Could Sirtuin Activities Modify ALS Onset and Progression?

Abstract

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease with a complex etiology. Sirtuins have been implicated as disease-modifying factors in several neurological disorders, and in the past decade, attempts have been made to check if manipulating Sirtuin activities and levels could confer benefit in terms of neuroprotection and survival in ALS models. The efforts have largely focused on mutant SOD1, and while limited in scope, the results were largely positive. Here, the body of work linking Sirtuins with ALS is reviewed, with discussions on how Sirtuins and their activities may impact on the major etiological mechanisms of ALS. Moving forward, it is important that the potentially beneficial effect of Sirtuins in ALS disease onset and progression are assessed in ALS models with TDP-43, FUS, and C9orf72 mutations.

Introduction

Amyotrophic lateral sclerosis (ALS) is a rapidly progressive and fatal motor neuron disease with the demise of cortical and spinal motor neurons, and death often results from the eventual loss of respiratory capacity (Zarei et al. 2015; Riva et al. 2016). There is no curative treatment, or any effective therapeutic intervention. Other than physical therapy and assisted ventilation at late stages, the only specific FDA-approved drug is the glutamate transmission moderating riluzole (Miller et al. 2012), which prolongs median survival by 2–3 months, but does not reverse motor neuron degeneration. ALS is a clinically complex disease (Turner and Swash 2015). Although largely sporadic and idiopathic, about 10% of ALS cases are hereditary, with more than 20 gene mutations now known to predispose individuals to familial ALS (Iguchi et al. 2013; Renton et al. 2014). The genetic and clinical view of ALS has changed drastically since the initial identification of mutations in Cu–Zn superoxide dismutase (or SOD1) (Rosen et al. 1993) as a prevalent cause of familial ALS. One major new understanding came with the recognition of a key pathological feature (ubiquitinated inclusions of the RNA/DNA-binding protein TAR DNA-binding protein 43 (TDP-43)) that is shared between familial/sporadic ALS and frontotemporal dementia (FTD) (Arai et al. 2006; Neumann et al. 2006; Mackenzie et al. 2007). This was followed by the identification of ALS causing mutations in another RNA-binding protein, Fused in Sarcoma (FUS) (Kwiatkowski et al. 2009; Vance et al. 2009). The ALS–FTD link and the renewed clinical perception of these as separate ends of a “disease spectrum” with multisystem disorder is further consolidated by the identification of the C9orf72 non-coding hexanucleotide repeat expansion as the most common mutation in familial ALS and FTD (DeJesus-Hernandez et al. 2011; Renton et al. 2011; Bennion Callister and Pickering-Brown 2014). Despite the identification of causative mutations and extensive investigations into the pathological mechanisms underlying ALS, there has been a paucity of advances made in terms of therapeutic interventions (Bartus et al. 2016; Stoica and Sena-Esteves 2016; Mazzini et al. 2016).

The class III histone deacetylases, or Sirtuins (named after its founding member Silent information regulator 2 (Sir2) (Imai et al. 2000) in S. cerevisiae) are NAD⁺-dependent protein deacetylases (Guarente 2007; Haigis and Sinclair 2010) with a large set of histone and non-histone substrates (Feige and Auwerx 2008). The human genome hosts seven Sirtuin paralogues (SIRT1-7) with differing primary subcellular localizations and sites of action (Haigis and Sinclair 2010). SIRT1, SIRT6, and SIRT7 are predominantly nuclear, while SIRT3, SIRT4, and SIRT5 are mitochondria localized (Gertz and Steegborn 2016). SIRT2, on the other hand, is largely cytoplasmic and is a tubulin deacetylase (Li and Yang 2015). Yeast Sir2 is a part of transcription repressor complexes functioning in expression silencing of multiple yeast gene loci (Imai et al. 2000), and the other Sirtuin orthologues likewise function in transcription regulation via deacetylation of histone, transcription factors, and as components of transcriptional complexes (Feige and Auwerx 2008; Bosch-Presegué and Vaquero 2015).

