Peroxisome proliferator-activated receptor γ (PPARγ) is one of the master regulators of adipocyte differentiation and is closely related to the development of obesity (24). Under fasting conditions, SIRT1 is activated to promote CDK2-associated cullin 1 (CACUL1) binding to PPARγ-responsive site without affecting CACUL1 expression, in turn repressing the transcriptional activity and adipogenic potential of PPARγ in 3T3-L1 cells (24). It has been reported that ECM in AT accelerates early adipogenesis, and SIRT1 mediates proadipogenic events triggered by ECM in subcutaneous AT (25). Conversely, SIRT7 promotes adipogenesis in mice by inhibiting the autocatalytic activation of SIRT1 (81). Strikingly, another report suggested that SIRT1 plays a positive role in the early stage of mouse embryonic stem cell (mESC) differentiation to adipocyte and exhibits a negative effect on the late stage of adipogenesis (50). SIRT1 deficiency increases the acetylation of nuclear receptor corepressor 1 (NCOR1) to inhibit adipogenesis; thus, SIRT1 plays a positive role during the early stage of mESCs to adipocyte differentiation (50).

In AT-derived mesenchymal stem cells (AT-MSCs), SIRT1 activated beige adipocyte differentiation in elderly AT-MSC via p53/p21 pathway, but not SIRT3 (52). On the contrary, another study indicated that the effects of SIRT1 on brown AT are not related to the differentiation process of brown adipocytes (29). The expression patterns of SIRT3 are correlated with brown adipocyte differentiation, and SIRT3 directly deacetylates PPAR coactivator-1α (PGC-1α) to promote the differentiation of fully thermogenic competent brown adipocytes (59). Depletion of SIRT3 decreases the protein level of forkhead box O 3a (FOXO3a) and subsequently impairs the ability of AT-derived human MSCs to undergo adipogenic differentiation, resulting in adipocyte dysfunction and IR (63).

It is worth noting that SIRT2 inhibits adipocyte differentiation. Apparently, the effects of SIRT1 and SIRT2 are similar in adipogenesis. Overexpression of SIRT1 and SIRT2 reduced lipid droplet number, attenuated lipid accumulation and adipocyte conversion, and prevented the proper induction of adipogenic markers, including PPARγ2 and CCAAT/enhancer-binding protein α (C/EBPα), in visceral AT-derived stem cells. Thus, the decreased expression of SIRT1 and SIRT2 promotes the differentiation capacity of visceral AT stem cells from obese humans, which is associated with fostering visceral AT expansion (26). SIRT2 directly deacetylates FOXO1 to affect FOXO1 acetylation/phosphorylation, increase its nuclear translocation, and suppresses adipogenesis (56). Moreover, SIRT2 promotes FOXO1 binding to PPARγ and subsequently inhibits its transcriptional activity to suppress adipogenesis (55).

In bovine AT, SIRT4 interacts coordinately with the transcription factors, including E2F transcription factor-1 (E2F-1), C/EBPβ, and homeobox A5 (HOXA5), to promote adipocyte differentiation (65). SIRT4 promotes branched-chain amino acid catabolism through methylcrotonyl-CoA carboxylase 1 (MCC1), resulting in the greatest enhancement of adipogenesis, and SIRT5 only modestly promotes adipogenesis, but not SIRT3 (64).

A recent study reported that SIRT5 inhibition stimulates brown-like adipogenesis in 3T3-L1 preadipocytes (67). SIRT5 inhibits the differentiation of bovine preadipocytes and simultaneously inhibits lipid synthesis and lipid deposition in adipocytes by activating the adenosine 5′-monophosphate (AMP)-activated protein kinase (AMPK) and repressing the mitogen-activated protein kinase (MAPK) (68). Similarly, SIRT6 inhibits preadipocyte differentiation and synergizes with SIRT5 to reduce lipid deposition in preadipocytes through the activation of the AMPKα pathway (71). On the contrary, another study showed that SIRT6 promotes mitotic clonal expansion during adipogenesis by inhibiting the expression of kinesin family member C (KIFC) and enhancing casein kinase 2 (CK2) activity, while SIRT6 deficiency leads to a severe adipogenesis defect and reduced expression of adipogenic markers, including PPARγ, C/EBPα, adipocyte protein2 (aP2), and adiponectin (79).

