A medicinal chemistry perspective on SIRT3 modulators in disease treatment

Abstract

As the primary mitochondrial deacetylase, SIRT3 plays a central role in key biological processes such as cellular energy metabolism, oxidative stress response, and apoptosis regulation. It is also closely associated with the development and progression of neurodegenerative diseases, cardiovascular disorders, and various cancers. With advancing research on its biological functions, SIRT3 has emerged as a promising therapeutic target, leading to the discovery of multiple types of SIRT3 modulators. From a medicinal chemistry perspective, this review systematically summarizes the progress, challenges, and opportunities in the development of SIRT3-targeting drugs. We comprehensively analyze the structure–activity relationships, mechanisms of action, and in vitro/in vivo activities of SIRT3 modulators. Furthermore, this review discusses future research directions, including the design of highly selective modulators, organelle-specific delivery systems, and protease-mediated SIRT3 degradation, offering theoretical and experimental foundations for therapeutic strategies targeting SIRT3-related diseases.

This review analyzes the structure–activity relationships, mechanisms, and activity of SIRT3 modulators, and discusses future directions like selective design, targeted delivery, and protease-mediated degradation for treating SIRT3-related diseases.

1. Introduction

1.1. Sirtuin family members

Sirtuins (SIRTs) are class III histone deacetylases dependent on nicotinamide adenine dinucleotide (NAD⁺), comprising seven family members (SIRT1–7).¹ All SIRTs share a conserved catalytic core domain, which is flanked by member-specific N-terminal and C-terminal extensions. Due to variations in these terminal regions, different SIRT members exhibit distinct subcellular localizations.² SIRT1, SIRT6, and SIRT7 are primarily nuclear; SIRT3, SIRT4, and SIRT5 are mainly mitochondrial; whereas SIRT2 is predominantly cytoplasmic.³ Based on phylogenetic classification of their conserved catalytic domains, SIRTs can be divided into four distinct classes. Class I includes SIRT1–3, which exhibit strong deacetylase activity; class II consists solely of SIRT4, which has weak deacetylase activity but demonstrates notable ADP-ribosyltransferase activity; class III contains only SIRT5, which shows weak deacetylase activity and prefers to remove negatively charged acyl groups, displaying strong deglutarylase, desuccinylase, and demalonylase activities; class IV comprises SIRT6 and SIRT7, both possessing deacetylase and ADP-ribosyltransferase activities; in addition, SIRT6 also exhibits deacylase activity toward long-chain fatty acyl groups.^4–7 Different SIRT family members perform varied functions in genome maintenance, metabolic regulation, and stress response modulation, playing crucial roles in maintaining cellular homeostasis and influencing the initiation and progression of diseases.^8,9

1.2. Distribution, structure and function of SIRT3

Sirtuin 3 (SIRT3) serves as the primary mitochondrial deacetylase and exhibits important functions in regulating cellular energy metabolism and stress responses.¹⁰ Initially identified in humans by Feinberg et al. in 2002, the SIRT3 gene is located on chromosome 11p15.5.^11–14 The full-length SIRT3 protein, with a molecular weight of approximately 44 kDa and also referred to as the long form, is found in the nucleus and cytoplasm but lacks catalytic activity in this state. Upon translocation into mitochondria, its mitochondrial targeting sequence is cleaved by mitochondrial processing peptidase, converting it into a shorter, active deacetylase known as the short form, which has a molecular weight of around 28 kDa. Consequently, SIRT3 is widely expressed in tissues rich in mitochondria, such as the kidney, heart, brain, and liver.^3,10,15

SIRT3 is a 399-amino acid protein consisting of a large N-terminal domain containing a mitochondrial targeting sequence and a smaller C-terminal domain. The N-terminal domain adopts a Rossmann fold responsible for NAD⁺ binding, while the C-terminal domain comprises a helical module and a zinc-binding motif formed by an extended loop from the N-terminal domain. The N-terminal region can be divided into three sub-sites: the adenine–ribose binding pocket, the nicotinamide–ribose binding site, and a deep catalytic center.^15–17 The N- and C-terminal domains are connected by four loops, forming a cleft where the acetylated protein substrate binds.¹⁶ In 2009, Perni et al. resolved the crystal structure of SIRT3 in complex with acetyl-CoA synthetase 2 containing an acetyl-lysine residue (AceCS2-K_ac; PDB: 3GLR).¹⁸ The K_ac side chain of AceCS2 is stabilized within SIRT3 by residues F294, V324, H248, and F180, and engages in hydrogen bonds or van der Waals interactions with residues including G295, L298, E323, E325, and V292. The carbonyl oxygen of the acetyl group is oriented toward the NAD⁺ binding site (Fig. 1).¹⁹

Fig. 1. (A) and (B) Cocrystal structure of AceCS2-K_ac bound to SIRT3 (PDB 3GLR); (C) mechanism of deacetylation exerted by SIRT3.

The deacetylation reaction catalyzed by SIRT3 (Fig. 1) begins with the binding of an acetylated substrate to the enzyme's catalytic site. The association of AceCS2-K_ac induces the formation of an antiparallel β-sheet in SIRT3, narrowing the interdomain distance and facilitating the precise positioning of NAD⁺ within the conserved binding pocket. The amide group of the substrate interacts with an aspartate residue. Upon substrate binding, a conformational change occurs in the flexible loop of SIRT3.^15,17 The carbonyl oxygen of AceCS2-K_ac carries out a nucleophilic attack on the C1 carbon of the NAD⁺ ribose, leading to cleavage of the nicotinamide–ribosyl bond and transfer of the acetyl group to ADP-ribose, forming an alkylamidate intermediate while releasing nicotinamide. The flexible loop undergoes uncoiling and folds downward to shield this intermediate, while V366 rotates to encapsulate the adenine moiety of ADP-ribose. H224 promotes the attack of the ribose 2′-OH group on the imidate carbon, converting the alkylamidate intermediate into a bicyclic species. This bicyclic intermediate is subsequently hydrolyzed by a water molecule, yielding deacetylated AceCS2 and the byproduct 2′-O-acetyl-ADP-ribose.^15,17 Following deacetylation, the flexible loop returns to its original conformation.¹⁸

SIRT3, a mitochondria-localized deacetylase, plays a pivotal role in multiple key biological processes. In the regulation of cellular oxidative stress, SIRT3 enhances the scavenging capacity of reactive oxygen species (ROS) by increasing the activity of antioxidant enzymes such as superoxide dismutase 2 (SOD2), and promotes the nuclear translocation of forkhead box O3a (FOXO3a), thereby strengthening the cellular antioxidant defense system.^20,21 In the control of apoptosis, SIRT3 acts through several signaling pathways: it modulates the glycogen synthase kinase-3 beta (GSK-3β)/Bax, Bax/Bcl-2, and Bad/Bcl-xL ratios, and suppresses the opening of the mitochondrial permeability transition pore by deacetylating cyclophilin D, thus inhibiting mitochondria-mediated apoptosis.^22,23 With regard to energy metabolism, SIRT3 activates various metabolic enzymes including long-chain acyl-CoA dehydrogenase (LCAD), pyruvate dehydrogenase E1 alpha 1 subunit (PDHA1), and 3-hydroxy-3-methylglutaryl-CoA synthase 2 (HMGCS2), thereby promoting the tricarboxylic acid cycle, urea cycle, and fatty acid oxidation.^24–27 In inflammatory responses, SIRT3 significantly mitigates inflammation by maintaining ROS homeostasis, inhibiting NOD-, LRR- and pyrin domain-containing protein 3 (NLRP3) inflammasome activation, and regulating the NF-κB signaling pathway.²⁸