In the past two decades, extensive investigations in multiple cellular and animal models have implicated Sir2p and its homologues in a myriad of physiological and pathophysiological roles. Sirtuins regulate some key cellular and systemic processes that include specific cell/tissue survival (Hao and Haase 2010; Matsushima and Sadoshima 2015; Petegnief and Planas 2013), energy expenditure and metabolic control (Houtkooper et al. 2012; Nogueiras et al. 2012; Chang and Guarente 2014), as well as tissue homeostasis and repair (Tang 2011; Ng and Tang 2013; Poulose and Raju 2015). Sirtuins’ activities are dependent on the availability of NAD⁺, which changes according to cellular energy and redox status. Sirtuins are therefore energy and redox sensors, and have indeed been strongly associated with healthspan and lifespan extension effect of dietary or caloric restriction (Cohen et al. 2004; Guarente 2013). A prominent and somewhat controversial aspect of pharmacological Sirtuin regulation is its activation by resveratrol (Pallàs et al. 2013; Hubbard et al. 2013), a natural stilbenoid found in abundance in the skin of grapes and berries which has been shown to promote healthspan and lifespan in yeast (Howitz et al. 2003) and animal models (Wood et al. 2004; Lagouge et al. 2006; Baur et al. 2006). Sirtuin-activating compounds (STACs) with higher specificity and bioavailability than resveratrol have been synthesized and these have much touted clinical promise as therapeutics against aging-associated disorders (Milne et al. 2007; Bonkowski and Sinclair 2016).

Sirtuin activities are often associated with neuronal survival (Min et al. 2013; Herskovits and Guarente 2013; Ng et al. 2015). SIRT1, in particular, is highly expressed in the mammalian brain (Zakhary et al. 2010; Michán et al. 2010) and its activation was shown to confer survival benefits in acute physical trauma and ischemic injury (Zhao et al. 2012a, b; Koronowski and Perez-Pinzon 2015). SIRT1 and the mitochondrial SIRT3 have also been shown to confer benefits in a range of neurodegenerative diseases (Srivastava and Haigis 2011; Min et al. 2013; Ng et al. 2015), prominently Alzheimer’s disease (AD) (Wong and Tang 2016) and Parkinson’s disease (PD) (Tang 2016a, b). On the other hand, SIRT2 activity has been negatively implicated in experimental stroke (Krey et al. 2015), PD (Outeiro et al. 2007), and Huntington’s disease (HD) models. Like that of AD and PD, a number of in vitro and in vivo studies have linked Sirtuins to a potential role in modifying ALS neuropathology. These studies were, however, largely confined to the use of the human SOD1G93A mutant expressed in cells or transgenic mice. This has been a major limitation in the field. In the paragraphs that follow, we summarize the results obtained from these studies and discuss the potential of Sirtuins functioning as modifiers of ALS disease onset and progression.

Sirtuin Levels and Activity Changes in ALS Models and Post-mortem Samples

Transgenic overexpression of SOD1, such as SOD1G93A, recapitulates key neurodegenerative features of ALS, and is thus widely used in cellular and animal models for the disease (Jaarsma et al. 2000). Expression profiling analysis over the years has documented temporal and regional changes in a good number of genes during disease progression (Chen et al. 2004; Ferraiuolo et al. 2007; de Oliveira et al. 2013), and these include markers for other neurodegenerative diseases such as amyloid precursor protein (APP) (Koistinen et al. 2006; Rabinovich-Toidman et al. 2015), tau (Barańczyk-Kuźma et al. 2007), DJ-1 and PTEN-induced putative kinase 1 (PINK1) (Lev et al. 2009; Knippenberg et al. 2013), as well as cyclin-dependent kinase 5 (Cdk5) (Nguyen et al. 2001). SIRT1 is enriched in CNS neurons (Zakhary et al. 2010; Michán et al. 2010), and its levels are known to be reduced in brain tissues of neurodegenerative disease subjects such as Huntington’s disease (Pallàs et al. 2008) and AD (Julien et al. 2009; Lutz et al. 2014). SIRT1 level was found to be reduced in a SOD1G93A expressing motor neuron–neuroblastoma hybrid cell line, ventral spinal cord 4.1 (VSC 4.1), and resveratrol treatment appeared to elevate SIRT1 levels (Wang et al. 2011). Interestingly, a significant increase in SIRT1 expression was observed in the cerebral cortex, hippocampus, thalamus, and spinal cord of symptomatic SOD1G93A mice (Lee et al. 2012). SIRT1 levels in the spinal cords of mutant SOD1G37R mice correlated with disease progression, and become significantly upregulated when neurodegeneration was severe (Kim et al. 2007).

Another study had however found SIRT1 levels to be decreased in the spinal cord, but increased in muscles during disease progression in mice (Valle et al. 2014). SIRT2 transcripts are elevated in the spinal cord in both SOD1G93A and SOD1G86R mice, but this was not reflected in the protein levels (Valle et al. 2014). Analysis of transcript and protein levels of post-mortem human brain tissues instead revealed a significant reduction in SIRT1 and SIRT2 levels in the homogenates of the primary motor cortex (Körner et al. 2013). However, in situ hybridization and immunohistochemistry revealed neuron-specific upregulation of SIRT1, SIRT2, and SIRT5 in the spinal cord (Körner et al. 2013). These discrepancies observed are probably due to differences in Sirtuin level changes in different cell types that vary over time during disease progression. SIRT1, for example, may be transiently upregulated in neurons and glia during the early phase of ALS because of redox and toxic stress, but as neurons and accompanying cells are nearing their time of demise, the levels would likely be diminished.