In summary, SIRT1 regulates adipogenesis either positively or negatively, which depends on the stage of adipocyte differentiation, whereas SIRT7 promotes adipogenesis by inhibiting the autocatalytic activation of SIRT1 ( Figure 1 ). SIRT2 exerts an inhibitory effect on adipogenesis by deacetylating FOXO1. SIRT3 and SIRT4 promote adipogenesis, while SIRT5 and SIRT6 impair preadipocyte differentiation and lipid deposition. Manipulation of the activity of SIRT1–6 might be a promising strategy to control adipocyte fate.

It was reported that upregulation of SIRT1 triggers lipolysis and loss of fat, including the hydrolysis of TGs and the release of free fatty acid (FFA), by repressing PPARγ (27). ATGL protein is considered the rate-limiting lipolytic enzyme since the rates of lipolysis are directly proportional to the levels of the ATGL protein. SIRT1 controls fat storage and mobilization, at least in part by regulating lipolysis in adipocytes via FOXO1-mediated expression of ATGL (28). Another study indicated that SIRT1 activates AMPK, which plays a crucial role in adipocyte lipolysis (20). Consistently, in white AT, SIRT1 knockdown not just fully recovered the resveratrol-elevated ATGL and positive regulatory domain containing 16 (PRDM16) protein but also reduced the resveratrol-elevated AMPK phosphorylation, suggesting that SIRT1 is the upstream factor of AMPK to control lipid metabolism (57).

SIRT2 increases lipolysis in mature adipocytes (55). SIRT2 mediates the increase in fatty acid oxidation (FAO) upon hypoxia-inducible factor-1α (HIF-1α) inactivation via PGC-1α, but not SIRT1 (58).

SIRT3 is able to prevent the loss of brown AT during obesity and metabolic disorders. Knockout of SIRT3 obviously promoted lipid droplet accumulation in brown AT (21). Interestingly, SIRT3 plays a minimal role in AT mitochondrial biology and systemic metabolism in adipocytes from SIRT3 knockout mice. Loss of SIRT3 in adipocytes showed no obvious change in metabolic responses to HFD feeding and aging (115). Overexpression of SIRT3 activated macroautophagy in mature adipocytes by activating the AMPK–unc-51-like kinase 1 (ULK1) pathway, which in turn resulted in smaller lipid droplet size and reduced lipid accumulation (60).

HFD feeding increases SIRT4 levels in mice (116), while in nutrient-replete conditions, SIRT4 is active to repress FAO by deacetylating malonyl CoA decarboxylase (MCD), an enzyme that produces acetyl-CoA from malonyl-CoA and stimulates lipogenesis (66).

SIRT5 deficiency impairs FAO, glutamate dehydrogenase (GDH) activity, and metabolic flexibility in brown adipocytes (69). In addition, a recent study reported that SIRT5 deficiency stimulates brown adipogenesis and ATGL function, which reduces intracellular lipid storage by promoting lipolysis and ultimately affects brown AT function (67). Strikingly, one study reported that SIRT5 deficiency did not cause any significant metabolic abnormalities under either chow or HFD conditions (70). Consistently, another study suggested there were no differences in the expression of genes related to fatty acid synthesis or transport, lipolysis, mitochondrial oxidative phosphorylation, or glucose transport in brown AT-specific SIRT5 knockout mice under standard housing conditions (22°C, normal chow diet) (69). Contradictory results might be due to different experimental subjects and conditions. Therefore, more studies are needed to confirm the effects of SIRT5 on lipid mobilization.