1.3. Role of SIRT3 in progression of diseases

1.3.1. Role of SIRT3 in non-oncological diseases

SIRT3 is highly expressed in the brain and nervous system under physiological conditions; however, its expression declines markedly during aging and in neurodegenerative disorders, representing a significant risk factor for disease pathogenesis. The protective effects of SIRT3 are mediated primarily through the inhibition of neuronal apoptosis. In a mouse model of Huntington's disease (HD), upregulation of SIRT3 reduced striatal neuronal degeneration, enhanced neuronal survival, and improved motor function.²⁹ The underlying mechanisms include: deacetylation of p53, preventing its mitochondrial translocation and subsequent cytochrome C release; deacetylation of cyclophilin D to suppress mitochondrial permeability transition, thereby preserving mitochondrial function.^30,31 Furthermore, SIRT3 binds to ATP synthase peripheral stalk subunit OSCP (ATP5O) to boost electron transport chain activity and increase ATP production, activates FOXO3a to promote the expression of autophagy-related genes and suppress Bax-mediated apoptosis, and enhances mitophagy via the PINK1–Parkin pathway, collectively contributing to the inhibition of cell death.^32,33

SIRT3 plays an important protective role in cardiac diseases such as heart failure and cardiac hypertrophy. SIRT3 knockout mice exhibit microcirculatory dysfunction and impaired diastolic function, indicating that loss of SIRT3 may lead to microvascular rarefaction and promote the progression of heart failure.³⁴ In a hypertensive heart failure model, SIRT3 expression is downregulated, accompanied by mitochondrial protein hyperacetylation. Reduced SIRT3 levels in failing myocardium result in increased acetylation and decreased activity of pyruvate dehydrogenase (PDH) and ATP synthase.^35,36 SIRT3 enhances antioxidant defense by deacetylating FOXO3a and promoting its nuclear translocation, thereby upregulating antioxidant enzymes and reducing ROS levels; it also directly activates SOD2.³⁷ In diabetic cardiomyopathy, SIRT3 suppresses TP53-induced glycolysis and apoptosis regulator (TIGAR) expression via p53 deacetylation and indirectly upregulates 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase 3 (PFKFB3) expression, collectively increasing fructose-2,6-bisphosphate levels and enhancing glycolysis in response to hyperglycemia.^38,39 Endothelial SIRT3 deficiency impairs PFKFB3-mediated glycolysis, increases oxygen consumption and ROS production, and ultimately leads to myocardial hypoxia and apoptosis.³⁶

1.3.2. Role of SIRT3 in cancer

SIRT3 plays a dual role in tumorigenesis, with its specific function depending on the genetic background and characteristics of the tumor microenvironment.⁴⁰ In certain cancers that rely primarily on oxidative phosphorylation (OXPHOS) for energy, such as head and neck squamous cell carcinoma and diffuse large B-cell lymphoma, SIRT3 exhibits oncogenic activity: it inhibits apoptosis through antioxidant effects or activates glutamate dehydrogenase (GDH) to enhance the tricarboxylic acid cycle, thereby supporting tumor growth.^41,42 In multiple myeloma, SIRT3 also promotes tumor progression by deacetylating isocitrate dehydrogenase 2 (IDH2).⁴³ Conversely, SIRT3 expression is frequently downregulated in various cancers including pancreatic, colorectal, and clear cell renal cell carcinomas.^44,45 Its tumor-suppressive mechanisms primarily involve the deacetylation of hypoxia-inducible factor-1α (HIF-1α), promoting its prolyl hydroxylase-mediated degradation, thereby disrupting the Warburg effect and reversing glycolytic dominance in cancer cells.³⁹ Additionally, SIRT3 inhibits the malate–aspartate shuttle by deacetylating glutamic–oxaloacetic transaminase 2 (GOT2), restricting glycolytic flux; it also inactivates cyclophilin D, reducing glucose uptake and the production of lactate, pyruvate, and acetyl-CoA, ultimately impairing tumor metabolic adaptation.^46,47 In clear cell renal cell carcinoma, SIRT3 further enhances mitochondrial function, induces metabolic reprogramming, decreases glucose uptake, and increases sensitivity to chemotherapeutic agents.⁴⁸

1.4. Current status in the development of SIRT3-targeting drugs

As a member of the sirtuin family, the biological functions of SIRT3 were not fully understood in the early stages of research, which limited progress in the development of SIRT3-targeting drugs. In recent years, advances in understanding SIRT3's deacetylase activity and crystallographic features have promoted the discovery of both small-molecule inhibitors and activators of SIRT3. SIRT3 inhibitors, including peptide-based and small-molecule compounds, have demonstrated promising efficacy in various cancers such as multiple myeloma (MM) and head and neck squamous cell carcinoma (HNSCC). Meanwhile, SIRT3 activators, comprising synthetic small molecules, natural product derivatives, and peptide-based activators, have shown beneficial effects in models of cancer, heart failure, and Parkinson's disease. However, compared to drugs targeting acetyltransferases and other deacetylases, no SIRT3-targeting molecule has yet entered clinical trials.

The development of SIRT3-targeting drugs faces several technical challenges. Firstly, the physiological and pathological functions of SIRT3 are complex and highly dependent on cellular context and disease stage. For example, in triple-negative breast cancer, SIRT3 acts as a tumor suppressor, where its activation can induce apoptosis and inhibit proliferation; however, in cancers such as diffuse large B-cell lymphoma (DLBCL) and acute myeloid leukemia (AML), SIRT3 promotes tumor growth and survival.⁴⁹ This two-faced nature necessitates careful consideration when deciding whether to activate or inhibit SIRT3, thereby complicating drug development. Secondly, the high structural similarity among sirtuin family members makes it difficult to design compounds that selectively target SIRT3 without affecting other isoforms, raising concerns about off-target effects and potential toxicity. Finally, as SIRT3 is located in the mitochondrial matrix, drugs must be able to specifically accumulate within mitochondria to exert their effects, adding another layer of complexity to drug design.

This review aims to summarize recent advances in the development of SIRT3-targeting inhibitors and activators, with a primary focus on their structure–activity relationships, thereby providing a foundation for future drug discovery efforts.

2. SIRT3 inhibitors

2.1. Peptide-based inhibitors

Research on peptide-based SIRT3 inhibitors remains largely in the early stages. Since 2006, Zheng and colleagues have conducted extensive studies on N^ε-thioacetyl-lysine peptides, leading to the development of several sirtuin-targeted inhibitors (Fig. 2).⁵⁰ Starting from a pan-SIRT1/2/3 tripeptide inhibitor containing N^ε-thioacetyl-lysine, the team designed and synthesized a series of derivatives with varied side chains at the −1 and +1 positions. Among these, compounds 3, 4, and 7 exhibited significantly enhanced selectivity toward SIRT3 while maintaining inhibitory potency, whereas compounds 5 and 6 showed strong inhibitory efficacy against SIRT3.⁵¹ Cyclization of compounds 3, 4, and 7 yielded derivatives 3′, 4′, and 7′, which led to a notable reversal in selectivity among SIRT1/2/3.⁵² However, the inherent limitations of natural peptides, such as poor chemical and physical stability, as well as short plasma half-life, must be addressed through conventional design or emerging technologies before they can be developed into viable therapeutics.

Fig. 2. Structures, biological activity data, and development paths of peptide-based inhibitors.