Studies on the Roles of Sirtuins in Cellular Models of ALS

While intervention-type experiments with either resveratrol/STACs or genetic overexpression of Sirtuin genes have been performed with the major neurodegenerative disorders AD and PD, ALS has received comparatively less attention. The reported works with cellular models are summarized in Table 1. One of the first investigations in this regard by Kim and colleagues showed that resveratrol treatment significantly attenuated SOD1G93A-mediated neurotoxicity in transfected primary neurons (Kim et al. 2007). This attenuation was also observed with the co-transfection of SIRT1, but not with a deacetylase-dead SIRT1-H363Y mutant. In another report, death of the motor neuron-like VSC 4.1 cells transfected with SOD1G39A was attenuated by resveratrol treatment (Wang et al. 2011). Interestingly, resveratrol elevated the SOD1G39A-depressed SIRT1 levels (compared with cells transfected with wild-type SOD1 or empty vector) in VSC 4.1 cells. While Sirtuin transcript and protein levels have been shown to be altered by resveratrol treatment in other contexts (Franco et al. 2010; Schirmer et al. 2012), it is possible that this effect may be indirect and at least partially due to the improved cell survival and health.

Table 1.

A summary of reports linking manipulation of Sirtuin levels or activities in ALS cellular and animal models

ALS model	Intervention	Reference
Neuronal culture models
Rat cortical primary neurons transfected with SOD1G93A	1. Resveratrol attenuated mutant SOD1G93A-mediated neurotoxicity 2. Overexpression of SIRT1 (but not a deacetylase-dead mutant) protected against SOD1G93A toxicity	Kim et al. (2007)
Motor neuron-like ventral spinal cord 4.1 cells (VSC 4.1) transfected with SOD1G93A	Resveratrol treatment promoted neuronal viability, which was partially blocked by SIRT1 inhibition	Wang et al. (2011)
Exogenous expression of SIRT3 in motor neuron cultures from SOD1G93A mice	SIRT3 and its substrate PGC-1α protected against SOD1G93A-induced mitochondrial defects and cell death	Song et al. (2013)
Co-culture of SOD1G93A spinal cord astrocytes with motor neurons	Enhancing NAD+ salvage pathway and overexpression of SIRT3 (but not SIRT1) attenuated the neurotoxic phenotype of SOD1G93A astrocytes	Harlan et al. (2016)
Mouse models
Transgenic SOD1G93A mice	Resveratrol (single dose) treatment did not improve survival or motor performance	Markert et al. (2010)
Transgenic SOD1G93A mice	Resveratrol delayed disease onset and extended survival, evidence for HSP1 and p53 deacetylation	Han et al. (2012)
Transgenic SOD1G93A mice	Resveratrol significantly extended lifespan and promoted survival of spinal motor neurons	Mancuso et al. (2014)
Transgenic SOD1G93A mice	Resveratrol significantly delayed the disease onset and prolonged the lifespan of ALS mice	Song et al. (2014)
Transgenic SOD1G93A mice with low or high mutant protein expression	Crossing mutant SOD1 mice with PrP-SIRT1 transgenics with overexpression of SIRT1 in the brain and spinal cord. Sirt1 transgene conferred longer lifespan for low levels of mutant SOD1 mice, but did not alter time of symptomatic onset	Watanabe et al. (2014)

Other than SIRT1, the mitochondrial SIRT3 was also shown to promote neuronal survival against SOD1G93A. Expression of the mutant caused mitochondrial fragmentation and arrested mitochondrial transport in culture rat spinal cord motor neurons, and this defect is countered by the expression of both SIRT3 and its substrate Peroxisome proliferator-activated receptor γ Coactivator-1α (PGC-1α) (Song et al. 2013). Motor neuron demise in SOD1G93A-associated ALS is a complex phenomenon, and has non-cell autonomous components. It has been demonstrated that astrocytes are critical determinants of ALS progression (Yamanaka et al. 2008). Mutant SOD1 expressing astrocytes caused degeneration of co-cultured motor neurons (Nagai et al. 2007; Di Giorgio et al. 2007). In fact, astrocytes generated from post-mortem tissue (Haidet-Phillips et al. 2011) or by reprograming fibroblasts (Meyer et al. 2014) of both familial and sporadic ALS patients, exhibited non-cell autonomous toxicity towards co-cultured neurons. It was recently shown that enhancing the NAD⁺ salvage pathway (which would presumably activates Sirtuins) in astrocytes isolated from SOD1G93A mice or spinal cord-derived astrocytes from post-mortem ALS subjects attenuated their toxicities toward co-cultured motor neurons. Furthermore, overexpression of SIRT3 (but not SIRT1) rescued the motor neuron toxicity conferred by SOD1G93A astrocytes (Harlan et al. 2016).