SIRT6 overexpression is associated with the downregulation of a selective group of PPAR-responsive genes and genes associated with lipid storage, including angiopoietin-like protein 4 (ANGPTL4), adipocyte fatty acid-binding protein (FABP4 or aP2), and diacylglycerol acyltransferase 1 (DGAT) (72). Fat-specific SIRT6 knockout sensitized mice to HFD-induced obesity, which was attributed to adipocyte hypertrophy instead of adipocyte hyperplasia by decreasing expression of ATGL. Furthermore, SIRT6 deficiency suppresses the expression of ATGL by regulating FOXO1 acetylation and subcellular localization, thereby inhibiting lipolytic activity (73).

In conclusion, SIRT1 triggers lipolysis by repressing PPARγ, activating AMPK, and increasing the expression of ATGL. SIRT2 mediates the increase in FAO via PGC-1α. SIRT3 reduces lipid droplet size and lipid accumulation by activating the AMPK-ULK1 pathway. SIRT4 represses FAO by deacetylating MCD. The role of SIRT5 in lipid mobilization is controversial and needs further investigation. SIRT6 deficiency leads to decreased lipolytic activity by increasing phosphorylation and acetylation of FOXO1. In contrast, little is known about SIRT7 in lipid mobilization ( Figure 2 ). Understanding the roles of Sirtuins in lipid mobilization would likely provide key insights into developing therapeutics against obesity and obesity-induced metabolic disease.

SIRT1 has been reported to alleviate inflammation in a variety of tissues and cells (93). SIRT1 in AT plays a protective role against inflammation through multiple mechanisms. Macrophage infiltration and the gene expression of inflammatory cytokines were elevated in the heterozygous SIRT1 knockout mice fed with a moderate-fat diet (30). AT macrophages surround and ingest dying or dead adipocytes to form crown-like structures (CLSs). Myeloid SIRT1 deficiency promotes massive macrophage infiltration in AT and increases the number of CLSs (31, 32). Myeloid SIRT1 regulates pro-inflammatory cytokines and macrophage infiltration in AT from HFD mice (32). In addition, mice with AT-selective overexpression of human SIRT1 (H363Y), a dominant-negative mutant that inhibits endogenous SIRT1 activity, exhibited elevating inflammation (33). AT-specific-SIRT1 (H363Y) overexpressed mice exhibits hyperglycemia, dyslipidemia, and ectopic lipid deposition at a much younger age than their wild-type littermates (33). However, the pro-inflammatory effect of HFD is the triggering signal of SIRT1 cleavage.

In general, SIRT1 regulates both pro-inflammatory and anti-inflammatory cytokines. The pro-inflammatory signals that activate caspase-1 through the nucleotide-binding oligomerization domain, leucine-rich repeat and pyrin domain-containing 3 (NLRP3) inflammasome, such as TNF-α, cause the cracking of SIRT1. Indeed, AT-specific SIRT1 knockout mice obviously develop IR (34). In visceral AT, a negative correlation between the mRNA level of SIRT1 and IL-1β was observed (35). In addition, SIRT1 blocks the infiltration of macrophages and promotes the polarization toward anti-inflammatory M2 macrophages, which in turn ameliorates inflammation in AT (46); concomitantly, SIRT1 deacetylates the transcription factor NFATc1, thereby enhancing the binding of NFATc1 to the IL-4 gene promoter and finally modulating macrophage polarization (53).