2.2. Small-molecule inhibitors

2.2.1. Nicotinamide derivatives

Nicotinamide acts as an endogenous negative regulator of sirtuins. As a product of the sirtuin-catalyzed deacetylation reaction, it reduces sirtuin activity by competitively inhibiting NAD⁺ binding, and its interaction with sirtuin can also promote the reverse reaction of deacetylation.⁵³ In 2012, the Genazzani team systematically modified the structure of nicotinamide (Fig. 3).⁵⁴ For example, introducing a methyl group at the 4-position of the pyridine ring yielded compound 9, which exhibited a 16-fold increase in inhibitory activity against SIRT3 (IC₅₀ = 23 μM). Replacing the amide bond in nicotinamide with its bioisostere 1,2,3-triazole resulted in compound 8, showing a 10-fold enhancement in SIRT3 inhibition (IC₅₀ = 38 μM). Compound 8 was the first reported selective SIRT3 inhibitor, showing no significant inhibition against SIRT1 or SIRT2 at 1 mM. Biological assays demonstrated that this compound could effectively inhibit the proliferation of various tumor cells, including HeLa, SK-MEL-28, and CCRF-CEM. At 1 mM, compound 8 significantly increased the acetylation levels of mitochondrial proteins in HeLa cells; at 100 μM, it also markedly reduced ATP content and elevated ROS levels in SK-MEL-28 cells. However, due to its relatively weak activity, compound 8 not only fails to specifically target SIRT3 but also inhibits other NAD⁺-dependent enzymes. As a result, such endogenous small molecules and their derivatives may cause multiple off-target effects, and their precise mechanisms of action in vivo require further in-depth investigation.

Fig. 3. Structures, biological activity data, and development paths of compounds 8–9.

2.2.2. Cambinol analogs

Cambinol, identified through phenotypic screening of a compound library, is a non-selective inhibitor of SIRT1/2 with IC₅₀ values of 56 μM and 51 μM against SIRT1 and SIRT2, respectively, while exhibiting only weak inhibition of SIRT3 (IC₅₀ > 200 μM). In 2014, Simon and colleagues conducted systematic structural modifications and structure–activity relationship studies on cambinol (Fig. 4).⁵⁵ Their research revealed that introducing specific functional groups at different positions could significantly alter inhibitory selectivity among sirtuin isoforms. For example, replacing the pyrimidinone moiety with a pyrazolone, together with introducing a methyl group at the 4-position of the phenyl ring and a methoxy group at the γ-position of the naphthalene ring, yielded compound 10, which showed selective inhibition toward SIRT1 (>7.4-fold selectivity). Further introduction of a phenyl group onto the naphthalene ring increased molecular rigidity, resulting in compound 11, which exhibited selective inhibition of SIRT3 (>5.3-fold selectivity). In contrast, replacing the pyrimidinone with an azacycle and introducing a bromine atom on the naphthalene ring led to compound 12, which demonstrated selective inhibition of SIRT2.

Fig. 4. Structures, biological activity data, and development paths of compounds 10–12.

2.2.3. Purine-2,6-dione derivatives

In 2020, Liu et al. identified a novel class of purine-2,6-dione-based sirtuin inhibitors through fluorescence-based screening of an in-house compound library (Fig. 5).⁵⁶ Among them, compound 13 exhibited potent inhibitory activity against SIRT3 (IC₅₀ = 0.79 μM). Structural modification revealed that replacing the ethyl group at the R1 position with a phenethyl group (compound 14) further enhanced the inhibitory activity, yielding an IC₅₀ of 0.37 μM. Molecular docking simulations indicated that the R1 substituent of compound 14 is surrounded by multiple hydrophobic amino acid residues, and the introduction of the phenethyl group enhances molecular hydrophobicity, thereby improving potency. Compound 14 demonstrated moderate binding affinity for SIRT3 (K_d = 19.9 μM) and low cytotoxicity at 20 μM. However, this compound acts as a pan-inhibitor, showing IC₅₀ values of 0.17 μM, 1.35 μM, and 0.45 μM against SIRT1, SIRT2, and SIRT5, respectively.

Fig. 5. Structures, biological activity data, and development paths of compounds 13–14.

2.2.4. Thieno[3,2-d]pyrimidine derivatives

In 2013, Perni et al. conducted an DNA-encoded library-based screen to identify novel SIRT3 modulators (Fig. 6).⁵⁷ Four compounds were initially identified, among which compound 15 exhibited the strongest SIRT3 inhibition, with IC₅₀ values of 3.6, 2.7, and 4.0 nM against SIRT1, SIRT2, and SIRT3, respectively. To improve its physicochemical properties, the team systematically truncated the structure while retaining the thienopyrimidine-6-carboxamide core. Among the truncated analogs, compound 16, in which the thiophene moiety was replaced by an acetamide group, achieved an optimal balance between inhibitory activity against SIRT1/2/3 (IC₅₀ = 23–110 nM) and molecular weight (M_W = 347). Further optimization led to compounds 17 and 18, where replacing the acetamide with tert-butylamide and sulfonamide groups increased potency by 1.7- and 7.8-fold, yielding IC₅₀ values of 33 nM and 7.2 nM against SIRT3, respectively. The crystal structure of SIRT3 in complex with 15 (PDB: 4JSR) revealed that the compound occupies both the nicotinamide-binding pocket and the acetyl-lysine substrate-binding site (Fig. 6). The 6-carboxamide group forms hydrogen bonds with I230 and D231, and via water molecules, with A146 and I154. The thienopyrimidine core engages in π–π stacking with F157, while its N1 atom hydrogen-bonds with F157; the N3 atom is solvent-exposed, interacting only with water. The ethylpiperidine group binds in a hydrophobic pocket formed by Y165, F180, I230, I291, and F294, with its amide oriented toward the acetyl-lysine site. The terminal acetamide of the 2-thiophene moiety hydrogen-bonds with E296 and extends toward the protein surface. Compounds 15 and 17 showed good selectivity profiles against kinases, nuclear receptors, ion channels, transporters, and GPCRs, with no significant inhibition of hERG or CYP enzymes. Compound 17 exhibited favorable properties including low log D (2.73), good solubility (297 μM), and stability in liver microsomes. Although 15 remained the most potent SIRT3 inhibitor (IC₅₀ = 4 nM) in the series, it lacked selectivity among sirtuin isoforms.

Fig. 6. (A) Structures, biological activity data, and development paths of compounds 15–21; (B) structure of compound 15, and cocrystal structure of compound 15 bound to SIRT3 (PDB 4JSR).

To investigate the lack of selectivity of compound 15, Yao et al. performed structural superposition analysis of the SIRT1–3 subtypes. The results indicated that the amino acid residues forming hydrogen bonds with 15 are highly conserved among SIRT1–3. In addition, the flexibility of the compound's side chain induced a conformational fit upon binding to different sirtuin isoforms, contributing to the poor selectivity. Consequently, the team retained the thienopyrimidine-6-carboxamide core and introduced a piperazine group as a linker (Fig. 6A).⁵⁸ At a concentration of 1.0 μM, compound 19 with a 3,5-dichloro substitution at the terminal end maintained over 50% inhibition. However, as the piperazine group did not fully occupy the catalytic cavity, a larger bridged piperazine scaffold (20) was used to replace the piperazine moiety. Compared to 19, compound 20 showed reduced activity, exhibiting only 39.1% inhibition at 1 μM. This decrease in potency may be attributed to conformational distortion caused by ring strain, which altered the orientation of the side chain. The team then employed a bulkier bridged piperazine structure with lower ring strain. The 4-trifluoromethyl-substituted derivative 21 demonstrated the strongest SIRT3 inhibition in the series, with an IC₅₀ of 43 nM, and exhibited approximately 9-fold and 11-fold selectivity over SIRT1 and SIRT2, respectively. Compound 21 directly interacted with SIRT3 in tumor cells, significantly increasing the acetylation level of SOD2, elevating ROS levels, and reducing ATP production. It also upregulated Bax expression, thereby inducing apoptosis and markedly inhibiting the growth of various tumor cell lines, for example, showing an IC₅₀ of 0.32 μM in MOLM13 cells. Furthermore, in a mouse xenograft model of AML, intraperitoneal administration of 21 at 5 mg kg⁻¹ slowed disease progression without significantly affecting body weight, indicating a favorable safety profile.

