Human Sirtuin Regulators: The “Success” Stories

The human sirtuins are a group of NAD+-dependent protein deacylases. They “erase” acyl modifications from lysine residues in various cellular targets including histones, transcription factors, and metabolic enzymes. Through these far-reaching activities, sirtuins regulate a diverse array of biological processes ranging from gene transcription to energy metabolism. Human sirtuins have been intensely pursued by both academia and industry as therapeutic targets for a broad spectrum of diseases such as cancer, neurodegenerative diseases, and metabolic disorders. The last two decades have witnessed a flood of small molecule sirtuin regulators. However, there remain relatively few compounds targeting human sirtuins in clinical development. This reflects the inherent issues concerning the development of isoform-selective and potent molecules with good drug-like properties. In this article, small molecule sirtuin regulators that have advanced into clinical trials will be discussed in details as “successful” examples for future drug development. Special attention is given to the discovery of these compounds, the mechanism of action, pharmacokinetics analysis, formulation, as well as the clinical outcomes observed in the trials.

The human sirtuins, SIRT1-SIRT7, have been intensively investigated for their enzymatic activities and biological functions. There are numerous wonderful review articles highlighting the significance of these enzymes in regulating normal physiology and pathophysiology (Chang and Guarente, 2014;Herskovits and Guarente, 2014;Kumar and Lombard, 2018;Chang et al., 2020;Jaiswal et al., 2021). For example, overexpression of SIRT1 has been observed to increase carcinogenesis in prostate and thyroid tumors in mice with a deficiency of tumor suppressor PTEN (phosphatase and tensin homolog) (Herranz et al., 2013). SIRT1 also plays an important role in the development of drug resistance in chronic myeloid leukemia (CML) cells . It has been shown to activate error-prone DNA repair which can lead to increased incidence of genetic mutations . Thus, a SIRT1 inhibitor would be beneficial in combination with established chemotherapeutics to reduce drug resistance. Sirtuin activators can also play an important role in disease modulation. SIRT1 is considered to have a neuroprotective role in the brain, as it helps to regulate apoptosis and BDNF (brain-derived neurotrophic factor) expression (Luo et al., 2001;Zocchi and Sassone-Corsi, 2012). Activation of SIRT1 may be beneficial for the treatment of Alzheimer's disease (AD), in which SIRT1 levels are typically reduced (Lutz et al., 2014). AD is characterized by the presence of amyloid plaques containing amyloid beta (Aβ) and neurofibrillary tangles (NFTs) containing hyperphosphorylated tau (Ittner and Götz, 2011). SIRT1 overexpression was shown to increase α-secretase activity, thus reducing the formation of Aβ (Endres and Fahrenholz, 2012). SIRT1 also prevents AD pathology through the deacetylation of tau (Min et al., 2010). Tau acetylation inhibits the degradation of tau and is detected in early stages of diseases with abnormal tau accumulation (Min et al., 2010). Therefore, SIRT1 serves as a possible therapeutic target for the treatment of AD.
Naturally, the development of small molecule regulators targeting human sirtuins has become a hot topic of research. Despite all the efforts over the last few decades, the success stories were scarce. Many small molecule sirtuin inhibitors and activators can only be called "chemical probes" at the present time due to the lack of isoform selectivity, moderate potency, limited bioavailability, and poor pharmacokinetic (PK) and pharmacodynamic (PD) profiles. There is a clear gap between the pre-clinical probe discovery and clinical drug candidate development. There is also a gap in the amount of research effort put into studying the various sirtuin isoforms. SIRT1 is by far the most studied isoform, with over 11,000 articles indexed in PubMed. In comparison, the other two most studied isoforms, SIRT2 and SIRT3, together have only around 3,700 articles. This disparity in research translates to fewer small molecule modulators targeting the other isoforms. Thus, the modulators that have entered clinical trials are primarily focused on SIRT1.
In this review, we will "tell the tales" of several human sirtuin regulators that have advanced into clinical investigation for the treatment of various diseases. The focus of the discussions will be the discovery of these compounds, their mechanism of action (MOA), and the rationale and outcome of the clinical trials. Although we brand these compounds as the "success" stories, they are not without controversy or limitation. On the flip side, the lessons we learn from these examples may help guide the design and development of the next generation of sirtuin regulators as therapeutic candidates.