On the other hand, the role of other Sirtuin paralogues in ALS is unclear. SIRT2, which is known to act negatively in promoting degeneration in other neurodegenerative disease models (Outeiro et al. 2007; Luthi-Carter et al. 2010), may not have a significant role in ALS. This was illustrated by the observation that the SIRT2 inhibitor AGK2 did not rescue cells from the toxicity of SOD1G93A (Valle et al. 2014). Furthermore, HDAC6 deletion, but not SIRT2 deletion, attenuated motor neuron degeneration in SOD1G93A mice (Taes et al. 2013).

Studies on the Roles of Sirtuins Using in Vivo Models of ALS

Multiple pathogenic processes contribute to symptoms in ALS (Ferraiuolo et al. 2011), and measurements of motor activity and overall survival in transgenic mice overexpressing mutant SOD1 (Mancuso and Navarro 2015) are common methods to assess disease progression in vivo. An early attempt using a single-dose resveratrol treatment of SOD1G93A mice did not result in any measurable benefit in terms of motor performance and survival extension (Markert et al. 2010). However, subsequent reports were more positive with regards to a beneficial effect of resveratrol that were administered in multiple doses (see Table 1). Han et al. reported both a delay in disease onset and extended survival using a continuous dosing regimen of resveratrol via peritoneal injection into SOD1G93A mice (Han et al. 2012). The authors observed a deacetylation of heat shock factor 1 (HSF-1) and consequential upregulation of Heat shock protein (HSP) 25 and HSP70 in spinal cord lysates. Deacetylation of p53, a major SIRT1 substrate (Vaziri et al. 2001; Luo et al. 2001), served as a proxy for SIRT1 activation (Han et al. 2012). In another report, Mancuso et al. gave SOD1G93A mice a resveratrol-enriched diet and observed delayed disease onset, improved survival of spinal motor neurons, and better preserved motor neuron function (Mancuso et al. 2014). Resveratrol appeared to increase SIRT1 levels and activity as indicated by increased p53 deacetylation, as well as normalizing autophagic flux and mitochondrial function. The latter observations thus linked known SIRT1 and SIRT3 activities in autophagy (Ng and Tang 2013) and mitochondria biogenesis (Tang 2016a, b) with resveratrol treatment. Song et al. made similar observations pertaining to ALS disease onset and survival also with daily peritoneal administration of resveratrol (Song et al. 2014), and found that resveratrol increased both SIRT1 and Peroxisome proliferator-activated receptor gamma coactivator-1α (PGC-1α) levels in the spinal cord.

In a complementary approach, Watanabe et al. crossed a SIRT1 overexpressing transgenic line (SIRT1 under the prion precursor protein (PrP) promoter, with high CNS levels) with two SOD1G93A lines with differing mutant SOD1 protein expression levels (Watanabe et al. 2014). High CNS SIRT1 expression was able to extend the lifespan and slowed disease progression in the line with lower levels of SOD1G93A, but not the one with a higher level. For the lower SOD1G93A expressing line, the authors observed that the SIRT1 overexpression caused hypoacetylation of HSF-1 and elevated the expression of an inducible heat shock protein, HSP70i, with the latter apparently reducing the formation of insoluble SOD1 dimers (Watanabe et al. 2014). In the mouse line with higher SOD1G93A, however, further upregulation of HSP70i by SIRT1 was not observed, and the expression of this chaperone protein appeared to be already at its maximum. Therefore, in agreement with the results of Han et al. (Han et al. 2012), reduction of mutant SOD1 toxicity via the enhanced expression of chaperone proteins of the HSP family could be an important mode of SIRT1’s neurotoxicity attenuation and survival promotion mechanism.