2.2.5. Thiomyristoyl lysine derivatives

In recent years, Lin et al. have developed a series of sirtuin inhibitors containing a thiomyristoyl group (Fig. 7). In 2016, they reported the SIRT2-selective inhibitor TM, which exhibited an IC₅₀ of 0.028 μM against SIRT2 with no significant inhibition of SIRT1 or SIRT3.⁵⁹ Subsequent structural modifications were made to TM to improve its properties. To enhance solubility and activity, a trimethylammonium group was introduced at the carboxyl terminus, yielding the SIRT1/2/3 inhibitor NH4-6 (SIRT3 IC₅₀ = 2.3 μM).^60,61NH4-6 displayed good aqueous solubility (clog P = 3.86) and better cellular permeability than TM at 50 μM. Introduction of a hydroxyl group at the meta-position of the phenyl ring in TM afforded JH-T4 (SIRT3 IC₅₀ = 2.5 μM), which, however, showed toxicity toward normal cells due to its inhibitory activity to multiple sirtuin isoforms.⁶² To improve mitochondrial targeting and penetration, Lin et al. replaced the carbobenzoxy group in JH-T4 with a triphenylphosphonium (TPP) mitochondrial-targeting moiety, resulting in YC8-02.⁴² Compared to JH-T4, YC8-02 demonstrated significantly enhanced mitochondrial accumulation. It also exhibited superior SIRT3 inhibitory activity (IC₅₀ = 0.53 μM) and potently suppressed the survival and proliferation of DLBCL cell lines, while showing negligible effects on human solid tumor cell lines or cord blood cells. In vivo, intraperitoneal administration of YC8-02 at 30 mg kg⁻¹ induced tumor regression. However, the high hydrophobicity of the TPP group resulted in poor aqueous solubility of YC8-02, limiting its bioavailability. Additionally, there remains room to further optimize the SIRT3 inhibitory potency of YC8-02 for improved therapeutic efficacy.

Fig. 7. (A) Structures, biological activity data, and development paths of thiomyristoyl lysine derivatives as SIRT3 inhibitors; (B) structure of NH6-10, and cocrystal structure of NH6-10 bound to SIRT3 (PDB 9CBT).

In 2024, Lin et al. performed structural optimization on YC8-02 to improve its solubility and potency, leading to the development of SJ-106C, a SIRT3 inhibitor featuring a triethylammonium-based mitochondrial-targeting group.⁶³SJ-106C exhibited an IC₅₀ value of 0.49 μM against SIRT3, representing an approximately 3-fold improvement over YC8-02. With a clog P of 5.9 compared to 11.7 for YC8-02, SJ-106C also demonstrated significantly enhanced solubility. The researchers resolved the co-crystal structure of SIRT3 in complex with NH6-10 (IC₅₀ = 2.8 μM; PDB: 9CBT), revealing that NH6-10 forms a covalent adduct with NAD⁺ (Fig. 7). Electrostatic interactions were observed between E325 and the positively charged triethylammonium moiety at the C-terminus of NH6-10. In cellular assays, SJ-106C showed IC₅₀ values of 0.65 μM and 0.45 μM against Karpas 422 and OCI-LY7 cells, respectively. It exhibited superior mitochondrial accumulation compared to YC8-02 and JH-T4, and at concentrations of 2.5 μM and 5 μM, it increased the acetylation levels of mitochondrial proteins. SJ-106C demonstrated good bioavailability in mice. Intraperitoneal administration at doses of 50 mg kg⁻¹ or 100 mg kg⁻¹ significantly inhibited the growth of OCI-LY7 xenograft tumors, with no notable effect on mouse body weight.

2.2.6. Quinoline derivatives

In 2022, Zhang et al. designed and synthesized a series of compounds to identify potent SIRT3 inhibitors (Fig. 8).⁶⁴ They introduced a 4-acrylamidophenyl-quinoline scaffold and performed structural modifications by incorporating various substituents at the 4-carboxyl position of the quinoline ring. The introduction of an ortho-fluoro group on the phenyl ring (compound P6) or increasing the steric bulk of the phenyl substituent (compound P19) led to enhanced inhibitory activity. Among them, P6, containing a 2-fluorophenyl group, exhibited the strongest activity, showing 65.15% inhibition of SIRT3 at a concentration of 10 μM. The IC₅₀ value of P6 against SIRT3 was determined to be 7.2 μM, while it showed weaker inhibition of SIRT1 (IC₅₀ = 32.6 μM) and SIRT2 (IC₅₀ = 33.5 μM). In cellular assays, P6 induced G0/G1 phase cell cycle arrest in MLL-rearranged leukemia cells and effectively suppressed the growth of THP-1 and MOLM-13 cell lines. Mechanistic studies indicated that the anticancer effect of P6 does not rely on apoptosis induction but rather operates through tumor phenotype elimination and induction of cellular differentiation. However, the acrylamide group of P6 did not exhibit significant interactions within the SIRT3 active site, and its nature as a Michael acceptor raises concerns about potential off-target effects.

Fig. 8. Structures, biological activity data, and development paths of quinoline derivatives as SIRT3 inhibitors.

In 2024, Zhang et al. performed structural optimization on compound P6 (Fig. 8).⁶⁵ Replacement of the acrylamide group in P6 with an imidazole moiety yielded S27, which exhibited a 1.4-fold improvement in SIRT3 inhibition (IC₅₀ = 5.06 μM) and showed significant anti-proliferative activity against tumor cells. Based on this result, the team further modified the structure of S27. Introduction of a methyl group at the 6-position of the quinoline ring led to compound T5, which demonstrated enhanced activity with a 1.8-fold increase in potency compared to S27 (IC₅₀ = 2.88 μM). Both S27 and T5 significantly increased the acetylation levels of histone H4 and p53. They also exhibited notable anti-proliferative effects against MM1.S and RPMI-8226 cell lines. At concentrations of 5 μM or 20 μM, S27 and T5 induced clear morphological changes in the tested cells, accompanied by suppression of tumor phenotype and a shift of MM1.S and RPMI-8226 cells toward a more normal cellular state. Additionally, S27 was found to reverse the stimulatory effect of IL-6 on multiple myeloma cell proliferation, thereby inhibiting the growth of RPMI-8226 cells. In pharmacokinetic evaluations, S27 and T5 displayed favorable properties including high human intestinal absorption (HIA), 20% and 30% oral bioavailability (F20% and F30%), and desirable apparent volume of distribution (VD).

