Keywords

1 Introduction

Cellular quality control networks are comprised of molecular chaperones, degradation machineries, and cytoprotective stress responses, such as the heat shock response (HSR) and the unfolded protein response (UPR) (Akerfelt et al. 2010; Haynes et al. 2013; Hetz 2012). The HSR enables the cell to adjust the expression of chaperones and other cytoprotective genes under proteotoxic conditions, thereby ensuring stress survival, recovery, and adaptation (Akerfelt et al. 2010). At the molecular level, this involves transcriptional regulation of heat shock (HS) genes in a manner proportional to the intensity, duration, and type of stress (Abravaya et al. 1991a, 2010; Gasch et al. 2000). HSF1, as the “master regulator” of the HSR, functions both as a stress sensor and transcriptional regulator of downstream stress response genes that are essential for protecting and remodeling the cell (Akerfelt et al. 2010).

The HSR has been considered a universal molecular response to various stress stimuli. Nonetheless, there are several examples in which the HSR is poorly or incompletely activated, such as the limited response seen in early development, as well as in different tissues of aged animals and in late-onset neurodegenerative diseases, such as Huntington’s disease and Alzheimer’s disease (Ben-Zvi et al. 2009; Bienz 1984; Heydari et al. 1993; Labbadia et al. 2011; Prahlad et al. 2008; Sprang and Brown 1987; Taylor and Dillin 2011). These observations suggest that the HSR can be spatially and temporally regulated during the lifetime of a multicellular organism.

In the past decade, lifespan-enhancing pathways were shown to be potent modifiers of proteostasis and to suppress protein aggregation and toxicity. Moreover, HSF1 has been shown to be an important factor influencing lifespan (Kenyon 2010b; Labbadia and Morimoto 2014; Tissenbaum 2012). Recent studies mainly conducted in Caenorhabditis elegans have uncovered several cell-nonautonomous pathways that modulate cellular quality control systems during the lifespan of the organism, in particular, HSF1. These pathways include cell-nonautonomous regulation of the heat shock response by neurons and germline stem cells, as well as transcellular signaling of proteostatic deficiencies in which the expression of misfolded proteins in one tissue can induce a systemic HSR (Prahlad and Morimoto 2009; Shai et al. 2014; van Oosten-Hawle and Morimoto 2014). Several of these pathways are also linked to lifespan, suggesting that proteostasis networks may show differential sensitivity at different stages over the lifespan of an organism. Here, we review changes in proteostatic function with age, focusing on cell-nonautonomous regulation of HSF1 during aging (mainly in C. elegans) to offer insight into how multicellular organisms adjust stress responses during key life stages and how HSF1 function, in turn, can impact lifespan.

2 The Loss of Proteostasis Is a Hallmark of Aging

Aging is often associated with the time-dependent loss of cellular function, increased susceptibility to environmental and physiological insults, and increased vulnerability to disease (Gems 2015; Partridge 2014; Riera and Dillin 2015). One of the hallmarks of aging is age-dependent loss of proteostasis. In agreement, aggregation and toxicity of aggregation-prone proteins associated with disease, such as huntingtin, is enhanced with age, while different protein quality control machineries show decreased function (Labbadia and Morimoto 2015a; Shai et al. 2014; Taylor and Dillin 2011). Proteostasis loss and age-dependent aggregation can be recapitulated in different model systems in a time frame related to their aging. In fact, a subnetwork of chaperones was shown to be required in both human brain aging, as well as in brain samples, and C. elegans models of age-associated diseases (Brehme et al. 2014). Thus, quality control systems and, specifically, the decline in proteostasis capacity are suggested to play important roles in aging and age-associated diseases (Brehme et al. 2014; Labbadia and Morimoto 2015a).

2.1 When Does Proteostasis Collapse?

The failure of proteostasis networks to maintain the proteome of aged animals is commonly explained by the slow accumulation of damage over time. Protein damage and misfolding in aged individuals is, therefore, suggested to result from a constant difference between the error rate and the efficiency of cellular quality control network in repairing or removing misfolded proteins over the course of an organism’s life, leading to a gradual accumulation of damaged proteins over time. An alternative explanation for the failure of proteostatic networks in old individuals is that the function of cellular proteostatic networks decline with age (Taylor and Dillin 2011). For example, translation fidelity may decline with age, resulting in an increased load of damaged proteins as the individual ages (Kirstein-Miles et al. 2013). Likewise, the expression and function of cellular quality control networks may be differentially regulated over the lifespan of the organism (Bar-Lavan et al. 2012). Such changes can remodel cellular folding capacity and stress tolerance, thus increasing the risk for age-associated pathology. What distinguishes between these mechanisms is the rate of damage accumulation during adulthood. While the first scenario suggests a passive accumulation of damage over time that can be modulated by the efficacy of the proteostatic network, the second suggests that quality control function is differentially regulated over the lifespan of the organism.