Another link between Sirtuins and ALS was demonstrated through an important SIRT1 substrate, namely PGC-1α (Rodgers et al. 2005; Lagouge et al. 2006), which could in turn regulate SIRT3 expression (Kong et al. 2010; Giralt et al. 2011). SIRT1’s deacetylation of PGC-1α is a key aspect of metabolic regulation and mitochondrial homeostatic activity exerted by the deacetylase (Rodgers et al. 2005; Lagouge et al. 2006; Gerhart-Hines et al. 2007; Tang 2016a, b). In this regard, Pasinetti’s group has shown that overexpression of PGC-1α in a double transgenic model significantly improved motor function and survival of SOD1G93A mice (Zhao et al. 2011). The authors also showed that a ketogenic diet enriched in medium chain fatty acids, the levels of which are reciprocally influenced by Sirtuin activities (Shimazu et al. 2010), improved mitochondrial respiration and alleviated motor symptoms of SOD1G93A mice (Zhao et al. 2012a, b).

Sirtuins as Modifiers of ALS Disease Onset and Progression: Potential Mechanisms

The investigations described above, albeit largely SOD1G93A based and therefore limited in scope, do provide tantalizing evidence that Sirtuin activities, particularly those of SIRT1 and SIRT3, could potentially act in modifying ALS disease onset and progress. Earlier perceptions of ALS pathology have largely focused on two main classes of etiological impairments that are interconnected, namely (1) mitochondrial dysfunction and oxidative stress, and (2) protein aggregation and impairment in the clearance of toxic proteins or their aggregates. These lead to, or are accompanied by defects in axonal transport, neuroinflammation, and excitotoxicity that cause neuronal death and axon–muscle detachment. With the findings of the involvement of mutations in the RNA-binding proteins TDP-43 and FUS, as well as the hexanucleotide repeat expansion in C9orf72 in familial and sporadic ALS, defects in RNA metabolism move into focus as an important etiological factor. In the following sections, we discuss how Sirtuin activities may confer benefit, or work against, these etiologically important impairments for ALS.

Mitochondrial Dysfunction and Oxidative Stress

Although the pathology of ALS is complex (Rossi et al. 2016), one common and prominent pathological manifestation of the disease is mitochondrial dysfunction (Carrì et al. 2016). While oxidative stress could cause mitochondrial dysfunction, the mitochondria is itself a major source of cellular reactive oxygen species (ROS) and damaged mitochondria could produce more neuronal demise-promoting oxidative stress in many neurodegenerative diseases (Federico et al. 2012). Some proteins mutated in familial ALS (and also found in sporadic cases) are known to cause mitochondrial damage or dysfunction. The most obvious in this regard is mutant SOD1, whose association with mitochondria impairs mitochondrial dynamics and function (Magrané et al. 2009; Song et al. 2013; Tafuri et al. 2015). Mutation in a mitochondrial-localized protein, Coiled-coil-helix-coiled-coil-helix domain 10 (CHCHD10) which has an apparent role in stabilizing mitochondrial crista structure, has been recently associated with ALS/FTD (Bannwarth et al. 2014). The ER-resident Vesicle-associated membrane protein-associated protein B (VAPB), which is mutated in a familial form of ALS (Nishimura et al. 2004), functions in ER–mitochondria association by interacting with mitochondrial protein tyrosine phosphatase-interacting protein-51 (PTPIP51) to regulate calcium levels (De Vos et al. 2012). Interestingly, this interaction is disrupted by both mutant TDP-43 (Stoica et al. 2014) and FUS (Stoica et al. 2016). In general, therefore, ALS mutations and pathologies elicit an entangled web of reciprocal feedbacks between mitochondrial dysfunction and oxidative stress.

Sirtuins could counter mitochondrial dysfunction and oxidative stress in ALS, and it has been proposed to do also for AD, PD, and a number of other neurodegenerative disorders. SIRT1 and SIRT3 are known to be important regulators of mitochondrial biogenesis and function. As indicated above, SIRT1’s deacetylates and activates a key transcription factor, PGC-1α (Nemoto et al. 2005; Rodgers et al. 2005; Lagouge et al. 2006), thus regulating PGC-1α-controlled genes that in turn control metabolism and mitochondrial biogenesis (Tang 2016a, b; Brenmoehl and Hoeflich 2013; Yuan et al. 2016). PGC-1α is required for the induction of several ROS-detoxifying enzymes, and the SIRT1 activation of PGC-1α may thus also be important in countering mitochondrial oxidative stress (St-Pierre et al. 2006). SIRT3, on the other hand, appears to be an important factor downstream of PGC-1α in terms of mitochondrial function (Kong et al. 2010). A key regulator of mitochondrial metabolic enzymes (Hirschey et al. 2010), SIRT3 acts as well in mitochondria protective pathways such as the mitochondrial unfolded protein response (Papa and Germain 2014) and anti-oxidative damage responses (Someya et al. 2010). One way by which SIRT3 counters mitochondrial oxidative damage is through deacetylation of FOXO transcription factors and upregulation of genes that regulate mitochondrial homeostasis (Tseng et al. 2013). SIRT1, SIRT3, PGC-1α, and FOXO proteins may therefore constitute a transcription control network, whose levels and activities could modulate ALS mutant protein-induced mitochondrial dysfunction and oxidative damage. In this regard, it has been proposed that modulation of Sirtuin activities may influence ketogenic therapeutic approaches to improve mitochondrial function in ALS (Pasinetti et al. 2013). Collectively, these could essentially promote the survival of motor neurons in ALS.