In 2024, Zhang et al. continued structural modification of compound S27 (Fig. 8).⁶⁶ The imidazole group was retained at the 4-position of the phenyl ring, while aniline and piperazine moieties were introduced onto the carboxylic acid group. The introduction of an iodine atom at the para-position of the phenyl ring (compound A7) and a trifluoromethyl group at the para-position of the benzoyl moiety (compound B15) led to improved activity. A7 emerged as the most potent compound (IC₅₀ = 3.76 μM), exhibiting a 1.3-fold increase in activity compared to S27. Both A7 and B15 demonstrated significant anti-proliferative effects. Treatment with A7 induced marked morphological changes in MM1.S and RPMI-8226 cells, characterized by suppression of the tumor phenotype and promotion of differentiation toward a normal cell state. A7 significantly increased the expression of the differentiation antigen CD49e on the cell surface, as well as the levels of κ-IgLG and λ-IgLG in MM1.S and RPMI-8226 cells. Additionally, A7 and B15 reversed the IL-6-induced proliferation of multiple myeloma cells. At a concentration of 5 μM, both compounds significantly enhanced the growth-inhibitory effect of ixazomib. Furthermore, the combination of A7 with ixazomib synergistically promoted apoptosis beyond the effect of ixazomib alone.

Since the bioactive molecules bearing a 2-phenylquinoline core exhibited no in vivo anticancer activity in animal models, Zhang et al. further modified their structures to optimize the physicochemical properties of the compounds in 2025.⁶⁷ Specifically, the 2-phenylquinoline moiety was replaced with a phenylpyridine scaffold; various substituents were introduced to the amino group on the benzene ring, and a 4-(4-methylpiperazin-1-yl)anilino group was attached to the carboxyl group on the pyridine ring (Fig. 8). During the structural modification, it was found that the fluorinated benzene ring could significantly enhance the enzyme inhibitory activity. Ultimately, compound D16 was designed, which acts as a highly selective SIRT3 inhibitor with potent in vivo antitumor activity.

2.2.7. Other types of small-molecule inhibitors

In 2013, Lahtela-Kakkonen and colleagues conducted a virtual screening campaign using a sequential workflow combining ligand shape-based matching and flexible ligand docking, with the aim of identifying novel scaffold structures for SIRT3 inhibitors (Fig. 9).⁶⁸ Based on docking scores, the top 40 compounds were selected for experimental validation. At a concentration of 200 μM, these compounds exhibited SIRT3 inhibition rates ranging from 6% to 74%, with eight showing over 35% inhibition. Two novel scaffold families for SIRT3 inhibition were identified through this process. Structural optimization revealed that introducing a chlorine atom onto the phenyl ring of the 5-amino-2-phenyl-benzoxazole core enhanced inhibitory activity, whereas other halogen or electron-donating substituents led to reduced potency. The 2,4-dichloro-substituted derivative 23 displayed the strongest SIRT3 inhibition (71%). The influence of halogen substitution on binding affinity was attributed to two main mechanisms: direct halogen–protein interactions and indirect effects modulating overall binding performance. The ortho-chloro substituent adopted a conformation favorable for halogen bonding, located within a beneficial region of the molecular interaction field, and formed a hydrogen bond with the amino group of L298. In contrast, in fluoro-substituted analogs, the greater distance between the fluorine atom and the amino group of L298 hindered hydrogen bond formation. The phenyl moiety was observed to be solvent-exposed. The binding site induced a slight conformational twist in the 5-amino-2-phenyl-benzoxazole core, a geometry particularly favorable for ortho-chloro-substituted derivatives.

Fig. 9. Other types of small-molecule SIRT3 inhibitors.

In 2015, Scholle et al. employed SAMDI-MS technology to identify an active SIRT3 substrate (GYKAcRGC) and used this peptide as the detection target to screen a library of 100 000 small molecules.⁶⁹ This effort led to the identification of 306 SIRT3 inhibitors. Among them, the most active compound, SDX-437, exhibited an IC₅₀ value of 700 nM against SIRT3 and showed no significant inhibition of SIRT1 at 100 μM, suggesting its potential as a useful chemical tool for further investigation of SIRT3 function.

In 2015, Lahtela-Kakkonen et al. conducted a virtual screening of a compound library utilizing a potential binding site located within the zinc-binding domain of SIRT3, situated between residues F186, Q260, D290, and E296, with the aim of identifying novel sirtuin inhibitor scaffolds. This effort resulted in the identification of 26 candidate compounds.⁷⁰ At a concentration of 200 μM, these compounds showed inhibition rates ranging from 6% to 59% against SIRT3. Among them, compound 24 exhibited 32% inhibition at 200 μM and was found to form hydrogen-bond interactions with Q260 and D290. Although 24 demonstrated the lowest activity in the series, its relatively small molecular size rendered it a promising starting point for subsequent structural optimization and structure–activity relationship studies.

In 2016, Kapila et al. developed a novel SIRT3 inhibitor, LC-0296, which exhibited IC₅₀ values of 67 μM, 33 μM, and 3.6 μM against SIRT1, SIRT2, and SIRT3, respectively, showing approximately 20-fold and 10-fold selectivity over SIRT1 and SIRT2.⁷¹ Experimental results demonstrated that LC-0296 could inhibit the proliferation and promote the apoptosis of HNSCC cells by increasing intracellular ROS levels. It also enhanced the sensitivity of HNSCC cells to radiotherapy and cisplatin treatment, while showing no inhibitory effect on normal human oral keratinocytes.

In 2018, Li et al. developed a novel DNA-encoded dynamic library approach and screened it against six protein targets.⁷² Six types of linkers were selected to bridge the small molecules identified from the screening. Four bridged compounds were successfully identified for SIRT3. Among them, the 77/39 pair demonstrated strong binding ability. When these two molecular units were directly connected (forming 77–39), the binding affinity was significantly enhanced, showing a K_d value of 2.14 μM against SIRT3. In contrast, the introduction of a linker between the two units led to a notable decrease in binding affinity. 77–39 exhibited 14-fold selectivity over SIRT2. At a concentration of 100 μM, 77–39 increased the acetylation level of mitochondrial proteins without affecting their expression and significantly reduced intracellular ATP levels.

In 2021, Parkesh et al. conducted a high-throughput screen of pharmacologically active compounds and identified several approved drugs with inhibitory activity against SIRT3.⁷³ Minaprine, an FDA-approved aminophenylpyridazine antidepressant, exhibited IC₅₀ values of 8.47 μM, 12.9 μM, and 13.2 μM against SIRT1, SIRT2, and SIRT3, respectively. Todralazine, a hydralazine-derived antihypertensive agent, showed IC₅₀ values of 21.1 μM, 18.7 μM, and 15.3 μM against SIRT1, SIRT2, and SIRT3. 8-Azaguanine, an FDA-approved antileukemic drug, acted as a selective SIRT3 inhibitor with an IC₅₀ of 51.6 μM. Ampicillin, a β-lactam antibiotic, also functioned as a selective SIRT3 inhibitor, displaying an IC₅₀ of 23.4 μM. Amprolium, an FDA-approved veterinary pyrimidine-based anticoccidial, inhibited SIRT1 and SIRT3 with IC₅₀ values of 16.9 μM and 18.3 μM, respectively.

In 2024, Al-Harrasi et al. identified novel SIRT3 inhibitors from an in-house database through an integrated computational approach.⁷⁴ Approximately 800 compounds were subjected to virtual screening by docking into the active site of human SIRT3, from which eight molecules were selected based on their favorable interactions with the enzyme. Among them, MI-212 showed the highest docking binding energy (−8.37 kcal mol⁻¹). These eight molecules demonstrated strong binding affinity for SIRT3 by forming multiple hydrogen bonds and π–hydrogen interactions with residues in the active site, all maintaining the lowest energy conformation. In vitro assays revealed that both MI-44 and MI-217 significantly inhibited the growth of MDA-MB-231 cells, with IC₅₀ values of 7.4 μM and 6.2 μM, respectively, and induced cellular apoptosis.