2.2 Age-Dependent Changes in HSR Activation

Recent studies mostly conducted with C. elegans demonstrated that cellular quality control networks are modified as animals undergo transition to a reproductively mature state. Specifically, rapid changes in the regulation and activation of stress responses were identified. When thermoresistance and induction of heat shock genes were monitored over time in adult C. elegans, both declined sharply 12 h following transition to adulthood (Shemesh et al. 2013). Thus, within hours of reaching adulthood, most animals lost the ability to mount an effective HSR and survive the insult. This finding suggests the existence of very strong negative regulation of the HSR upon transition to adulthood, inhibiting HSF1 function. When other stress responses were examined, such as activation of UPR in the endoplasmic reticulum (UPRER), UPR in the mitochondria (UPRmt), and oxidative stress in response to different perturbations, an early decline in the mounting of a stress response was also observed (Ben-Zvi et al. 2009; Labbadia and Morimoto 2015b; Taylor and Dillin 2013). Likewise, survival from acute oxidative stress and activation of an NRF2-dependent oxidative stress response in adult Drosophila melanogaster was strongly dampened, as compared to young animals (Rahman et al. 2013). Altered temporal regulation was also apparent for C. elegans c-Jun N-terminal kinase (JNK) signaling. While the JNK homolog KGB-1 enhances DAF-16 nuclear localization and transcriptional regulation during C. elegans development, this function was reversed upon transition to adulthood (Twumasi-Boateng et al. 2012). Changes in the activation of JNK signaling were also observed in adult Drosophila and old mice, although expression modulation was not monitored early in adulthood in either case (Girardot et al. 2006; Hsieh et al. 2003; Simonsen et al. 2008; Tsakiri et al. 2013). These data suggest that stress-induced transcriptional activation is strongly dampened early in adulthood and exhibits switch-like behavior associated with reproduction onset (Fig. 5.1).

Fig. 5.1
figure 1

Proteostasis is remodeled upon transition to adulthood. Changes in stress response activation, expression and function of chaperones, the ubiquitin-proteasome system, and autophagy modulate cellular proteostasis capacity, resulting in accumulation of misfolded and aggregated proteins during adulthood. These changes coincide with the onset of oocyte biomass production and can be modulated by GSC arrest

Full size image

It is important to note that stress response activation can also be remodeled at other stages over the lifespan of an organism. For example, activation of UPRmt can be transmitted to other tissues, albeit only during development (Durieux et al. 2011). Likewise, heat shock response can be modulated during C. elegans development by including 5-fluoro-2′-deoxyuridine in the growth medium (Feldman et al. 2014). These observations strongly support the hypothesis that quality control networks can be remodeled during development and with age and suggest that many different pathways modulate cellular responses to stress.

2.3 Age-Dependent Changes in Proteostatic Network Composition

Changes in proteostatic function over time are associated with changes in expression of genes encoding quality control machinery components, such as ribosomal proteins, chaperones, and proteasome- and autophagy-associated proteins (Taylor and Dillin 2011). Relying on a quantitative mass spectrometry proteomics-based approach designed to follow changes in protein expression during C. elegans adulthood, it was uncovered that ribosomal abundance strongly declined with age and that ribosome stoichiometry was disrupted, while the abundance of specific small HS proteins (sHSP), proteasome subunits, and some E3 ligases increased with age. These changes were even observed early in adulthood (Walther et al. 2015). In agreement, C. elegans epidermal growth factor (EGF)-mediated signaling was shown to upregulate the expression of genes associated with the ubiquitin-proteasome system and downregulate the expression of some chaperones upon transition to reproductive adulthood (Liu et al. 2011). Additional studies observed changes in the expression of genes encoding proteostasis-mediating components during aging. For example, in adult Drosophila, changes in the activation of JNK signaling and in the levels of proteasome subunits were observed (Girardot et al. 2006; Simonsen et al. 2008; Tsakiri et al. 2013). Likewise, expression levels of autophagy-associated genes declined with age in adult Drosophila and rats (Kiffin et al. 2007; Simonsen et al. 2008). Thus, the composition of the proteostatic network is also remodeled in early adulthood (Fig. 5.1).

