci (this animal in [B] represents the “low-expression” category), despite the high level of GFP::DAF-16. Rectangle: coinjection marker Podr-1::rfp expression in head neurons. Young adults, 250X magnification. Scale bar: 50 µm.

(C) Knockdown of daf-16 or mdt-15 shortened the life span of daf-2(−) mutants to a greater extent than it affected wild-type (see Table S4). Log-rank test, ***p < 0.001.

(D) mdt-15 RNAi shortened the life span of daf-16(−); daf-2(−) mutants, but had a greater life-shortening effect on daf-16(−); daf-2(−) animals expressing daf-16 in the intestine (see Table S4). Log-rank test, **p < 0.01; ***p < 0.001.

MDT-15 is a transcriptional mediator that regulates expression of lipid and other metabolic genes (Taubert et al., 2006, 2008). mdt-15 is upregulated by DAF-16 in daf-2(−) mutants (Murphy et al., 2003) and in long-lived germline-defective animals (McCormick et al., 2012). mdt-15 RNAi shortened the life span of wild-type by ~20% (Figure 5C and Table S4), as reported (Taubert et al., 2006). However, we found that mdt-15 RNAi shortened the life span of daf-2(−) mutants by ~45%. Similarly, mdt-15 RNAi shortened the life span of daf-16(−); daf-2(−) mutants by ~10%–20%, but it shortened the life span of daf-16(−); daf-2(−) mutants expressing daf-16 in the intestine by ~30% (Figures 5D and Table S4). Thus, mdt-15 is important for wild-type longevity, and even more important for the extended life spans of daf-2(−) mutants. At the tissue level, mdt-15 RNAi reduced the ability of intestinal daf-16(+) to delay muscle deterioration (Figure 6A), while having no significant effects on the sarcomeres of daf-16(−); daf-2(−) mutants. However, mdt-15 may affect additional processes required for movement, as mdt-15 RNAi decreased the motility of both daf-16(−) and daf-16(+) animals (Figure S6C).

DAF-16 Can Act at a Distance to Protect Animals from Amyloid Paralysis

When expressed in C. elegans’ muscles, the human Alzheimer’s protein Aβ(1–42) aggregates and paralyzes the animals during early adulthood (Link, 1995). This paralysis is attenuated by insulin/IGF-1-pathway mutations (Cohen et al., 2006; Florez-McClure et al., 2007). We wondered whether DAF-16 could also act at a distance to counteract Aβ toxicity. In multiple independent lines, we found that Aβ-containing daf-16(−); daf-2(−) adults expressing intestinal daf-16(+) moved much better than did Aβ-containing wild-type animals or daf-16(−); daf-2(−) mutants (Figures 6B, 6C, and S6E and Table S3). (As a control, we introduced the Aβ transgene back into wild-type and found that it still induced paralysis [Table S3].) The ability of DAF-16 to counteract Aβ toxicity correlated with its expression levels in the intestine (Figure S6D & S6E). Thus, intestinally-expressed DAF-16 can counteract Aβ-dependent muscle dysfunction. Likewise, mdt-15 RNAi, similar to daf-16 RNAi, also abolished the ability of intestinal DAF-16 to delay Aβ-dependent paralysis (Figure 6D and Table S3).

DISCUSSION

Understanding how the insulin/IGF-1 endocrine system coordinates the rate of aging among different tissues is fundamentally important, as this pathway appears to influence the rate of aging throughout the animal kingdom, from worms to man. In this study, we analyzed the expression of a diverse collection of DAF-16-regulated genes in vivo to better understand how components of this signaling network map across the tissues of the animal. One could imagine two extreme cases: in one, DAF-16 would act at the end of an insulin/IGF-1 signal-transduction pathway, regulating downstream genes that affect only the health and longevity of the cells in which they are expressed. At the other extreme, since DAF-16 activity within a single tissue can delay the aging of other tissues and increase the life span of the whole animal, DAF-16 could regulate only downstream signaling genes whose products then activate daf-16-independent life-extension pathways in other cells. We find that both mechanisms operate.

Limitations of Our Gene Set

Microarray analysis (Halaschek-Wiener et al., 2005; Lee et al., 2009; McElwee et al., 2003; Murphy et al., 2003) and direct DNA binding assays (Oh et al., 2006; Schuster et al., 2010) can identify C. elegans genes that are likely to be regulated directly versus indirectly by a transcription factor. Our in vivo imaging analysis of DAF-16-regulated genes complements these approaches and allows us to investigate the tissue specificity and cell autonomy of gene expression. The genes we analyzed have diverse functions, affecting protein homeostasis, innate immunity, and metabolism, and many have been shown to contribute to the long life spans of daf-2(−) mutants.