3. SIRT3 activators

3.1. Peptide-based activators

In 2016, Lu et al. discovered that exendin-4 (EX4) promotes the expression of SIRT3.⁹⁷ Their study demonstrated that in diabetic rats after one month of disease onset, the retina exhibited significantly increased cell death and elevated ROS levels, along with a marked reduction in b-wave amplitude and oscillatory potentials (OPs), as well as decreased expression of SIRT1 and SIRT3. Four-day intravitreal EX4 treatment significantly reduced retinal cell death and ROS levels, restored visual function, and returned SIRT1 and SIRT3 expression to near-normal levels. In R28 cells, hydrogen peroxide treatment increased ROS levels and cell death while downregulating SIRT1 and SIRT3 expression; these effects were reversed by the administration of EX4.

3.2. Small-molecule activators

3.2.1. Benzofuran derivatives

In 2021, Ouyang et al. identified a small-molecule activator of SIRT3 through structure-based drug design and high-throughput screening (Fig. 10).⁴⁹ The study initially identified two selective binding sites on SIRT3, designated as pocket L and pocket U. Screening against pocket U led to the discovery of compound ZINC03830212, and its co-crystal structure with SIRT3 was resolved (Fig. 10; PDB: 5H4D). The binding site of ZINC03830212 largely overlapped with pocket U. Upon binding, it induced a conformational change in the loop containing F157, enhancing π–π interactions between F157 and the nicotinamide group of NAD⁺. The iodobenzene ring of ZINC03830212 formed a π–anion interaction with E177, while the diethylamine group engaged with residues such as F157 and I179. The benzofuran core served primarily as a structural scaffold. Maintaining the binding mode of the ZINC03830212 scaffold to pocket U, a series of derivatives with varying terminal groups were synthesized. Compound 25, containing a morpholine group, showed a 2.6-fold increase in activity compared to ZINC03830212 (EC₅₀ = 1.23 μM). Elongating the linker to incorporate a pyrrolidine ring further improved potency in derivative 26 (EC₅₀ = 0.69 μM). The most potent compound, 27, featuring an alkane linker and an iodine atom on the phenyl ring, exhibited an EC₅₀ of 0.21 μM. Compound 27 significantly enhanced cellular SIRT3 activity, promoted the deacetylation of SOD2 and p53, and induced mitophagy via the SIRT3–Parkin pathway, leading to potent anti-proliferative effects in MDA-MB-231 cells. It showed no effect on the activity or expression levels of SIRT1, SIRT2, or SIRT5 at concentrations up to 100 μM. In vivo, 27 administered at doses of 25, 50, and 100 mg kg⁻¹ significantly suppressed tumor growth. While no notable change in body weight was observed in the low-dose groups, mice treated with 100 mg kg⁻¹ showed significant weight loss after one week, along with some pulmonary toxicity.

Fig. 10. (A) Structures, biological activity data, and development paths of compounds 25–27; (B) structure of ZINC03830212, and cocrystal structure of ZINC03830212 bound to SIRT3 (PDB 5H4D).

3.2.2. 1,4-Dihydropyridine derivatives

In 2009, Mai et al. designed novel sirtuin modulators based on a 1,4-dihydropyridine scaffold (Fig. 11).⁷⁵ They prepared derivatives containing 3,5-diethyl ester, 3,5-dicarboxyl, and 3,5-dicarbamoyl substituents and evaluated their activity against SIRT1/2/3 at a concentration of 50 μM. Compounds bearing a benzyl group at the N1 position consistently activated SIRT1/2/3, with the diethyl ester-substituted derivative MC2562 showing the strongest activity, exhibiting an EC₁₅₀ of approximately 50 μM for SIRT3. At 50 μM, MC2562 reduced the acetylation level of α-tubulin and induced cell cycle arrest at the G1/S phase, though it did not show significant antitumor activity in the MCF-7 cell line. MC2562 demonstrated anti-senescence effects, increased mitochondrial density, and stimulated the activity of the mTFA promoter containing a nuclear respiratory factor 1 (NRF1)-binding site in a peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α)-dependent manner.

Fig. 11. Structures, biological activity data, and development paths of 1,4-dihydropyridine derivatives as SIRT3 activators.

In their 2022 study, Mai and colleagues aimed to obtain selective SIRT3 activators through structural optimization of MC2562.⁷⁶ Replacing the benzyl group with a benzoyl moiety produced MC2791, which exhibited superior agonistic activity (∼320% at 200 μM). Alternatively, introducing methoxy groups (MC2789) slightly attenuated SIRT3 activation (269%) but provided a compound with measurable affinity (K_d = 32 μM). Functional assays in MDA-MB-231 cells at 50 μM showed that both analogs increased glutamate dehydrogenase activity, and MC2791 additionally decreased SOD2 acetylation, confirming their biological activity.

In 2023, Mai et al. developed a series of derivatives based on MC2791 and MC2789.⁷⁷ Removal of one meta-methoxy group from the phenyl ring of MC2789 yielded compound 28, which exhibited the strongest activity (387% activation), representing a 1.4-fold increase over MC2789 (269%). The binding affinity (K_d) of 28 for SIRT3 was determined to be 29 μM. In CAL-62 and MDA-MB-231 cell lines, 28 reduced the protein levels of HIF-1α, endothelial PAS domain-containing protein 1 (EPAS-1), and Snail family transcriptional repressor 1 (SNAIL).

In 2025, Li et al. continued structural optimization of MC2789 and synthesized a series of novel SIRT3 activators.⁷⁸ Replacing the benzene ring of MC2789 with a pyridine moiety and introducing chlorine atoms at the 2- and 3-positions yielded compound 29, which exhibited enhanced activity with approximately 260% activation of SIRT3 at a concentration of 50 μM and a binding affinity (K_d) of 51.51 μM. 29 inhibited glucose uptake and lactate production in MDA-MB-231 and BT-549 cells, thereby suppressing cell migration, colony formation, and proliferation. Furthermore, 29 induced a decrease in mitochondrial membrane potential, leading to reduced ATP levels and promoting apoptosis in both BT-549 and MDA-MB-231 cells.

3.2.3. Quinoxaline derivatives

In 2025, Ouyang et al. employed an integrated strategy combining computational simulations and experimental validation to screen for potential SIRT3 activators (Fig. 12).⁷⁹ Virtual screening against predicted binding sites identified five lead compounds, among which SKLB-4A exhibited the most potent activation, showing 118.19% activation at 100 μM, with an EC₅₀ of 21.95 μM and a K_d of 4.7 μM against SIRT3. The team subsequently optimized the structure of SKLB-4A to develop novel SIRT3 activators. Replacing the methoxy group in SKLB-4A with a fluorine atom yielded SKLB-4B, which showed improved activity. Further substitution of the cyano group in SKLB-4B with a methyl ester resulted in SKLB-11A, which achieved 188.14% activation of SIRT3. The crystal structure of the SIRT3–SKLB-11A complex (Fig. 12; PDB: 9KTK) revealed that the pyrrole ring of SKLB-11A inserts into the cleft between two structural domains, engaging in π–π stacking with H248. SKLB-11A also forms a hydrogen bond with V292 and is surrounded by F180 and H248. A significant conformational change was observed in the side chain of L298, whose orientation deviates from previously reported SIRT3 structures, suggesting that SKLB-11A may modulate SIRT3 activity through an allosteric mechanism. Functionally, SKLB-11A suppressed ROS elevation, maintained mitochondrial membrane potential stability, and induced mitophagy, thereby promoting the clearance of damaged mitochondria. In a doxorubicin-induced cardiac dysfunction model, SKLB-11A effectively ameliorated cardiac impairment, reduced serum levels of myocardial injury markers, and preserved myocardial structural integrity. At effective doses, SKLB-11A showed no significant toxicity and exhibited favorable pharmacokinetic properties.