2.4 Age-Dependent Changes in Protein Folding and Aggregation

If proteostatic networks are remodeled during adulthood, then one would expect proteostatic capacity to decline within a specific window of time during adulthood. Changes in the activity of quality control networks would impact their interactions with the proteome and, therefore, change rates of protein misfolding. One such example is the onset of aggregation and toxicity in a C. elegans model of Huntington’s disease. In this model, animals expressing 35 repeats of poly-glutamine (polyQ) start to accumulate aggregates following the onset of reproduction (Morley et al. 2002). For animals cultivated at 20 °C, this occurs ~4 days postembryo, while for animals cultivated at 25 °C, this transpires 2.5 days postembryo (Morley et al. 2002; Shemesh et al. 2013). Thus, aggregation-prone proteins quickly became insoluble upon transition to adulthood.

The functions and folding of metastable temperature-sensitive (ts) mutant proteins are also disrupted early in adulthood. Temperature-sensitive mutant proteins assayed for changes in their folding capacity, such as ras, the acetylcholine receptor, and perlecan (UNC-52), all showed a rapid loss of function early in adulthood (days 2–6 of adulthood) (Ben-Zvi et al. 2009; Shemesh et al. 2013). In C. elegans muscle, expression of a ts version of myosin gave rise to age-dependent mislocalization of myosin and resulted in uncoordinated movement. Here, myosin misfolding was detected as early as day 3 of adulthood (Ben-Zvi et al. 2009). Likewise, neurons and coelomocytes of animals expressing a ts version of DYN-1 showed age-dependent dysfunction and DYN-1 mislocalization as early as day 2 of adulthood (Ben-Zvi et al. 2009). In such animals, the misfolding of metastable proteins observed early in adulthood was thus general but affected different proteins in different tissues.

Age-dependent protein misfolding associated with functional decline was also apparent for wild-type proteins, although their aggregation started, for the most part, later in adulthood (Ben-Zvi et al. 2009; Haithcock et al. 2005; Shemesh et al. 2013). When wild-type myosin was monitored over time, it showed age-dependent mislocalization and misfolding, similar to what was seen with metastable myosin, albeit later in adulthood than the nonnative protein (Ben-Zvi et al. 2009; Shemesh et al. 2013). An unbiased examination of age-dependent aggregation during C. elegans adulthood supported this finding. Using mass spectrometry to systematically examine age-dependent changes in protein solubility, David et al. identified several hundred proteins that became insoluble over time. Some wild-type proteins aggregated as early as day 3 of adulthood (David et al. 2010). Many of these proteins were also identified in another independent study (Reis-Rodrigues et al. 2012). A recent study using quantitative mass spectrometry to analyze changes in the proteome during C. elegans adulthood showed that many proteins underwent more than twofold changes in abundance, linked to widespread aggregation (Walther et al. 2015). Thus, upon transition to reproductive adulthood, there is an onset of age-dependent decline in cellular proteostasis, coinciding with a decline in HSR activation (Fig. 5.1).

The rate of protein clearance is also modulated early in C. elegans adulthood. Monitoring the clearance of GFP tagged with an uncleavable ubiquitin tag as a degradation reporter (UbG76V-GFP) demonstrated that there is a sharp increase in protein degradation upon transition to adulthood. While UbG76V-GFP was completely degraded during development, it started to accumulate when animals became reproductive adults (Liu et al. 2011). At later stages, protein degradation was shown to decrease in worms, flies, and rats (Hamer et al. 2010; Keller et al. 2000; Liu et al. 2011; Tonoki et al. 2009). Changes in autophagy early in adulthood were not monitored, although reduced autophagic activity was observed in adult Drosophila and rats (Kiffin et al. 2007; Simonsen et al. 2008). Given that the two systems are linked for the removal of aggregated proteins (Cha-Molstad et al. 2015), these observations raises interesting questions about how such interactions are modulated with age to change protein clearance.