Nuclear factor-κB (NF-κB) acts as a key regulator of inflammation to induce pro-inflammatory cytokines, which in turn increase adiposity and AT dysfunction (121). SIRT1 acts as a negative regulator of the inflammatory pathway and a positive regulator of insulin signaling in adipocytes by deacetylating NF-κB and inhibiting binding to the promoter of its target genes (36–40). Consistently, SIRT1 knockdown in white AT leads to NF-κB nuclear translocation by reducing histone H3 lysine 9 (H3K9) deacetylation (122). Interestingly, loss of SIRT1 leads to compensatory SIRT6 deacetylase activity on H3K9, demonstrating that SIRT1 and SIRT6 act on NF-κB through different mechanisms (40). Moderate SIRT1 overexpression ameliorates the effects of LPS on brown AT inflammation by the reduced acetylation of NF-κB, STAT3, and p38 MAPK (29, 39, 123). It is generally believed that the cluster of differentiation 40 (CD40)/CD40 ligand (CD40L) pathway is an integral part of the onset and maintenance of inflammatory reactions in obesity. SIRT1 modulates TNF-α-induced expression of CD40 partially via the NF-κB pathway in 3T3-L1 adipocytes (41). SIRT1 is a key upstream regulator in AT inflammation, by controlling the gain of pro-inflammatory transcription in response to inducers including fatty acids, hypoxia, and endoplasmic reticulum stress. The activation of SIRT1 by small molecules reduces the inflammatory response induced by FFA in macrophages and alleviates inflammation in white AT (38). SIRT1 suppresses NF-κB signaling pathway, which might provide another way to maintain AMPK activity against inflammatory challenges (42).

Phosphorylation of protein kinase B (Akt) activates the mammalian target of rapamycin (mTOR) signal in macrophages and then triggers inflammation and IR in obese mice (124). SIRT1 interacts with Akt2 and inhibits the mTOR/S6K1 pathway to attenuate AT inflammation in mice (43). Adiponectin, an adipocyte-derived relaxation factor with anti-inflammatory activity, promotes nitric oxide production in the endothelium. SIRT1 upregulates adiponectin mRNA expression in 3T3-L1 adipocytes via a FOXO1-C/EBPα transcriptional complex (44). SIRT1 improves the release of adiponectin from perivascular AT to fight against inflammatory insult (42).

Strikingly, adipocyte-specific deletion of SIRT1 exacerbates the detrimental effects of acute HFD feeding but shows protective effects in the context of chronic HFD exposure. Consistent with the more glucose tolerant and less insulin state of the adipocyte-specific SIRT1 knockout mice after chronic HFD, these mice possessed lower circulating levels of MCP-1 and TNF-α but increased levels of IL-10 and arginase in epididymal AT, which were mediated through hyperacetylation and dephosphorylation of PPARγ Ser273 along with reduced CDK5 activity. Additionally, increased p65 acetylation (NF-κB) was detected in these mice. Therefore, in the case of chronic HFD-induced obesity, inhibition of SIRT1 in adipocytes might result in improved metabolic functions (45). During the onset of obesity, SIRT1 deficiency in adipocytes (rather than myelocytes) accelerates peripheral IR by regulating macrophage infiltration and polarization, which has nothing to do with obesity (53).

Most studies have shown that SIRT1 plays a protective role in AT inflammation, but there are still very few studies with the opposite observation. The types of diet and the stage of obesity could affect the roles of SIRT1 in AT inflammation. Therefore, more studies are needed to further confirm the roles of SIRT1 in AT inflammation in the future.

The roles of other Sirtuins in AT inflammation have been seldom studied. SIRT3 mediates a metabolic switch in macrophages by deacetylating pyruvate dehydrogenase E1α (PDHA1) lysine, which in turn promotes NLRP3 inflammasome activation (61). Jun N-terminal kinase (JNK) in macrophages contributes to the accumulation of macrophages and plays a key role in the metabolic response to obesity, including IR (125). SIRT6 deficiency causes elevated inflammation in AT of mice by increasing the occupancy of c-Jun, downstream of JNK, on the gene promoters of IL-6 and MCP-1 (73). Another study suggested that SIRT6 regulates at least two stages of adipose inflammation: augmenting the migration potential of macrophages toward AT-derived chemoattractants and facilitating pro-inflammatory M1 polarization; SIRT6 deletion in macrophages promotes the activation of NF-κB and production of IL-6, resulting in STAT3 activation and the positive feedback circuits for NF-κB stimulation (74). Furthermore, adipocyte SIRT6 drives macrophage polarization toward M2 by increasing the production of the canonical type 2 cytokine IL-4 by adipocytes in a cell autonomous manner, which in turn attenuates pro-inflammatory responses in AT (80). Consistently, adipocyte SIRT6 decreases the M1 composition in white AT, and SIRT6 deficiency in adipocytes leads to an aggravating inflammatory reaction in white AT (23). Eosinophils secrete Th2 cytokine IL-4/IL-13 promoting M2 macrophage polarization. Myeloid-specific SIRT6 deficiency affects both eosinophils and M2 macrophage content in subcutaneous AT upon cold exposure, confirming a close link between eosinophils and M2 macrophages (76).