For the benefit of the general audience, we would like to briefly discuss the basic theories behind the development of small molecule drug candidates toward clinical trials. The candidate compounds are normally small molecules, either natural products or synthetic compounds, with desired biological activity toward target proteins or enzymes in the in vitro setting. These candidates are the results of rounds of optimization for improved potency, selectivity, and solubility. For example, Lipinski's rule of five (Ro5) has been the golden standard to prioritize the drug-like properties of orally active compounds (Lipinski et al., 2001). The "druggability" of the candidates will then be analyzed through ADME (absorption, distribution, metabolism and excretion) studies. These studies will assess the bioavailability, distribution, stability, and elimination of the candidate compounds. In the following sections, C max of certain sirtuin modulators will be discussed. This critical parameter in ADME analysis describes the maximum concentration of a candidate compound in targeted tissue/organ after administration. The results from ADME studies will guide the further optimization of the candidate molecules. The best dosage, administration route, and formulation need to be evaluated as well. The formulations of resveratrol (RSV) will be discussed in detail in the next section. The active ingredient, RSV, has been combined with variety of substances to improve its bioavailability.

Resveratrol
In the 1980s, epidemiologists observed that developed countries with increased wine consumption had decreased deaths due to ischemic heart disease (IHD) (St Leger et al., 1979). This later came to be known as the "French Paradox" because of the lowered IHD mortality rates in France despite having no difference in saturated fat intake or blood cholesterol levels (Burr, 1995). In Bertelli et al. (1995) implicated resveratrol (RSV, Figure 1) as the mediator of the cardioprotective effects of wine, thus spurring interest in the molecule as a potential therapeutic. The association between RSV and SIRT1 was discovered in 2003 by high-throughput screening (HTS) utilizing a Fluor de Lys deacetylation assay on a library of plant-derived polyphenols (Howitz et al., 2003). Howitz et al. (2003) found that RSV activated SIRT1 deacetylase activity by decreasing the K M for both NAD + and the acetylated peptide. In addition to its effects on SIRT1, RSV was also shown to act on a wide range of enzymes, including COX-1, cAMP degrading phosphodiesterases, and nuclear factor-κB (NF-κB) (Jang et al., 1997;Manna et al., 2000;Pezzuto, 2011;Park et al., 2012). Consequently, there is debate whether the observed effects of RSV treatment are due to SIRT1 activation. For example, RSV has been shown to promote autophagy and many have attributed this effect to SIRT1 activation (Morselli et al., 2010;Wu et al., 2011). But this assertion was disputed by a study that observed the direct inhibition of mTOR, an inhibitor of autophagy, by RSV (Park et al., 2016). Despite the promiscuous nature of RSV, a study by Price et al. (2012) found that the presence of SIRT1 was necessary for RSVmediated mitochondrial biogenesis and AMPK activation.
After the discovery of RSV as a SIRT1 activator, the validity of the study and its use of the Fluor de Lys assay was disputed. As the assay uses a synthetic peptide substrate that contains a fluorophore, some argued that it was not physiologically relevant and could produce false positives. Kaeberlein et al. (2005) were able to replicate the activation of SIRT1 by RSV, but only when the fluorescent moiety was present on the peptide substrates. Other studies confirmed that the fluorophore was necessary for RSV-promoted activation of SIRT1 (Borra et al., 2005;Beher et al., 2009;Pacholec et al., 2010). In one such study, Pacholec et al. (2010) used NMR, SPR, and ITC to prove that the sirtuinactivating compounds (STACs) were directly interacting with the fluorophore attached to the peptide, even in the absence of SIRT1. Altogether, the contradictory results called into question the reliability of the fluorometric assay and raised serious doubts concerning RSV's mechanism of action and its ability to bind to and activate SIRT1 (Schmidt, 2010).
In response to the controversy, scientists at Sirtris, a company established after the initial discovery of SIRT1 activation by RSV, provided an explanation for the importance of the fluorescent moiety. Dai et al. (2010) observed the formation of STACsubstrate complexes, but found no correlation between the potency of a STAC and its affinity for the fluorescent TAMRA group. They further found that the ability of a STAC to activate SIRT1 was dependent on the substrate structure and were able to induce RSV-activation of SIRT1 using substrates composed of natural amino acids (Dai et al., 2010). This finding was also seen by Lakshminarasimhan et al. (2013) in which they observed RSVactivation of SIRT1 after replacing the fluorophore with a large, hydrophobic residue. Hubbard et al. (2013) further found that the fluorophore or its hydrophobic replacement has a positional requirement and were able to show that hydrophobic motifs within PGC-1α and FOXO3a could facilitate SIRT1 activation. Additionally, they also determined that the conserved residue Glu230 was critical for RSV activation and found that the benefits of RSV were attenuated when SIRT1 Glu230 mutants were expressed in myoblasts (Hubbard et al., 2013).