Fig. 12. (A) Structure, biological activity data, and development paths of quinoxaline derivatives as SIRT3 activators; (B) structure of SKLB-11A, and cocrystal structure of SKLB-11A bound to SIRT3 (PDB 9KTK).

3.2.4. Other types of small-molecule activators

In 2024, Wang et al. demonstrated that MY-13 exhibits potent anti-tumor properties against colorectal cancer (Fig. 13).⁸⁰ By modulating SIRT3-mediated pathways involved in apoptosis and autophagic cell death, MY-13 significantly inhibited the proliferation and migration of human colorectal cancer cells, with IC₅₀ values of 5.23 μM in RKO cells and 14.71 μM in HCT-116 cells. The anti-proliferative effect of MY-13 was further validated in a subcutaneous xenograft tumor model, reinforcing its efficacy in suppressing tumor growth. MY-13 was identified as a potent SIRT3 activator, and its anti-cancer activity in colorectal cancer was shown to be dependent on SIRT3 activation.

Fig. 13. Other types of small-molecule SIRT3 activators.

In 2024, Yang et al. employed an integrated approach combining computational screening and biological evaluation to identify macrocyclic sulfonamide compound 30 as a promising SIRT3 activator with therapeutic potential in Parkinson's disease.⁸¹ Virtual screening of a macrocyclic sulfonamide library led to the identification of several candidate SIRT3 activators, among which compound 31 exhibited the strongest activity. Structure–activity relationship analysis revealed that macrocycles containing 16- to 19-membered rings showed moderate to good SIRT3 activation. Compounds with electron-donating groups displayed higher activity than those with electron-withdrawing groups. Alkyl side chains were generally associated with better activity than aryl substituents, and the introduction of amino acid fragments into aryl-containing derivatives reduced potency. Compound 31 exhibited a K_d of 11.3 μM and an EC₅₀ of 29.37 μM for SIRT3. Further studies showed that compound 30 significantly enhanced the thermal stability of SIRT3 in SH-SY5Y cells. In a Parkinson's disease mouse model, 30 facilitated the clearance of α-synuclein via SIRT3 activation, leading to improved motor function. Treatment with 30 also prevented the loss of dopaminergic neurons in the substantia nigra induced by AAV-A53T.

In 2024, Ouyang et al. identified a highly effective SIRT3 activator targeting the previously characterized pocket L.⁸² A virtual screening strategy based on pocket L was employed, leading to the selection of the top 20 candidate compounds, among which 2-APQC exhibited the most potent SIRT3 activation. 2-APQC demonstrated SIRT3-activating capability comparable to that of the known activator honokiol, with a K_d value of 2.756 μM and no significant cytotoxicity at 40 μM. By binding to and activating SIRT3, 2-APQC modulates multiple signaling pathways, including mTOR-p70S6K, JNK, and TGF-β/Smad3, and effectively attenuates isoproterenol-induced cardiac hypertrophy, fibrosis, and heart failure in both in vitro and in vivo models.

3.3. Natural products and their derivatives

Isosteviol is a natural ent-beyerane diterpenoid with broad pharmacological activities. In 2024, Zhao et al. designed and synthesized a series of isosteviol derivatives and evaluated their protective effects in a zebrafish cardiomyopathy model (Fig. 14).⁸³ At concentrations of 5 μM or 15 μM, isosteviol itself exhibited only weak activity. Introduction of a piperazine group at the C19 carboxyl position (compound 32) enhanced activity, increasing the survival rate of zebrafish from 60% to 80% at 5 μM. Further modification by attaching a benzoic acid group to 32 yielded compound 33, which showed even stronger activity, raising the survival rate from 50% to 95% at 5 μM. Compound 33 effectively alleviated cardiac edema, restored normal heart structure, and improved doxorubicin-induced cardiac dysfunction, significantly reducing the mRNA levels of myocardial markers Anp, Bnp, and cTnT. 33 exhibited very low toxicity even at concentrations above 20 μM. However, its poor solubility limited further application. To address this, the researchers synthesized the sodium salt form 34, which significantly improved solubility, achieving 54.02 mg mL⁻¹ compared to 0.94 mg mL⁻¹ for 33. 34 demonstrated nearly equivalent therapeutic efficacy to 33 in zebrafish. Mechanistic studies indicated that at concentrations of 1 μM and 10 μM, 34 activated the SIRT3/SOD2 and optic atrophy 1 (OPA1) pathways, restored diminished mitochondrial membrane potential, and thereby protected cardiomyocytes from oxidative damage.

Fig. 14. Structures, biological activity data, and development paths of compounds 32–34.

In 2017, Wang et al. screened a compound library using a deacetylation activity assay to identify SIRT3 activators, leading to the discovery of C12, which effectively activated SIRT3 and increased SOD2 activity by 5-fold (Fig. 15).⁸⁴ The EC₅₀ of C12 for SIRT3 was determined to be 75.78 μM, with a K_d of 3.9 μM. However, activation of endogenous SIRT3 required a C12 concentration of approximately 300 μM, which may be attributed to its low bioavailability resulting in limited uptake by primary neuronal cells.

Fig. 15. Natural products and their derivatives as SIRT3 activators.

To enhance the efficacy of C12 against pulmonary fibrosis, Godugu et al. developed a novel compound, MitoC12, by conjugating a triphenylphosphonium cation with C12.⁸⁵MitoC12 exhibited higher accumulation in mitochondria compared to the cytoplasmic fraction, reduced bleomycin-induced oxidative stress in BEAS-2B cells, and inhibited TGF-β-induced pulmonary fibrosis in MRC-5 cells. It exerted protective effects against bleomycin-induced lung fibrosis by improving pulmonary function, attenuating inflammatory responses, and restoring lung tissue architecture. Mechanistic studies revealed that MitoC12 reduced collagen deposition and downregulated the expression of fibrotic markers such as TGF-β, collagen 1A/3A, alpha-smooth muscle actin (α-SMA), fibronectin, and vimentin. Furthermore, it activated SIRT3 to regulate SOD2 and 8-oxoguanine DNA glycosylase 1 (OGG1) function, thereby maintaining mitochondrial homeostasis and contributing to its anti-fibrotic effects.

In 2023, Liu et al. developed a novel SIRT3 activator, SZC-6.⁸⁶ Molecular docking was performed on predicted potential allosteric sites on the SIRT3 surface, leading to the selection and synthesis of 33 candidate compounds, among which SZC-6 exhibited significant activation. SZC-6 displayed a K_d of 15 μM and an EC₅₀ of 23.2 μM for SIRT3. In a mouse model of isoproterenol (ISO)-induced cardiac hypertrophy, SZC-6 alleviated hypertrophic responses, reversed ISO-induced diastolic and systolic dysfunction, and inhibited cardiac fibroblast proliferation and differentiation into myofibroblasts. Mechanistic studies revealed that SZC-6 enhanced ATP production and mitochondrial oxygen consumption rate while reducing reactive oxygen species levels in ISO-treated neonatal rat cardiomyocytes (NRCMs) through SIRT3 activation, thereby improving mitochondrial function. Additionally, SZC-6 promoted liver kinase B1 (LKB1) phosphorylation, leading to adenosine monophosphate-activated protein kinase (AMPK) activation and subsequent inhibition of Drp1-dependent mitochondrial fission.