Protein Misfolding, Aggregation Formation and Clearance Impairment

Another prominent pathological hallmark of ALS is misfolded protein accumulation and the formation of aggregates (Parakh and Atkin 2016). The presence of SOD1 mutant protein aggregates in transgenice mice, and neural as well as other human tissues, has been well documented (Watanabe et al. 2001; Jonsson et al. 2008). TDP-43 has an intrinsic propensity to aggregate, and this tendency is enhanced by ALS-associated mutations (Johnson et al. 2009; Nonaka et al. 2009). FUS is also known to form both cytoplasmic as well as nuclear aggregates (Sun et al. 2011; Nomura et al. 2014). The TDP-43/FUS family of proteins harbor prion-like domains that allow them to self-associate (Li et al. 2013) and to form functional ribonucleoprotein complexes (Kim et al. 2010), and TDP-43 mutations could promote its complex formation with FUS (Ling et al. 2010). The expanded nucleotide repeats within the C9ORF72 gene could generate abnormal dipeptide repeat proteins from repeat-associated non-AUG (RAN) translation of the repeat-containing transcripts (Cleary and Ranum 2013), and these form oligomers or aggregates (Wen et al. 2014; Porta et al. 2015; Schipper et al. 2016). The degree of ALS-associated protein misfolding and toxic oligomer formation (Fang et al. 2014) may therefore play a significant role in disease initiation. Insoluble aggregates, depending on its composition and nature, could however in some cases be neuroprotective (Cragnaz et al. 2014).

Accumulation of misfolded proteins and aggregate formation impair cellular regulation of protein biogenesis and turnover, or proteostasis, in multiple ways (Ruegsegger and Saxena 2016). For example, the aggregates could disrupt both axonal transport as well as membrane traffic within the soma, both of which would be detrimental to the motor neuron with a long axon. We focus here on two general ALS-linked proteostatic disruption mechanisms that could be helped by maintaining or elevating Sirtuin activities. Firstly, it is possible that global protein folding in the neuron is impaired by the sequestration of chaperone proteins by accumulating ALS protein aggregates. Mutant SOD1 interacts with HSP105 (Yamashita et al. 2007), and other chaperones such as HSP70 (Gifondorwa et al. 2007) and HSJ1a/DNAJB2 (Novoselov et al. 2013), confer survival benefits to SOD1G93A mouse. TDP-43 interacts with members of the HSP40/HSP70 family (Udan-Johns et al. 2014), and HSJ1a also acts to counter TDP-43 aggregation (Chen et al. 2016). FUS is known to interact with mitochondrial HSP60, which apparently aids pathogenic FUS entry into the mitochondria (Deng et al. 2015).

A second general way whereby ALS-associated protein aggregates could disrupt proteostasis is for these to impair or eventually overwhelm key cellular waste disposal mechanisms, namely the ubiquitin–proteasome pathway and macroautophagy. In fact, some ALS-associated mutations are components of these pathways. Mutations in a member of the ubiquitin family, ubiquilin-2, causes X-linked ALS (Deng et al. 2011). Protein aggregation in ALS is clearly connected to mitochondrial dysfunction and oxidative damage described in the section above, as damaged or aged mitochondria are cleared by a mitochondria-based process of autophagy, namely mitophagy (Hamacher-Brady and Brady 2016). Autophagy of damaged mitochondria requires their tagging with polyubiquitin chains via the activity of the E3 ubiquitin ligase Parkin (Durcan and Fon 2015). Mutations in optineurin, which harbors a polyubiquitin-binding domain and is an autophagy receptor for damaged mitochondria (Wong and Holzbaur 2014), causes familial ALS (Maruyama et al. 2010). In this connection, mutations in TANK-binding kinase 1 (TBK1), which enhances optineurin’s binding to ubiquitin chains to promote mitophagy of damaged mitochondria (Richter et al. 2016), also causes ALS (Freischmidt et al. 2015). Another autophagy receptor that is found to be linked to both familial and sporadic ALS is sequestosome-1 (SQSTM1)/p62 (Fecto et al. 2011), which is known to link mutant SOD1 to its clearance via autophagy (Gal et al. 2009). Autophagy-associated function has also recently been shown for C9orf72 (Yang et al. 2016; Sullivan et al. 2016; Webster et al. 2016; Sellier et al. 2016).