After enough evidence was presented confirming an interaction between RSV and SIRT1, the focus shifted to identifying how RSV interacts with SIRT1. When the association between RSV and SIRT1 was first reported, RSV was categorized as an allosteric modulator of the "K system" type (Monod et al., 1965;Howitz et al., 2003). For "K system" modulators, the K M value is affected while the V max remains the same. Identifying the region of SIRT1 where RSV binds was the first step in characterizing the interaction. SIRT1 is comprised of three major structured regions: an N-terminal domain (183-229), a catalytic domain (229-516), and a C-terminal regulatory region (641-665) (Davenport et al., 2014). The apo enzyme undergoes a conformational change following the binding of NAD + into a "closed" form that traps the substrate (Yuan and Marmorstein, 2012;Davenport et al., 2014). Using hydrogen deuterium exchange mass spectrometry (HDX-MS), Dai et al. (2015) were able to investigate the STAC binding domain (SBD) and determine that residues 183-229 were necessary for STAC activation of SIRT1. A crystal structure of a SIRT1/FdL peptide/RSV complex by Cao et al. (2015) indicated the binding of 3 RSV molecules for each SIRT1. Two of the RSV molecules form hydrogen bonds with the peptide and the SBD, thus bringing the two domains together in a "closed" conformation with improved substrate binding (Cao et al., 2015;Dai et al., 2018). The perspective of RSV acting as a stabilizing force between the substrate and SIRT1 has also been observed in a computational study in which RSV restored binding of "loose-binding" substrates (Hou et al., 2016).
Bioavailability is the primary obstacle in the development of RSV as a therapy. In human trials, Walle et al. (2004) observed that RSV had an absorption of ∼70%, but were unable to detect unmodified RSV 30 min after administration. After absorption, RSV is rapidly metabolized in the liver where it is conjugated to either sulfate or glucuronate (Springer and Moco, 2019). The RSV conjugates are produced by sulfotransferases (SULTs) and uridine-5 -diphospho-glucoronosyltransferases (UGTs) (Springer and Moco, 2019). Both SULTs and UGTs have genetic polymorphisms that can affect their ability to metabolize drugs, and thus can lead to high variability in levels of unmodified RSV between individuals (Ung and Nagar, 2007;Mehboob et al., 2017). The sulfate-conjugated metabolites, which are the primary conjugated form of RSV, were shown to have similar actions to RSV. Calamini et al. (2010) found that the sulfate conjugate of RSV was able to inhibit COX-1 and COX-2, and could activate SIRT1 only in the presence of a Fluor de Lys substrate. As the level of the conjugated forms of RSV surpass that of free RSV, it remains unclear which molecule is responsible for the experimentally observed effects of RSV supplementation.
Development of novel formulations of RSV emerged as its bioavailability became a prominent issue. One approach was to inhibit the enzymes responsible for conjugating RSV. Coadministration of RSV with quercetin was shown to decrease the formation of RSV-sulfate conjugates through the inhibition of SULT1A1, the major SULT isoform expressed in the liver and kidneys (De Santi et al., 2000). Multiple studies have observed the synergistic effect between RSV and quercetin from reducing adipose tissue weight following a high fat diet (Arias et al., 2016) to inhibiting the development of prostate cancer in a mouse model (Singh and Ahmad, 2015). In a similar fashion, Reen et al. (1993) combined RSV with piperine, an alkaloid previously shown to diminish the activity of UGTs in the intestines of rats. The study found that the addition of piperine to RSV resulted in a more than 1000% increase in the maximum serum concentration of RSV (Johnson et al., 2011). Despite the promising results seen in animal models, outcomes from human trials in healthy subjects were ambiguous (Table 1). Two trials co-administering RSV and quercetin found no improvement in RSV pharmacokinetics (la Porte et al., 2010;Huhn et al., 2018). Piperine supplementation was found to be useful in improving RSV bioavailability in one study examining cerebral blood flow (Wightman et al., 2014), but had no beneficial effect on serum levels of RSV in another study (Bailey et al., 2021).