2.5 Modulating HSF1 Levels Can Affect Proteostatic Collapse

If the levels and functions of proteostasis-related genes are limiting in adulthood, could modifying the levels of quality control transcription factors, such as HSF1, modulate proteostasis? When HSF1 levels were diminished using RNAi, the loss of function of metastable proteins, such as paramyosin and dynamin, and of aggregation-prone proteins, such as polyQ and Aβ, was exacerbated, resulting in an earlier decline in proteostasis. In contrast, overexpression of HSF1 proved to be protective for metastable and aggregation-prone proteins (Ben-Zvi et al. 2009; Cohen et al. 2006; Hsu et al. 2003; Morley and Morimoto 2004). Modulating HSF1 levels also affected the aggregation of wild-type proteins (Walther et al. 2015). The switch-like behavior of the HSR and the fact that over-expression of HSF1 can rescue proteostasis in adulthood (Ben-Zvi et al. 2009; Labbadia and Morimoto 2015b; Shemesh et al. 2013) suggest that HSF1 and other stress transcription factors are likely differentially regulated over the lifespan of the organism.

3 HSF1 Is a Lifespan Regulator

The involvement of HSF1 in setting the function of proteostasis in adulthood suggests that HSF1 can impact aging and lifespan. In agreement, HSF1 was shown to modulate age-associated phenotypes and shorten lifespan in C. elegans (Garigan et al. 2002). HSF1 knockdown induced tissue integrity decline of both somatic and germline stem cells, and wild-type animals treated with HSF1 RNAi were short-lived (Garigan et al. 2002). Conversely, the lifespan of animals overexpressing HSF1 was extended by 22–40 % (Hsu et al. 2003; Morley and Morimoto 2004). Thus, HSF1 is a lifespan regulator and it can promote longevity.

3.1 HSF1 Is Required for Insulin/IGF-1 Signaling (IIS) Pathway Activity

Downregulation of the insulin/IGF-1 signaling (IIS) pathway promotes longevity and suppresses protein aggregation and toxicity. Occupation of the IIS receptor (DAF-2 in C. elegans) initiates a conserved cascade of kinases that results in the phosphorylation and subsequent inactivation of the FOXO transcription factor (DAF-16 in C. elegans). In different model organisms, activation of the FOXO transcription factor, as occurs upon reduced IIS receptor function or levels, resulted in an increase in lifespan (Kenyon 2010b). In C. elegans, DAF-16 is required for DAF-2-mediated extended lifespan and improved proteostasis (Cohen et al. 2006, 2009; Hsu et al. 2003; Morley et al. 2002), as is HSF1. RNAi knockdown of HSF1 in a DAF-2 mutant background shortened the lifespan of the animal and resulted in a loss of the expected proteostatic improvement (Ben-Zvi et al. 2009; Cohen et al. 2006, 2009; Hsu et al. 2003; Morley et al. 2002).

3.2 HSF1 Is Required for Germline Stem Cell (GSC) Signaling

Endocrine signaling from germline stem cells (GSCs) and the somatic gonad links reproduction to aging. Removal of GSCs by laser ablation or induction of GSC arrest through mutations extended the lifespan of C. elegans (Arantes-Oliveira et al. 2002; Hsin and Kenyon 1999) and other model animals, such as the nematode Pristionchus pacificus (Hsin and Kenyon 1999; Rae et al. 2012) and Drosophila (Flatt et al. 2008). Transplantation of young ovaries into old mice also promoted longevity (Cargill et al. 2003; Mason et al. 2009). These studies established a role for the reproductive system, and specifically signals from the somatic gonad and GSCs, in lifespan (Kenyon 2010a). Because removal of the entire gonad did not extend lifespan, it would appear that longevity is not a simple consequence of sterility but rather depends on communication between the reproductive system and somatic tissues that modulate different signaling pathways, including those involved in metabolism and quality control, to extend lifespan (Antebi 2012). GSC-dependent extension of lifespan requires the activation of many different downstream signaling pathways, one of which includes HSF1 (Antebi 2012; Libina et al. 2003).

3.3 HSF1 Is Required in Part for Dietary Restriction (DR)

Dietary restriction (DR) is another manipulation that extends life in organisms spanning from yeast to mammals. In fact, DR was the first manipulation described that increased lifespan (McCay et al. 1989). There are many different protocols to induce DR, including limiting food intake (Greer and Brunet 2009). For example, mutant C. elegans with a defective eat-2 pharyngeal pump-encoding gene showed reduced food intake and thus serve as a genetic model of DR (Avery 1993). While different DR regimes all extend lifespan, they activate different nutrient-sensing pathways, including those involving TOR, AMP kinase, and sirtuins (Kenyon 2010b). Only some of these pathways, however, require HSF1 for longevity and reduction of proteotoxicity (Greer and Brunet 2009; Mouchiroud et al. 2011; Sutphin and Kaeberlein 2008; Vora et al. 2013; Zhang et al. 2009). Thus, HSF1 is required for lifespan extension in several different models of longevity that are linked but not always coupled to improved proteostasis.