Taken together, SIRT1, SIRT3, and SIRT6 participate in the regulation of AT inflammation, while other Sirtuins remain poorly understood and need further researches ( Figure 3 ).

SIRT1 modulates fatty acid metabolism and lipid mobilization in adipocytes via modification of the extracellular environment. Small adipocytes, less ECM between adipocytes, and reduced macrophage infiltration in AT were observed in SIRT1 null animals (46). Consistently, SIRT1 deficiency suppresses the expression of leptin, adiponectin, and matrix metalloproteinases 3/13 (MMP3/13) and elevates the expression of the pro-fibrotic collagen (Collagen 6A3) in adipocytes. Pathway analysis revealed SIRT1-dependent key transcription factors, including PPARα, sterol regulatory element-binding transcription factor 1/2 (SREBF1/2), and PGC-1α (51). Moreover, SIRT1 overexpression downregulates the genes related to ECM remodeling (i.e., collagens, metalloproteases, and integrins) accompanied by a lower degree of inflammation-related fibrosis in visceral AT (39).

Upon Angiotensin (Ang) II stimulation, adipocytes adjacent to the adventitia enlarge, while collagen IV deposition in perivascular AT increases. Myeloid SIRT3 deficiency resulted in severe loss of brown AT characteristics and increased expression of collagen VI, which ultimately aggravated Ang II-induced perivascular AT dysfunction (61).

SIRT1 participates in the regulation of AT fibrosis by controlling collagens and metalloproteases. The roles of other Sirtuins in AT fibrosis are seldom studied.

Pro-opiomelanocortin (POMC) neurons selectively regulate brown AT-like remodeling of perigonadal white AT, and SIRT1 deficiency affects the survival of POMC neurons (47). SIRT1 promotes AT browning in mice via the deacetylation of PPARγ on Lys268 and Lys293, which in turn recruits the transcriptional coactivator PRDM16 to PPARγ, resulting in selective induction of brown AT genes and repression of white AT genes (27, 48). It has been reported that SIRT1 regulates angiogenesis by modulating angiogenic factors (such as vascular endothelial growth factor, platelet-derived growth factor, and transforming growth factor-β), which in turn controls AT function (46). PGC-1α, highly expressed in brown AT, is a key factor of brown fat thermogenesis and white AT browning (20, 75). SIRT1 deficiency displays a lower thermogenic activity and a significant decrease of UCP1 and PGC-1α expression in brown AT from HFD-fed mice, which are accompanied by aggravated mitochondrial dysfunction. In line with this, oxidation genes were downregulated, including PPARα, peroxisomal acyl-coenzyme A oxidase (ACOX), and carnitine palmitoyltransferase 1b (CPT1b), and the mitochondria content was lower, ultimately leading to brown AT degeneration (54). Furthermore, SIRT1 is an endogenous activator of FGF21 in hepatocytes, which in turn systemically controls white AT browning and energy homeostasis (49). Another study reported that SIRT1 induces white AT browning following sleeve gastrectomy by activating AMPK (20).