However, our analysis was biased against certain types of genes. First, neural genes: Neurons are relatively resistant to RNAi. RNAi was used in part of the microarray analysis from which we selected genes to study, and in our initial assessment of the daf-2 dependence of transgene expression. However, previously, we showed that nonneuronal daf-2 RNAi doubles life span, and that nonneuronal daf-16 RNAi completely suppresses the life-span extension of daf-2(−) mutants (Libina et al., 2003). Moreover, expressing daf-16 exclusively in neurons in a daf-16(−); daf-2(−) background produces only a small increase in life span (Libina et al., 2003). Therefore, understanding the nonneuronal activities of DAF-16 is relevant for understanding life-span regulation.

Second, because our reporters were driven by upstream DNA sequences, we did not query potential regulatory sequences within introns or coding sequences of DAF-16-regulated genes. This is a concern, as the one translational fusion we did examine, from mtl-1, was highly daf-2 responsive, whereas a transcriptional fusion to the same promoter sequence did not respond as well. In addition, we note that DAF-16 could potentially influence gene expression at the level of translation (McColl et al., 2010), and this would not be assessed in our study.

Finally, important daf-2/daf-16-regulated genes with small induction ratios would probably have escaped detection in our assay—and we could have missed, simply by chance, key lifespan genes with properties that differ from the genes we examined. Nevertheless, our analysis of this small gene set suggests some interesting new features of this regulatory network.

Three Modes of DAF-16 Gene Regulation

Together, our findings provide molecular support for the idea that DAF-16 influences gene expression among the tissues of C. elegans in three ways (Figure 6E). First, DAF-16 can act within a tissue to regulate genes predicted to influence the health and longevity of that tissue. Second, DAF-16 activates downstream signal transduction cascades that act independently of daf-16 to regulate gene expression at a distance and to slow the aging of other tissues (Libina et al., 2003) (this study) (FOXO-to-FOXO(−) signaling). Third, as shown previously, DAF-16 regulates the expression of insulin-like genes (Murphy et al., 2003), allowing DAF-16 activity in one tissue to affect DAF-16 activity elsewhere in the animal (FOXO-to-FOXO signaling) (Murphy et al., 2007).

Cell-Autonomous Gene Regulation by DAF-16

Six of the nine daf-2/daf-16-responsive genes we examined were regulated in a strictly cell-autonomous fashion by DAF-16, as was the previously analyzed DAF-16-regulated gene sod-3. This finding argues against the model that DAF-16 directly regulates only downstream signaling genes, whose effects on other cells are completely responsible for life extension. Instead, the behavior of this gene sample, which includes genes with diverse functions, suggests that DAF-16 may activate numerous types of cell-protective genes cell autonomously. All but one promoter we analyzed contained canonical DAF-16-binding elements. The presence of DBEs suggests direct DAF-16 regulation, though our findings show that DBE-containing promoters, such as the dod-11 and hsp-12.6 promoters, can also be switched on by DAF-16 indirectly, through DAF-16’s activity in other tissues. Conversely, DAF-16 could also act via a promoter fragment that lacked canonical DBEs, apparently by binding to suboptimal DBE sites, to the DAE/GATA site, and to other, unidentified sequences. FOXO proteins possess chromatin-remodeling ability (Hatta and Cirillo, 2007), so DAF-16 may utilize multiple binding sites to create a permissive environment for gene expression.

All but one of the twenty daf-2-dependent transgenes we examined (plus the control sod-3 transgene) were expressed in the intestine, and eleven were expressed mainly or exclusively in the digestive tract. This intestinal enrichment was highly significant statistically (p = 1.1E-11; Table S1). Why might DAF-16 regulate so many intestinal genes? First, the intestine seems to be particularly vulnerable to aging, undergoing extensive tearing and deterioration (McGee et al., 2011). It is also a major entry port for toxins and bacterial pathogens, and becomes packed with bacteria with age (Garigan et al., 2002; McGee et al., 2011). Bacterial packing and intestinal deterioration are both reduced greatly by daf-2 mutations (Garigan et al., 2002; McGee et al., 2011). Thus, DAF-16 may extend life and promote stress resistance in part by “bullet-proofing” the intestine. For instance, the intestinal metallothionein mtl-1 may protect the animal from heavy metals it ingests. Finally, our 23 DAF-16-regulated, intestine-expressed genes include at least 15 metabolic genes (Table S1), raising the possibility that DAF-16 could act in the intestine to help nourish the animals to improve systemic health.