3.4 HSF1 Is Differentially Regulated in Development and in Adulthood

Similar to the observed changes in proteostatic composition and function in adulthood, HSF1 regulation by the IIS pathway is also modulated upon transition to adulthood. Using conditional RNAi, Cohen et al. showed that HSF1 function in IIS-dependent reduction of aggregates toxicity was mostly required during larval development. Knocking-down HSF1 levels against a background of reduced IIS signaling (as achieved upon knockdown of the daf-2 insulin-like receptor-encoding gene) on the first day of adulthood did not reduce IIS-mediated protective effects. In contrast, downregulation of DAF-16 was essential only in adulthood (Cohen et al. 2010). Moreover, HSF1 knockdown during development also abolished IIS-mediated extension of lifespan, while HSF1 downregulation during early or late adulthood had only a mild or no effect on lifespan, respectively (Volovik et al. 2012). Thus, HSF1 function in proteostasis and lifespan when part of the IIS pathway shows switch-like behavior upon transition to adulthood (Fig. 5.1).

4 Cell-Autonomous and Cell-Nonautonomous Regulation of HSF1

HSF1 regulation has been studied in depth in both in vitro and in single-cell systems (see Chaps. 1, 2, 3 and 4). In short, upon accumulation of misfolded and damaged protein, the equilibrium of molecular chaperones is altered such that HSF1 monomers are released from repressive complexes that include molecular chaperones (Abravaya et al. 1992; Shi et al. 1998; Zou et al. 1998). Once free, HSF1 acquires DNA-binding activity through trimerization in a process that is also regulated by extensive posttranslational modifications that include phosphorylation (Guettouche et al. 2005; Holmberg et al. 2001; Kline and Morimoto 1997; Knauf et al. 1996; Sorger and Pelham 1988), sumoylation (Anckar et al. 2006; Hietakangas et al. 2003), and acetylation (Westerheide et al. 2009) that can enhance or suppress HSF1 DNA-binding activity (Akerfelt et al. 2010). In the nucleus, HSF1 binds to promoters of target loci (heat shock elements (HSEs)) found in promoter regions of stress-inducible genes, which are maintained in a state that is permissive for HSF1 binding and transcription regulation (Fujimoto et al. 2012; Guertin and Lis 2013; Shopland et al. 1995; Zelin et al. 2012).

In nonstressed metazoan cells, HSF1 is predominantly found in a monomeric non-DNA-binding state, repressed by transient interactions with chaperones including Hsp90, Hsp70, and Hsp40 (Abravaya et al. 1992; Shi et al. 1998; Zou et al. 1998). When the stress signal is dampened, or when damaged proteins are no longer expressed, the HSR is attenuated. Attenuation is associated with the release of HSF1 trimers from DNA and subsequent conversion of the trimers to HSF1 monomers (Anckar and Sistonen 2007; Morimoto and Santoro 1998; Wu 1995). HSR attenuates, in part, due to the accumulation of chaperones that bind to HSF1 and acetylation of HSF1 in the DNA-binding domain (Abravaya et al. 1991b, 1992; Shi et al. 1998; Yao et al. 2006). The combination of these posttranslational modifications and chaperone interactions thus subjects HSF1 to multiple levels of control and feedback loops so as to precisely regulate chaperone levels in the cell. Regulation of the HSR, therefore, occurs through HSF1-dependent and HSF1-independent mechanisms (Fig. 5.2a). However, the mode of regulation whereby HSR is linked to chaperone levels in the cell cannot explain the poor or incomplete activation of the HSR during development and aging. Such variability in HSR induction suggests the existence of additional, cell-nonautonomous modes of HSR regulation in multicellular organisms. Accordingly, we discuss below five novel pathways that are suggested to regulate HSF1 activation, efficacy, or attenuation in a cell-nonautonomous manner (Fig. 5.2b).