How could Sirtuin activities counter proteostasis impairment in ALS? Firstly, SIRT1 deacetylates an important transcriptional controller of the heat shock response, namely HSF-1 (Anckar and Sistonen 2007). HSF-1 deacetylation by SIRT1 promotes and stabilizes its binding to the promoter of HSP70 (Westerheide et al. 2009). HSF-1 activity and heat shock response are also regulated by SIRT1 activity regulators, and SIRT1-deficient mouse embryonic fibroblasts are defective in HSP70-dependent protein quality control (Tomita et al. 2015). HSF-1 augments the expression of synaptic proteins that are essential for synaptic function and stability (Hooper et al. 2016), and SIRT1-activated HSF-1 would therefore likely confer protection against neuromuscular junction failure in ALS (Fig. 1). A potential role for SIRT1 activity in modifying ALS onset and progression resulting from transgenic SOD1 mutants through HSF-1 was clearly illustrated by the benefits conferred by CNS SIRT1 overexpression (Watanabe et al. 2014). SIRT1 could of course act directly on some cytoplasmic chaperonins and enhance their modes of action, such as that documented for cytoplasmic chaperonin containing TCP-1 (CCT) (Gan and Tang 2010).

Fig. 1.

A schematic diagram illustrating the potential mechanisms whereby Sirtuin activities could modify ALS disease onset and progression. SIRT1 and SIRT3 have roles in promoting autophagy and mitophagy through their deacetylation of multiple substrates, including FoXO family transcription factors. Through their interactions and influence on mitochondria biogenesis and the expression of anti-oxidative stress enzymes, these Sirtuins promote mitochondrial health and clearance of damaged mitochondria. SIRT deacetylation of HSF-1 enhances the expression of heat shock protein (HSP) family chaperones that would curb protein misfolding and aggregate formation in the soma, as well as the expression of synaptic components that could improve synaptic integrity. SIRT1’s decetylation of histone (H) and chromatin silencing effect may (speculatively) attenuate undesirable transcription of mutant C9orf72 and the generation of RAN-translated dipeptides. Taken as a whole, Sirtuin activities are likely to confer disease benefit in terms of promoting motor neuron survival, attenuating disruption of axonal transport, and preserving the integrity of the neuromuscular junction (NMJ) (cartoons in boxes). Disease-associated effects or damages are indicated by black arrows (→) while Sirtuin effects are indicated by red arrows (→) (enhancement) and (⊣) (inhibition) (Color figure online)

Secondly, Sirtuins are known modulators of autophagy (Ng and Tang 2013; Tang 2016a, b). SIRT1 regulates autophagy by directly deacetylating key autophagy proteins, including Atg5, Atg7, and Atg8 (Lee et al. 2008). It could also regulate autophagy indirectly through its deacetylation of key transcription factor substrates such as FoXO (Sengupta et al. 2009; Hariharan et al. 2010), p53 (Tasdemir et al. 2008), the RelA/p65 (Criollo et al. 2010), or via modulation of AMPK activation (Wu et al. 2011; Song et al. 2015) (Fig. 1). SIRT3 is also involved in autophagy regulation (Liang et al. 2013; Pi et al. 2015; Duan et al. 2016). Sirtuin activities have also been implicated in the induction or regulation of organelle autophagy of the mitochondria, i.e., mitophagy (Tang 2016a, b). Mitophagy is also induced by the NAD⁺ precursor nicotinamide, and this is likely due to SIRT1 activation (Kang and Hwang 2009; Jang et al. 2012). Silencing of mitochondrial acetyltransferase diminished mitochondrial protein acetylation levels and promoted accumulation of autophagy mediators in the mitochondria, a process that is impaired by SIRT3 depletion (Webster et al. 2013). SIRT3 deacetylates FoXO3 (Tseng et al. 2013), which leads to the upregulation of mitophagy (Das et al. 2014). Sirtuin activities’ promotion of autophagy and mitophagy could thus be potentially beneficial for ALS etiology by enhancing the clearance of protein aggregates and damaged mitochondria. On the whole, these activities should promote the survival of diseased motor neurons.