Other groups tried to solve the bioavailability issue by focusing on the drug delivery method. The Fioretti lab developed a novel formulation of RSV as a solid dispersion on Magnesium dihydroxide microparticles. In their early studies, they found that the new formulation (later termed Revifast) was three times more soluble than the unmodified RSV and had enhanced bioavailability in rabbits (Spogli et al., 2018). In human trials, Revifast displayed an earlier peak in RSV as well as a twofold increase in free-RSV levels in plasma (Iannitti et al., 2020). Another method used to increase bioavailability of RSV is micronization. Sirtris developed SRT501, in which they reduced the particle size below 5 µm in order to improve solubility by increasing the surface area of RSV. In an early clinical trial, SRT501 was well tolerated in colorectal cancer (CRC) patients and had an improved C max compared to conventional RSV (Howells et al., 2011). Following that success, SRT501 was used as a treatment for patients with refractory or relapsed multiple myeloma. The clinical trial was terminated early due to severe side effects including nephrotoxicity (Popat et al., 2013). SRT501 was later discontinued for development as the company focused on other SIRT1-activating drug candidates.
RSV has been extensively studied in clinical trials. In the NIH clinical trial registry, over 150 trials using RSV are in various stages of completion. The trials concerned with sirtuin activation and the resulting therapeutic effects have been compiled in Tables 2-9. When summarizing the clinical outcomes of RSV (Figure 2), the issues with RSV are apparent. For most disease states RSV had a neutral effect, suggesting that the bioavailability issues are a major obstacle. Most of the trials in which RSV had a positive effect had higher dosages, typically over 500 mg/day, and the more recent trials are trending toward higher dosages as well. Although it is unclear if RSV can directly activate sirtuins when taken orally, it is likely that work with RSV will continue, as it is a readily available natural product with limited adverse reactions. Despite the early promise seen in the lab, the current reality of RSV has shown the need for developing better, more selective activators of sirtuins.

Resveratrol-Related Activators
Following the discovery of RSV, Sirtris began to develop new small molecule activators of SIRT1. From HTS, they discovered SIRT1 activators that were structurally distinct from RSV and had improved SIRT1 activating abilities (Milne et al., 2007). One of the compounds, SRT1720 (Figure 1), activated SIRT1 with an EC 1.5 = 2.9 µM and a maximum activation of approximately 4.5fold, whereas RSV had an EC 1.5 = 46.2 µM and a maximum activation of 2-fold (Milne et al., 2007). Using ITC, it was determined that the SRT STACs could only bind SIRT1 in the presence of the peptide substrate and they had a mechanism of action similar to RSV. It was further found that they used the same binding site as RSV. In a rodent model of insulin resistance, SRT1720 treatment resulted in a decrease in the blood glucose level and an increase in insulin sensitivity (Milne et al., 2007). Thus far, clinical trials of three compounds related to SRT1720 have been completed. SRT2104 (Figure 1) has the  greatest number of registered clinical trials (Table 10), but SRT2379 and SRT3025 (Figure 1) have also made it into the clinic ( Table 11). The single clinical trial to assess the safety and pharmacokinetics of SRT3025 was interrupted after the researchers found a prolongation effect of SRT3025 on the corrected QT interval, a warning that continuation could lead to  a potentially fatal proarrhythmia in the subjects (NCT01340911, GSK Study Register). Further development of SRT2379 was also terminated due to preclinical toxicities observed by the researchers (NCT01416376, GSK Study Register). For the lead compound, SRT2104, 5 out of the 8 clinical trials that focused on clinical outcomes had neutral or statistically insignificant results. A pharmacokinetic study found that the bioavailability of SRT2104 was 14% and exposure was improved when administered with food (Hoffmann et al., 2013). Most trials observed highly variable pharmacokinetics, leading some to have inconsistent clinical outcomes. A few trials observed beneficial effects of SRT2104 treatment on lipid profiles (Baksi et al., 2014),   (Krueger et al., 2015), and inflammation (van der Meer et al., 2015). McCallum et al. (2014) tried to improve the pharmacokinetics of SRT2104 by using different release formulations, but were unsuccessful. As of current, it appears that SRT2104 is no longer in development. Despite issues observed in clinical trials, SRT2104 continues to be used in studies as a SIRT1 activator (Miller et al., 2021).