Fig. 5.2
figure 2

Cell-autonomous and cell-nonautonomous regulation of HSF1. (a) Cell-autonomous HSR activation is a stepwise process. (i–ii) Stress-induced appearance of misfolded and damaged proteins, resulting in (iii–v) an equilibrium shift of chaperones from HSF1 to unfolded species, (vi) leading to formation of HSF1 trimers, which (vii) translocate to the nucleus, (viii) bind to HS gene promoters, and induce transcription of HS genes, leading to (ix) elevated levels of chaperones that bind to damaged proteins until the stress signal is alleviated. (b) Cell-nonautonomous signals can integrate into the cell-autonomous response. (1) HSF1 monomers are sequestered by the DHIC complex (1), an event that is regulated by IIS signaling. (2) HSF1 function is modulated by posttranslational modifications, such as acetylation that can be activated by cell-nonautonomous signals. (3) HSF1 binding can be modified by chromatin remodeling, while (4) HSF1 activation can be impacted by other transcription factors that modulate HSP levels and thus their repressive effects on HSF1 trimerization

Full size image

4.1 Neuronal Regulation of HSF1

The first example of cell-nonautonomous regulation of HSF1 was identified as a novel cross talk between an animal’s ability to perceive ambient temperature and the ability of the cell to mount a HSR (Prahlad et al. 2008; Prahlad and Morimoto 2011; Tatum et al. 2015). In C. elegans, two neurons termed AFD neurons are required to sense temperature and regulate temperature-dependent behaviors, such as feeding and reproduction (Lee and Kenyon 2009; Mori et al. 2007). For example, AFD neurons can modulate foraging behavior according to the history of food availability at a given temperature (Mori et al. 2007). Mutations in gcy-8, encoding AFD-specific receptor-type guanylyl cyclases, can disrupt thermotactic behavior (Inada et al. 2006). Work by Prahlad et al. demonstrated that the ability of gcy-8 mutant animals exposed to an acute HS treatment to mount an effective HSR was blocked in different somatic cells. This resulted in low expression of HS genes and reduced thermo-survival (Prahlad et al. 2008). In agreement, excitation of AFD neurons in the absence of stress using optogenetic tools was sufficient to activate HSF1, resulting in reorganization of HSF1 into nuclear puncta in germ cells (Tatum et al. 2015). In contrast, gcy-8 mutant animals were able to mount a HSF1-dependent heavy-metal stress response, revealing specificity to a given sensory input (Prahlad et al. 2008). While the ability to respond to acute HS was inhibited, the ability of gcy-8 mutant animals to upregulate chaperones and maintain proteostasis was enhanced in different somatic cells under chronic stress conditions, such as upon expression of aggregation-prone proteins. Activation of the AFD neurons by a single optogenetic event was sufficient to retard protein aggregation, suggesting that activation of HSR resulted in improved proteostasis (Tatum et al. 2015). Neuronal regulation of the HSR is mediated by the serotonergic neurons ADF and NSM as activated by signals from AFD neurons. Temperature increase resulted in serotonin release in an AFD-dependent manner, leading to preemptive activation of the HSR (Tatum et al. 2015). Thus, in C. elegans, the HSR and proteostatic maintenance of somatic cells depend on sensory neuronal function that modulates the organismal response to stress and can distinguish between acute and chronic stresses (Prahlad et al. 2008; Prahlad and Morimoto 2011; Tatum et al. 2015). Somatic tissues can also send feedback information on the cultivation temperature to the AFD neurons. This response requires HSF1 in both neuronal and nonneuronal cells (Sugi et al. 2011). It remains unknown how these signals are integrated into the cellular response to stress or, specifically, how these signals modulate HSF1. Given that C. elegans exposed to high concentrations of the dauer pheromone that signal crowdedness present reduced HSR activation, it would appear that other signals may also feed into the neuronal pathway, either directly or via a different route (Prahlad et al. 2008). Thus, signals are also likely to be sent to the AFD neurons, integrating environmental signals into the decision of whether or not to activate the HSR in response to temperature shift (Prahlad et al. 2008; Prahlad and Morimoto 2011; Sugi et al. 2011; Tatum et al. 2015).

4.2 IIS-Mediated Regulation of HSF1

IIS-dependent suppression of protein aggregation, as well as lifespan extension, required HSF1 activity, suggesting that cell-nonautonomous regulation of HSF1 plays a role in both proteostasis and lifespan modulation (Alavez et al. 2011; Ben-Zvi et al. 2009; Cohen et al. 2006, 2009; Gidalevitz et al. 2011; Hsu et al. 2003; Morley et al. 2002; Taylor and Dillin 2011). Indeed, Chiang et al. showed that HSF1 is not only held in repressive complexes by chaperones preventing its activation (Abravaya et al. 1992; Shi et al. 1998; Zou et al. 1998) (Fig. 5.1a ii–v) but that HSB-1 and DDL-1/2 also form repressive complexes with HSF1 (Chiang et al. 2012). Formation of these inhibitory complexes, termed DDL-1/2 HSF1 inhibitory complexes (DHIC), is regulated by phosphorylation of DDL-1. DDL-1 phosphorylation, in turn, is regulated by IIS signaling (Chiang et al. 2012). Thus, HSF1 activation can be regulated by its active sequestration into a repressive complex regulated by cell-nonautonomous signals. This mode of regulation may be shared by other regulators, either via phosphorylation of DDL-1 or by assembly of other repressive complexes that affect HSF1 activation in a specific tissue (Fig. 5.2b).