Insulin/IGF-1 signaling mutants are resistant to pathogenic bacteria (Evans et al., 2008; Garsin et al., 2003), possibly due to increased intestinal expression of the innate-immunity lysozyme gene lys-7. The intestine-specific GATA-factor ELT-2 may act with DAF-16 to protect the intestine from infection, as it is required for survival of wild-type animals exposed to pathogens, and for intestinal expression of lys-7 (Figure 3A and Table 1) and other lysozymes (Shapira et al., 2006).

Finally, we note that, because intestinal DNA makes up only a small fraction of the animal’s DNA, the intestinal enrichment of DAF-16-regulated genes could help explain why so few DAF-16-regulated genes from microarrays have been identified in chromatin profiling experiments (see Figure S4D, legend).

GATA Factors Direct Expression of DAF-16-Regulated Genes to Specific Tissues

DAF-16 is expressed widely, and we found that DAF-16 activates genes tissue specifically, at least in part, by functioning in combination with tissue-specific GATA factors. The intestinal GATA-factor ELT-2 and the hypodermal factor ELT-3 were required for intestinal and hypodermal expression, respectively, of many DAF-16-regulated genes. Our findings, both in vivo and in vitro, suggest that GATA factors regulate these genes by binding to their DAE/GATA site (as predicted previously by McGhee et al., 2009, from DNA sequence). We note that some DAE-containing DAF-16-regulated genes, such as lys-7 and mtl-1, are expressed exclusively in the intestine (Tables 1 and S1). Presumably, these genes contain additional sequences that prevent ELT-3 from activating them in the hypodermis. In addition, some DAF-16-regulated genes, like tps-1, were upregulated normally in both the intestine and the hypodermis in spite of GATA factor RNAi. Thus, DAF-16 does not absolutely require GATA factors to regulate gene expression in these tissues.

GATA Factors and Life Span

Since the intestinal GATA factor ELT-2 is needed for much intestinal DAF-16-regulated gene expression, it seemed likely that elt-2 knockdown would shorten the life span of daf-2(−) mutants. We found that this was the case. Our data are consistent with a very recent, independent report showing that elt-2 inactivation could disrupt cytoprotective gene expression and shorten the life span of daf-2(−) mutants (Shore et al., 2012). We found that elt-2 knockdown also shortened the life span of calorically restricted eat-2(−) mutants substantially. This life-span pathway is daf-16 independent but requires the FOXA transcription factor, pha-4 (Panowski et al., 2007). Interestingly, pha-4 expression is activated by ELT-2, and the two proteins have been shown to coregulate specific genes (Anokye-Danso et al., 2008).

Our findings also point to an important role for a new tissue, the hypodermis, in life-span regulation by DAF-16, as hypodermal-only daf-16 expression could extend the life span of daf-16(−); daf-2(−) mutants up to ~30%. Because ELT-3 was required for hypodermal expression of four DAF-16-regulated genes we examined, one might expect that knocking down ELT-3 expression would shorten the life span of daf-2(−) mutants. However, as observed by the McGhee group (Tonsaker et al., 2012), this was not the case. (Our findings differ from those from the Kim lab [Budovskaya et al., 2008] in some respects; please see Figure S5 for discussion.) This finding implies that DAF-16 regulates important hypodermal longevity genes independently of ELT-3.

DAF-16 Action at a Distance: FOXO-to-FOXO(−) Signaling

In this study, we extended the case for FOXO-to-FOXO(−) signaling from the organismal level to the level of individual tissues (aging muscles) and genes (specifically, two metabolic genes dod-11 and dod-8, and one chaperone gene, hsp-12.6). These findings put the concept of FOXO-to-FOXO(−) signaling on solid molecular footing. DAF-16 action in the intestine, which can extend life span substantially (by 50%–70%), affected dod-11, hsp-12.6, and dod-8 expression in multiple tissues. In addition, DAF-16 could act in other tissues to affect gene expression in the intestine and elsewhere. This latter finding can help to explain how, given the fragility of the intestine, daf-2(−) mutants can live 50% longer than wild-type if daf-16 is expressed only in nonintestinal tissues (Libina et al., 2003). In that case, perhaps DAF-16 can act at a distance to protect the intestine.