EX-527 (SELISISTAT)
Before 2005, many small molecule regulators targeting sirtuins showed modest potency, isoform selectivity, and solubility (Bedalov et al., 2001;Grozinger et al., 2001;Bitterman et al., 2002;Mai et al., 2005;Olaharski et al., 2005;Porcu and Chiarugi, 2005). Isoform-selective small molecule probes were highly sought after for a better understanding of the biological functions of these enzymes (Porcu and Chiarugi, 2005). Napper et al. (2005) performed an HTS of a library of 280,000 compounds against human recombinant SIRT1 using a fluorometric assay in order to expand available SIRT1 small-molecule probes. The initial screening identified indole as a viable scaffold for SIRT1 targeting. The subsequent SAR study led to the discovery of EX-527 (or Selisistat) and the structurally related compound CHIC-35 (Figure 3) as selective SIRT1 inhibitors with IC 50 values of 0.098 and 0.063 µM, respectively (Napper et al., 2005). EX-527 exhibited 100-fold selectivity for SIRT1 over SIRT2/3, Class I/II HDACs, and NAD + glycohydrolase (Napper et al., 2005). Kinetic studies indicated that EX-527 is an uncompetitive inhibitor of SIRT1 regarding NAD + ; thus, inhibition depends on the concentration of NAD + (Napper et al., 2005;Gertz et al., 2013). Additionally, EX-527 exhibits exceptional ADME properties such as oral bioavailability, metabolic stability, and membrane penetrability, thus elevating EX-527 from a smallmolecule probe to a therapeutic candidate (Napper et al., 2005). In contrast, the study of CHIC-35 has been limited to pre-clinical investigations, primarily on its anti-inflammatory effects (Lugrin et al., 2013;Asad and Sachidanandan, 2020). SAR studies of EX-527 revealed that the primary carboxamide was necessary for an effective inhibition and modifications to the lead compound were generally not well tolerated (Napper et al., 2005). The isoform selectivity of EX-527 was initially attributed to possible differences among the sirtuin active sites  (Gertz et al., 2013;Broussy et al., 2020). A recent structural biology study suggested that EX-527 occupies the C-pocket (nicotinamide binding site) and a neighboring hydrophobic pocket which are highly conserved in the sirtuin family (Gertz et al., 2013). Further kinetic analysis indicated that the isoform selectivity of EX-527 stemmed from the differences in the kinetics of catalysis rather than any significant structural variation (Gertz et al., 2013;Broussy et al., 2020).
EX-527 has been explored as a potential therapeutic for Huntington's disease (HD). HD is a neurodegenerative disease characterized by abnormal movements, personality changes, and cognitive decline (Bates et al., 2015). A hallmark of HD is the expression of the mutant Huntington gene (mHTT), which has an expansion of a cytosine-adenine-guanine (CAG) repeat (Bates et al., 2015;Tabrizi et al., 2020). This extension leads to the protein misfolding and the formation of protein aggregates in HD patients (Bates et al., 2015;Tabrizi et al., 2020). As such, a potential therapeutic approach has focused on the degradation and removal of these aggregates. Acetylation of mHTT can direct the protein to autophagosomes for degradation (Jeong et al., 2009), thus facilitating the removal of the mutant protein. SIRT1 has been pursued as a therapeutic target for the treatment of HD because it has been shown to deacetylate mHTT to prevent its degradation. Genetic loss or pharmacological inhibition of Sir2 (the Drosophila melanogaster homolog of human SIRT1) was found to be neuroprotective for mHTT-challenged fruit flies (Pallos et al., 2008). Indeed, Smith et al. (2014) demonstrated that in a Drosophila model of HD, 10 µM of EX-527 could rescue neurodegeneration at a comparable level as the genetic elimination of Sir2. Additionally, in the R6/2 mouse model of HD, EX-527 can restore neural functions at 5 and 20 mg/kg dosages (Smith et al., 2014). It is important to note that the role of SIRT1 in HD remains controversial, as some view it as protective in HD modulation (Duan, 2013).