4.3 SIRT1-Mediated Regulation of HSF1

Sirtuins are NAD+-dependent deacetylases that are implicated in lifespan regulation, although the mechanism of such action is not clear (Kenyon 2010b). Given that NAD+/NADH levels are determined by energy homeostasis, sirtuin function is linked to nutrient availability. SIRT1 was shown to deacetylate HSF1 and enhance DNA binding, suggesting that diet can modulate HSF1 release from DNA by removing the acetylation from the K80 acetylation site in the DNA-binding domain of HSF1. This deacetylation promotes the occupancy of HSF1 at the HS promoter sites, enhances HSR activation, and increases thermotolerance (Westerheide et al. 2009). In support of this idea, DR was shown to induce HS gene expression in a SIRT1- (sir-2.1 in C. elegans) dependent manner (Raynes et al. 2012). This suggests that signals that modulate SIRT1 function can affect HSF1. Likewise, other posttranslational modifications of HSF1 could, potentially, be regulated by cell-nonautonomous signals (Fig. 5.2b).

4.4 Germline Stem Cell-Mediated Regulation of HSF1

The coincidence of proteostatic collapse and the onset of reproduction, whereby the activation of different stress responses is strongly inhibited after egg laying begins, supports a link between proteostatic remodeling and reproduction. Several studies have examined the effects of inhibiting reproduction, and specifically GSC arrest, on proteostasis in C. elegans. The functions of several stress transcription factors, such as HSF1 and DAF-16, were modulated by GSC ablation or arrest (Arantes-Oliveira et al. 2002; Berman and Kenyon 2006; Ghazi et al. 2009; Hansen et al. 2005; Hsin and Kenyon 1999). Moreover, germline-less mutant animals were shown to have enhanced expression of proteostasis machinery genes, such as lgg-1, bec-1, and unc-51 required for autophagy (Lapierre et al. 2011) and the rpn-6.1 subunit of the proteasome (Vilchez et al. 2012). Whole genome microarray analysis of P. pacificus in germline-ablated animals identified ~3000 differentially expressed genes, including ribosomal and translation-, proteasome-, and protein folding- and refolding-associated genes (Rae et al. 2012). Moreover, the expression of stress-inducible genes, such as heat shock genes, in response to different assaults was modulated in germline-less animals (Ghazi et al. 2009; Shemesh et al. 2013). Thus, the proteostatic network, including the translational, chaperone, autophagic, and proteasomal machineries, is remodeled by GSC proliferation (Ghazi et al. 2009; Lapierre et al. 2011; Shemesh et al. 2013; Vilchez et al. 2012). The expression of proteostatic components is dependent on different downstream signaling pathways (Lapierre et al. 2011; Shemesh et al. 2013; Vilchez et al. 2012), suggesting that GSC signaling activates different regulatory programs to modulate somatic functions. This change in proteostatic composition is strongly associated with altered proteostatic function. Germline-less animals showed increased autophagosome numbers (Lapierre et al. 2011), increased levels of chymotrypsin-like proteasome activity, and decreased levels of highly polyubiquitinated proteins (Vilchez et al. 2012). GSC proliferation also modulated protein misfolding and aggregation in somatic cells. Germline-less animals displayed reduced aggregation, as well as reduced toxicity of polyQ proteins (Shemesh et al. 2013). Likewise, the functions of metastable and wild-type proteins that are compromised with age, such as UNC-52(ts) and myosin (Ben-Zvi et al. 2009), were rescued in germline-less animals (Shemesh et al. 2013). Finally, stress survival and the activation of stress responses, such as HS, were also rescued by GSC arrest (Alper et al. 2010; Arantes-Oliveira et al. 2002; Ermolaeva et al. 2013; Libina et al. 2003; Shemesh et al. 2013; Stiernagle 2006; TeKippe and Aballay 2010). Thus, inhibition of GSC proliferation reversed the decline in somatic protein quality control upon transition to adulthood.