In 2025, Liang et al. investigated the mechanism underlying lipophagy in cardiomyocytes and evaluated whether berberine could ameliorate diabetic cardiomyopathy (DCM) by modulating this pathway.⁸⁷ The study revealed impaired myocardial lipophagy in DCM, accompanied by significant downregulation of SIRT3, a key regulator of the process. Treatment with the SIRT3 activator nicotinamide riboside (NR) enhanced lipophagy and alleviated palmitic acid (PA)-induced lipotoxicity in H9C2 cells. Animal experiments further demonstrated that berberine significantly improved diabetes-induced cardiac dysfunction and myocardial hypertrophy in db/db mice, an effect dependent on SIRT3-mediated activation of lipophagy.

In 2025, Li et al. identified ginsenoside Rh1 as an effective SIRT3 activator and investigated its protective mechanism against mitochondrial dysfunction.⁸⁸ Using molecular docking, they screened ten rare ginsenosides for SIRT3 activation and found that Rh1 exhibited the strongest binding affinity to SIRT3. Rh1 significantly improved cardiac function, alleviated myocardial ischemia injury, reduced oxidative stress in hypoxic cardiomyocytes, and restored mitochondrial network morphology and respiratory function. Mechanistic studies revealed that Rh1 binds to SIRT3, upregulates Foxo3a expression, promotes its nuclear translocation, and reduces acetylation levels, thereby enhancing mitochondrial fusion, suppressing fission, and accelerating autophagy. The regulatory effects of Rh1 on oxidative stress, mitochondrial dynamics, and autophagy were reversed by SIRT3 siRNA, confirming the essential role of SIRT3 in mediating these protective actions.

In 2025, Li et al. identified rhynchophylline (Rhy) as a potent SIRT3 activator and further elucidated its protective mechanisms against mitochondrial damage in endothelial progenitor cells (EPCs) and endothelial dysfunction.⁸⁹ In a hypertensive rat model, Rhy not only reduced blood pressure and improved vasomotor function but also significantly enhanced the biological activity of EPCs in peripheral circulation. Rhy alleviated mitochondrial damage and reduced apoptosis by inhibiting the mitochondrial apoptotic pathway. SIRT3 knockout abolished Rhy's regulatory effects on mitochondrial homeostasis, thereby reversing its protective benefits against EPC dysfunction and endothelial injury. Rhy suppressed mitochondrial ROS generation and enhanced SOD2 activity in a SIRT3-dependent manner, whereas SOD2 silencing completely eliminated Rhy's inhibitory effects on oxidative stress and apoptosis. These findings demonstrate that Rhy, as a targeted SIRT3 activator, prevents mitochondrial dysfunction through the SIRT3/SOD2 signaling pathway.

4. Perspectives

4.1. Development of highly active and selective SIRT3 modulators

SIRT3, as a central regulator of mitochondrial metabolism, faces significant challenges in the development of both inhibitors and activators, particularly concerning potency and selectivity. Although numerous reported small molecules show promise, most exhibit activities only in the micromolar range and lack sufficient selectivity among sirtuin subtypes, especially SIRT1 and SIRT2, which limits their further investigation and application. For example, while compound 15 achieves an impressive IC₅₀ of 4 nM against SIRT3, it also potently inhibits SIRT1 and SIRT2. The difficulty in designing selective SIRT3 modulators stems largely from the high sequence homology between SIRT3 and other sirtuin family members, particularly SIRT1, SIRT2, and SIRT5. Future research should integrate structural biology and computational chemistry more deeply. Techniques such as cryo-electron microscopy and molecular dynamics simulations can help resolve SIRT3-specific conformations and identify unique allosteric binding pockets to guide rational drug design. Meanwhile, artificial intelligence-assisted virtual screening can accelerate the discovery of novel scaffolds, for example, by using deep learning to explore chemical space complementary to the catalytic core or allosteric sites of SIRT3, while avoiding conserved NAD⁺-binding regions to enhance selectivity.⁹⁰ In addition, studies on dynamic structure–activity relationships are essential to elucidate how ligand-induced conformational changes influence substrate channel opening and closure, potentially revealing selectivity switches. Ultimately, a multidimensional evaluation system, incorporating enzymatic assays, cellular acetylomics, transcriptomics, and phenotypic profiling, will be crucial to transition from pan-sirtuin modulation to SIRT3-specific regulation, providing high-precision chemical tools for disease treatment.

4.2. Exploration of organelle- and tissue-specific delivery systems

A key challenge facing current SIRT3 modulators is their off-target effects and systemic toxicity. For example, in addition to inhibiting sirtuins, nicotinamide derivative 8 also suppresses dehydrogenases and other NAD⁺-dependent enzymes; meanwhile, the activator 27 exhibits pulmonary toxicity in mice at high concentrations. A promising direction lies in developing precise delivery strategies. Utilizing membrane potential-driven carriers such as quaternary phosphonium or ammonium salts can enable mitochondrial-targeted delivery, while covalent conjugation or nano-encapsulation can enhance drug concentration in the mitochondrial matrix and reduce off-target effects in the cytoplasm.⁹¹ Based on disease microenvironment characteristics, stimuli-responsive carriers can be designed to achieve controlled drug release at pathological sites. Furthermore, antibody- or aptamer-coupled technologies can guide drugs to specific cell populations, minimizing damage to healthy tissues.⁹²

4.3. Mitochondrial protease-mediated SIRT3 degradation

Traditional inhibitors typically only block the enzymatic activity of SIRT3, whereas proteasome-targeted degradation can completely eliminate its function, offering new therapeutic strategies for drug-resistant or SIRT3-overexpression-related diseases. Within mitochondria, Lon peptidase 1 (LONP1) and caseinolytic protease proteolytic subunit (ClpP) serve as potential mediators for protein degradation.^93,94 Both ClpP and LONP1 are involved in degrading misfolded proteins inside mitochondria, and previous studies have reported the use of bifunctional molecules targeting ClpP to effectively degrade monomeric streptavidin and RNA polymerase.^95,96 The primary steps include identifying degradation signals of SIRT3 and screening molecular glues or bifunctional degraders that can induce the binding of SIRT3 to LONP1/ClpP. For example, small molecules can be designed with one end binding to SIRT3 and the other recruiting LONP1, leveraging the mitochondria-specific ubiquitin-like modification system to achieve targeted degradation. In disease models where SIRT3 overexpression drives cancer progression (such as DLBCL), a degradation strategy may be more effective than inhibition. Conversely, for diseases associated with SIRT3 deficiency (such as heart failure), inhibitors of LONP1/ClpP could be developed to stabilize functional SIRT3 protein. Further challenges include overcoming mitochondrial membrane barriers and substrate selectivity of proteases, which may be addressed through membrane-penetrating peptide-assisted delivery or the development of mitochondria-localized proteolysis-targeting chimeras.

5. Conclusion

SIRT3, as a mitochondrial deacetylase, plays a dual regulatory role in various diseases, highlighting the significant therapeutic potential of developing its modulators. To date, multiple structural classes of SIRT3 inhibitors and activators have been identified, some of which demonstrate promising activity and preliminary efficacy in both cellular and animal models. However, current SIRT3 modulators still face challenges such as insufficient potency, poor subtype selectivity, and difficulties in tissue-specific delivery. Future research should leverage structural biology, computational chemistry, and artificial intelligence to deeply explore SIRT3-specific binding pockets and allosteric mechanisms, facilitating the development of highly selective and potent modulators. In parallel, advances in mitochondrial-targeted delivery systems and protease-mediated degradation strategies will enhance the precision and safety of these agents. Through these efforts, SIRT3 modulators are expected to emerge as novel therapeutics for cardiovascular diseases, neurodegenerative disorders, and cancer.

Author contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

No primary research results, software or code have been included and no new data were generated or analysed as part of this review.