The aforementioned preclinical animal studies were essential for clinical trials involving EX-527 as an HD treatment option (Table 12). In a Phase 1 trial (NCT01521832), EX-527 was assessed for its safety in healthy human subjects. The study employed healthy men and women with two separate dosage regimens: a single dosage of 5, 25, 75, 150, 300, or 600 mg, and multiple dosages of 100, 200, or 300 mg/day (Westerberg et al., 2015). In this study, EX-527 was found to be welltolerated and safe after multiple doses of 300 mg/day and at a single dose of 600 mg. Based on the promising safety profile and dosing information, another clinical trial (NCT01485952) sought to investigate the feasibility of targeting SIRT1 as a potential treatment for HD (Sussmuth et al., 2015). In this study, human subjects with HD were treated with either 0, 10, or 100 mg/day of EX-527. There were improvements across the clinical, cognitive, and neuropsychiatric assessments from the baseline (day -1) to day 1 with no additional improvement at day 14 (Sussmuth et al., 2015).
A relatively new clinical application of EX-527 is improving in vitro fertilization (IVF) outcomes through the treatment of endometriosis (NCT04184323). Endometriosis is a chronic inflammatory reproductive disorder involving the growth of uterine endometrial cells outside of the uterine cavity (Zondervan et al., 2020). These delocalized endometrial growths can become lesions that lead to chronic localized pelvic pain and cramping with accompanying fertility issues. Standard treatment options for endometriosis involve mechanical lesion removal, hormonal therapy, or a combination of the two (Budinetz and Sanfilippo, 2010). However, reemergence of the lesions and complications associated with the hormonal therapy remain limitations. Thus, alternative treatment options are needed. A recent study demonstrated a KRAS activation-triggered SIRT1 overexpression in women with endometriosis, which has been suggested to contribute to infertility and the pathogenesis of endometriosis (Yoo et al., 2017). Targeting SIRT1 with small molecule inhibitors thus serves as a potential therapeutic treatment for endometriosis-mediated IVF failure. A planned clinical trial (NCT04184323) will seek to explore inhibition of SIRT1 by EX-527 as a possible treatment for the inflammation associated with endometriosis.

QUERCETIN
Quercetin (Figure 4) is a flavonoid phytoestrogen which has demonstrated activity in the management of brain, blood, salivary gland and uterine cancers (Dolatabadi, 2011;Sak, 2014;Sinaga et al., 2017), as well as viral infections such as HCV (Rojas et al., 2016) and Zika virus (Wong et al., 2017) and bacterial infections (Wang et al., 2018;Zeng et al., 2019;Oktyabrsky et al., 2020) in both in vivo and in vitro studies. The structure of quercetin consists of three rings and five hydroxyl groups at the 3, 5, 7, 3 and 4 -positions of the basic flavanol skeleton (Figure 4). The name "quercetin" was derived from the Latin word "Quercetum" which means oak forest. It is one of the most abundant flavonoids found in fruits and vegetables, and was discovered alongside other bioflavonoids by Albert Szent Gyorgyi in 1936 (Moskaug, Carlsen et al., 2004).
In the same HTS that identified RSV, quercetin was observed to increase SIRT1 activity by fivefold (Howitz et al., 2003). Studies have shown the inhibitory role of quercetin on the progression of breast, colon, prostate, and lung cancers (Baghel et al., 2012;Smith et al., 2016). Quercetin has been found to alleviate kidney fibrosis, intervertebral disc degeneration, and diabetic encephalopathy via activation of SIRT1-mediated pathways (Dong et al., 2014;Hu et al., 2020;Liu et al., 2020). Quercetin treatment of Herpes simplex virus-1 infected neuronal cell lines increased the survival of the cells by inhibiting viral production and improved neurodegenerative markers via SIRT1 activation (Leyton et al., 2015). It was also found that quercetin inhibits oxidative injury in human endothelial cells through SIRT1 activation, leading to the upregulation of the SIRT1/AMPK pathway . In addition, quercetin regulates oxidative stress in the body by directly reducing the level of reactive oxygen species (Oboh et al., 2016). In myocardial/ischemia-reperfusion (MI/R) injury in rats, quercetin supplementation increased the expression levels of SIRT1 and PGC-1α, leading to the activation of the SIRT1/PGC-1α pathway and subsequent reduction in MI/R-induced myocardial damage (Tang et al., 2019). In addition to SIRT1, quercetin has a mild stimulating effect on SIRT6 (You et al., 2019). When modified with a bulky trihydroxy benzoyl group at the 3-OH group, as in catechin gallate, it inhibits SIRT6 activity (Rahnasto-Rilla et al., 2018;You et al., 2019).