Sirtuins Influencing ALS Genes/Proteins and Disease Progression: Extending the Investigations

Despite the connections shown above and the work summarized in Table 1, it is clear that there is much more to be learned with regard to the connection between Sirtuins and ALS genes, proteins, and pathology. All the investigations reported thus far had focused on the use of the SOD1G93A mutation, and nothing is yet known about how manipulating Sirtuin levels or activities will influence ALS disease onset and progression using other ALS models. An interesting connection that has surfaced recently is a link between TDP-43 and SIRT1 (Yu et al. 2012). SIRT1 expression could apparently be influenced by a complex between TDP-43, the fragile X mental retardation protein (FMRP) and Staufen, both components of ribonucleoprotein particles that mediate transport and translation of neuronal RNAs (Barbee et al. 2006). It would be interesting and important to check if SIRT1 levels are affected by ALS-associated TDP-43 mutations, and how SIRT1 activation may in turn influence TDP-43 pathology.

Examining the effect of Sirtuin level and activity manipulations in FUS (Nolan et al. 2016)- and C9orf72 (Liu et al. 2016)-based ALS mouse models would also be pertinent. Other than the major disease-modifying mechanisms discussed above, there may be related or new avenues of Sirtuin action that may confer neuroprotective and survival extension benefits. For example, given the histone deacetylation and chromatin silencing effect of SIRT1, transcription of mutant repeats of C9orf72 may be speculatively suppressed, with consequential reduction of RNA metabolism perturbations and RAN dipeptide generation (Fig. 1). A much-debated mode of SOD1 and TDP-43 pathology is a prion-like misfolding property and its propagation (Grad et al. 2015). In this regard, it is worth noting that SIRT1 has been shown to protect against PrP-induced neuronal death (Bizat et al. 2010; Seo et al. 2012; Jeong et al. 2013). Clearly, SIRT1’s induction of chaperones and autophagy could play significant roles in mitigating pathological propagation of toxically folded proteins as well as enhancing their clearance.

A particular interesting aspect of motor neuron death in ALS pertains to the non-cell autonomous effect of glia, namely microglia and astrocytes on motor neurons (Clement et al. 2003; Ferri et al. 2004; Boillée et al. 2006). SOD1G93A-expressing astrocytes exhibit secretory toxicity towards motor neurons in co-culture (Di Giorgio et al. 2007; Nagai et al. 2007; Di Giorgio et al. 2008; Marchetto et al. 2008). Although the mechanistic nature of this non-cell autonomous effect is not yet particularly clear, the concept of glia-mediated neuronal demise via neuroinflammation is well known. In a humanized co-culture model, death of motor neurons induced by primary astrocytes from either sporadic or familial ALS was shown to occur by necroptosis (Re et al. 2014). SIRT1 is a suppressor of inflammation (Hwang et al. 2013) and the pro-inflammatory senescence-associated secretory phenotype (Hayakawa et al. 2015). Moreover, it was recently shown that enhancing the NAD+ salvage pathway, which likely enhances Sirtuin activities, could attenuate neurotoxicity of astrocytes expressing mutant SOD1 (Harlan et al. 2016). Understanding the role of Sirtuins in astrocytes and how these modulate astrocyte-based motor neuron toxicity should thus be an important future pursuit.

Sirtuin-Based Therapeutics for ALS?

Clinical trial for ALS therapeutics in the past decade has been largely negative and disappointing, and riluzole is the only approved disease-modifying pharmacological intervention thus far (Mitsumoto et al. 2014). No Sirtuin or Sirtuin activator-based ALS trials have yet been reported. Given the tolerability of resveratrol and its significant modification of CNS biomarkers reported in a recent AD trial (Turner et al. 2015), conducting such a trial for ALS would be pertinent and justified if clear beneficial effects of Sirtuin activities are eventually demonstrated in other ALS animal models. Resveratrol has multiple targets such as AMPK (Dasgupta and Milbrandt 2007) and cAMP phosphodiesterases (Park et al. 2012), and more specific natural and synthetic STACs (Hubbard and Sinclair 2014) may provide a protective effect that could be more exclusively attributed to Sirtuin activities. Another cautionary note for the use of pan-Sirtuin activators is that not all Sirtuin activities are neuroprotective. SIRT2, for example, has been negatively implicated in neuronal survival in Parkinson’s disease (Outeiro et al. 2007) and Huntington’s disease models (Pallos et al. 2008; Chopra et al. 2012). The role of SIRT2 activity, if any, in motor neuron demise is yet unclear (Taes et al. 2013), and deserve further investigation.