How can DAF-16 promote signaling across tissues? Others reported that the gene scl-1 was a candidate downstream signaling gene, but we were unable to confirm this in our studies (see Supplemental Discussion). However, we identified a new candidate, the DAF-16-regulated gene mdt-15. MDT-15 is a transcriptional mediator subunit that regulates genes involved in lipid metabolism, so it could potentially induce lipid signals that act across the tissues to affect life span. Loss of mdt-15 reduces dod-11 expression in many tissues, including several that do not appear to express mdt-15. Thus, MDT-15 appears to act on the sending end of an intercellular signaling pathway that is activated by DAF-16.

mdt-15(RNAi) animals are unhealthy. However, two findings suggest that mdt-15 plays an important role in aging. First, loss of mdt-15 accelerated age-dependent sarcomere deterioration in daf-2(−) animals expressing intestine-only daf-16(+), but not in daf-2(−) animals that were also daf-16(−) (Figure 6). Second, mdt-15 inhibition had a greater life-shortening effect on daf-2(−) mutants than it had on wild-type (Figure 5C). Rogers et al. (2011) recently reported similar, independent findings and also showed that mdt-15 was required for life-span extension by inhibiting the translation factor ifg-1/eIF4G. It will be interesting to learn more about this potential downstream signaling pathway in the future.

FOXO-to-FOXO Signaling

DAF-16 can act at a distance to upregulate DAF-16 activity elsewhere in the animal (Murphy et al., 2007). There are many situations in which inhibiting insulin or IGF-1 responsiveness in certain tissues extends life span. However, in the great majority of these cases, whether FOXO is required in other, wild-type tissues is not known. These examples include (1) the extension of life span caused by loss of daf-2 activity in the ectoderm (skin, neurons) of C. elegans (Apfeld and Kenyon, 1998), as well as (2) the suppression of longevity caused by neuron-only daf-2 expression (or intestine or neuron-only age-1/PI3K expression) in daf-2 (or age-1) mutant worms (Iser et al., 2007; Wolkow et al., 2000). Likewise, it is not known whether the ability of neuronal age-1(+) to influence intestinal hsp gene expression in C. elegans (Iser et al., 2011) requires intestinal FOXO activity. In mice, brain-specific loss of the IGF-1 receptor (Kappeler et al., 2008) or downstream IRS genes (Taguchi et al., 2007) can increase life span. Activating FOXO or inhibiting insulin signaling in adipose tissue can extend life span in flies (Giannakou et al., 2004; Hwangbo et al., 2004) and mice (Blüher et al., 2003). At least in flies, this condition appears to trigger FOXO-to-FOXO signaling (Hwangbo et al., 2004), but whether it might trigger FOXO-to-FOXO(−) signaling as well is not known. It would be interesting to carry out these experiments in a foxo(−) background to determine whether FOXO is required in responding tissues. Finally, given the importance of the C. elegans intestine, as well as the hypodermis, in life-span regulation, it would be interesting to ask whether insulin/IGF-1-pathway members could send life-extending signals from the intestines or the skin of higher organisms.

DAF-16 Can Suppress Symptoms of Age-Related Disease from a Distance

Long-lived insulin/IGF-1 mutants are resistant to many age-related diseases, so we asked whether DAF-16 might act at a distance in a disease setting. We found that, in a process that requires mdt-15, endodermal DAF-16 can partially suppress the paralysis caused by expressing Aβ in the muscles. DAF-16 and MDT-15 action could potentially reduce the expression of Aβ in muscles, or they could decrease the accumulation of toxic Aβ oligomer species. This finding is important, as it raises the possibility that systemic, FOXO-dependent signals, perhaps beneficial lipid signals, might be able to slow the progression of Alzheimer’s disease, and possibly other age-related diseases, in humans.

EXPERIMENTAL PROCEDURES

Microscopy

Transgenic animals were analyzed at the young-adult stage. For quantitative analysis, GFP fluorescence in the anterior quarter of the intestine was measured using the OpenLab software.

Gel-Shift Assays

Proteins were expressed and purified from bacteria. Radioactively labeled DNA oligos were mixed with proteins, resolved on a polyacrylamide gel, and subjected to autoradiography.

Life-Span Analysis

Lifespan assays were performed as described (Apfeld and Kenyon, 1998). RNAi was initiated at the young-adult stage. The prefertile period of adulthood was used as t = 0 for life-span analysis. STATA software (version 10.1) was used for statistical analysis.

Paralysis

Worms expressing human Aβ(1–42) in body-wall muscles were raised at 20°C and analyzed as adults. Worms that failed to move when touched with a platinum wire were scored as “paralyzed” (Link, 1995). To avoid mis-scoring, paralysis assays were terminated by day 8 of adulthood when wild-type animals begin to move more slowly. RNAi-treatment was initiated at the L4 stage, and young adults were scored at room temperature (~22.5°C), which accelerated paralysis and helped to distinguish Aβ paralysis from aging effects on motility.