Quercetin exists as a glycone or an aglycone in plants. When ingested, the glycone form can be hydrolyzed to the aglycone form that can be easily absorbed in the small intestine due to its hydrophobic nature (Massi et al., 2017). In human plasma, ingested quercetin glycosides are predominantly metabolized into quercetin 3-O-β-D-glucuronide and quercetin 3 -O-sulfate (Figure 4; D'Andrea, 2015; Moodi et al., 2021). Modifications such as glycosylation and methylation of the quercetin scaffold result in derivatives with distinct biological activities (Lesjak et al., 2018). For example, isoquercetin, the 3-O-glucoside of quercetin (Magar and Sohng, 2020), demonstrates FIGURE 4 | Chemical structures of quercetin and its derivatives.
Quercetin's therapeutic applications have been limited by its low bioavailability, poor solubility, and short half-life (Gugler et al., 1975;Ferry et al., 1996;KaŞıkcı and Bagdatlıoglu, 2016). Modifications have been made to improve these properties (Massi et al., 2017). In one study, the bioavailability of quercetin was increased by about 20 times through a phytosome delivery system (Riva et al., 2019). Several other studies have utilized conjugation to various amino acids and nanoparticle delivery systems to improve the bioavailability of quercetin. Acylated quercetin analogs synthesized by Duan et al. (2017) were about 10-fold more soluble in water than quercetin. Of the several quercetin human clinical trials targeting various disease states, only one ongoing study (NCT03943459) is aimed at investigating the activation of SIRT1 by quercetin in coronary disease.

CONCLUSION AND PERSPECTIVES
A simple search in PubMed provides hundreds of publications related to sirtuin inhibitor/activator development, demonstrating the critical roles these enzymes play in regulating diverse cellular events and the intense interest in pursuing them as therapeutic targets. Unfortunately, tremendous efforts have only resulted in a handful of small molecules in clinical studies as described in this review article. Translating sirtuin regulators from the bench to the clinics has been hampered by the lack of isoformselective candidate compounds with favorable pharmacological profiles. The catalytic domain is highly conserved between sirtuins and therefore represents a promiscuous target for NAD + or peptide-competitive inhibitors (Dai et al., 2018). In the case of activators, the binding sites are often not readily apparent by the inspection of a crystal structure, and there is no general and systematic strategy to identify and target these sites. Furthermore, for several human sirtuin isoforms, novel enzymatic activities were discovered recently (Du et al., 2011;Feldman et al., 2013;Jiang et al., 2013;Anderson et al., 2017). Potent regulators targeting specifically these new activities are still in the making. Pre-clinical investigations using animal models may differ in the genetic background or the assessment methods which have caused controversies and ambiguities that still need to be reconciled. In spite of numerous reports on the endogenous substrates of sirtuins and the pathways they regulate, our understanding of the biological functions of sirtuins is still in its infancy. For example, SIRT1 has been closely associated with cancer pathology, and has been suggested as either a tumor promoter or a suppressor in a context-dependent manner (Deng, 2009). All the research effort has only scratched the surface of sirtuin biology. A comprehensive and thorough picture of these intriguing enzymes still awaits description.
Of course, sirtuin-targeting drugs still hold great therapeutic potential, and progress in the field will accelerate the development of small molecule drug candidates. Apart from their highly conserved catalytic core, sirtuins harbor structurally unique N-or C-terminal extensions that can be targeted for selectivity or even specificity. The conformational plasticity of the active site that explains the isoform selectivity of EX-527 (Gertz et al., 2013;Broussy et al., 2020) has also been suggested as a novel target for inhibitor development. The clinical success of sirtuin-targeting medications requires a clear understanding of the "sirtuin-dependency" of the disease, robust lead compounds that are potent and selective with ideal drug-like properties, PK/PD profiling and improvement, as well as advances in formulation. The combined efforts from all these aspects will bring more sirtuin regulators into the clinic for treating diseases with considerable unmet medical needs.

AUTHOR CONTRIBUTIONS
AC, DW, and YC: conceptualization. AC, DW, DD, and YC: writing. YC: project administration and funding acquisition. All authors contributed to the article and approved the submitted version.

FUNDING
This work was supported in part by CHE-1846785 from NSF (to YC), 2020 VCU CCTR Endowment Fund (subaward of UL1TR002649 from National Center for Advancing Translational Sciences to VCU) (to YC), Jeffress Trust Award in Interdisciplinary Research from Thomas F. and Kate Miller Jeffress Memorial Trust and Bank of America (to YC), 2020 Presidential Research Quest Fund from VCU (to YC), and startup funds from VCU (to YC).