Labbadia and Morimoto directly examined how HSF1 function is remodeled upon transition to adulthood and in GSC-arrested animals. They examined different regulation nodes in the HSF1 activation cycle, such as HSF 1 levels, nuclear localization, and DNA binding (Fig. 5.2a). Using chromatin immunoprecipitation coupled to qPCR to address specific HS genes, they found that HSF1 and RNA polymerase II association with HS gene promoters following HS was reduced between the first and second days of adulthood (Labbadia and Morimoto 2015b). At the same time, proteostasis is remodeled (Labbadia and Morimoto 2015b; Shemesh et al. 2013). HSF1 transcriptional repression was associated with reduced chromatin accessibility and specifically, with an increased histone methylation repressive marker, H3K27me3, on HS, as well as other stress genes. Moreover, changes in the expression levels of the H3K27me3 demethylase jmjd-3.1 were correlated with the onset of proteostasis collapse (Labbadia and Morimoto 2015b; Shemesh et al. 2013). Thus, the promoter regions of HSF1 target genes are more repressed on the second day of adulthood. In contrast, the levels of jmjd-3.1 and H3K27me3 markers on the promoter of HS genes in GSC-arrested animals remained low (Labbadia and Morimoto 2015b), in agreement with increased HS gene activation and HS survival collapse (Labbadia and Morimoto 2015b; Shemesh et al. 2013). Taken together, these data strongly suggest that HS activation is actively repressed upon transition to adulthood by cell-nonautonomous signals from the reproductive system (Fig. 5.2b). These signals globally change the chromatin accessibility of HSF1 and likely of other transcription factors, resulting in remodeling of several different signaling pathways upon transition to adulthood by GSC inhibition, with these signaling pathways drastically remodeling somatic proteostasis. This switch-like mechanism links reproduction to maintenance of the soma. However, the nature of the signals sent from the reproductive system to the soma and how they modulate jmjd-3.1 levels remain unknown.

4.5 Tissue-to-Tissue HS Signaling

As noted above, HSF1 is regulated by chaperones, such as Hsp90 and Hsp70 that form a repressive complex with HSF1 and inhibit its trimerization (Abravaya et al. 1992; Shi et al. 1998; Zou et al. 1998). However, chaperone levels are also regulated cell-nonautonomously by transcellular chaperone signaling that is regulated by the FOXA transcription factor PHA-4 in C. elegans (van Oosten-Hawle et al. 2013). Enhanced expression of Hsp90 in one tissue was sufficient to elevate the levels of Hsp90 in other tissues and thus block the induction of thermo-protective HSR in the distal tissues. Conversely, knocking down the expression of Hsp90 in a single tissue was sufficient to induce HS genes in distal tissues and protect the animals from stress. Such transcellular chaperone signaling was dependent on PHA-4 regulation of chaperone expression (van Oosten-Hawle et al. 2013). Thus, PHA-4 and possibly other transcription factors that modulate chaperone levels can impact the cell’s ability to mount an effective stress response (Fig 5.2b).

5 Perspectives

The findings that proteostasis is actively remodeled upon transition to adulthood and that HSF1 plays a role in proteostatic maintenance and lifespan open many questions. For example, how are regulation of proteostasis and lifespan linked? How does the decline in stress activation impact lifespan? While we do not yet know the answers to these questions, it is interesting to note that downregulation of any of the GSC-dependent signaling pathways is required for increased lifespan (Antebi 2012), even though these pathways differentially modulated proteostatic functions and stress response regulation (Lapierre et al. 2011; Shemesh et al. 2013; Vilchez et al. 2012). This suggests that only the combined actions of many protective pathways is sufficient for lifespan enhancement, while proteostatic function can be modulated by partial activation of these pathways. This is in agreement with chemical and genetic manipulation of lifespan-extending pathways in which lifespan, proteostasis, and stress resistance were mechanistically dissociated (El-Ami et al. 2014; Tissenbaum 2012; Van Raamsdonk and Hekimi 2012; Yu and Driscoll 2011).

Recent studies have identified several signaling pathways that can regulate HSF1, as well as defining points in HSF1 regulation that can be modulated by cell-nonautonomous signals. Still, many parts of this puzzle are missing. For example, how does neuronal regulation modulate HSF1 function? How are the different cell-nonautonomous signals integrated at the level of the cell, as well as at the level of the organism, to mount an effective HSR? Finally, focusing on lifespan modulation and treatment of age-dependent diseases, the most important question asks whether the age-dependent decline in stress activation and proteostasis can be reversed.