SIRT3 deletion aggravates brown-to-white adipocyte conversion induced by high salt via inhibiting mitochondrial biogenesis and perilipin-1 expression; however, restoring SIRT3 prevents high salt-induced brown AT to white AT conversion by improving mitochondrial respiration (62). Consistently, knockout of SIRT3 promotes the accumulation of lipid droplets in brown AT and blocks the inhibitory effect of capsaicin on HFD-induced brown AT whitening (21). Myeloid SIRT3 deficiency reduces UCP1 and perilipin-1 protein levels and promotes collagen IV deposition in brown AT, which in turn exacerbates perivascular AT remodeling and AT dysfunction (61).

SIRT5 deficiency in mice resulted in less browning capacity in subcutaneous AT and a slight imbalance in energy and glucose homeostasis, which might be related to the inhibition of isocitrate dehydrogenase (IDH) activity and reduction of α-ketoglutarate concentration (22). In inguinal AT from SIRT5 knockout mice, the expression of thermogenic genes including UCP1, cell death-inducing DFFA-like effector a (CIDEA) and cytochrome c oxidase subunit 7A1 (COX7A1), and FAO genes, including CPT1b and medium-chain acyl-coenzyme A dehydrogenase (MCAD), was downregulated, while the expression of adipogenic genes (PPARγ and C/EBPα) was not changed (22). Another report suggested that mice lacking SIRT5 in brown AT showed difficulty converting fuel from FFA to glucose after overnight fasting. Strikingly, there was no difference in the expression of genes related to browning in brown AT-specific SIRT5 knockout mice (69).

SIRT6 regulates AT browning in response to either cold exposure or β-adrenergic agonist (75). Similarly, SIRT6 deficiency in POMC neurons impairs the browning and lipolytic functions of AT in HFD-fed mice by modulating leptin signaling (77). Additionally, SIRT6 deficiency in mice resulted in impaired AT browning and reduced expression of UCP1, accompanied by downregulation of p38 MAPK/ATF2 signaling (23). SIRT6 deficiency leads to impairment of thermogenesis and browning by decreasing UCP1 and PGC-1α in subcutaneous white AT of myeloid SIRT6 knockout mice (76). SIRT6-deficient mice showed a significant increase in glucose uptake in brown AT, and HIF-1α is required to recruit SIRT6 to glycolytic gene promoters (78).

Taken together, emerging evidence has demonstrated that the Sirtuin family plays crucial roles in AT browning, comprehensively contributing to metabolic functions ( Figure 4 ).

Selisistat, a selective SIRT1 inhibitor, was shown to be safe (128) and well tolerated in healthy volunteers and Huntington’s disease (HD) patients in short-term studies (129). Strikingly, a study showed that administration of selisistat over 14 days showed no pro-inflammatory effects (129), although some pre-clinical studies suggested its pro-inflammatory effect.

Resveratrol was identified as the most potent SIRT1 activator (130). One study showed that 150 mg/day trans-resveratrol (99.9%) supplementation in obese subjects for 30 days decreased the expression levels of inflammation-related genes, plasma levels of several inflammatory markers, and leukocyte numbers, and reduced AT lipolysis (131). However, resveratrol supplementation (75 mg/day) in non-obese, postmenopausal women with normal glucose tolerance did not change body composition, resting metabolic rate, plasma lipids, or inflammatory markers (132). Longevinex is a modified form of resveratrol; it had no effect on insulin sensitivity or the inflammation markers (IL-6) in 34 patients diagnosed with metabolic syndrome (133).

Because resveratrol suffered from low bioavailability and potency, as well as low target specificity, synthetic Sirtuin activators are emerging, such as SRT501, SRT2104, SRT2379, and SRT3025 (127). SRT501 has entered phase III clinical trials for the treatment of type 2 diabetes, and the pharmacokinetics and safety study of SRT2379 evaluated in healthy male volunteers have been completed (130). The promising clinical data on SRT2104 revealed that small-molecule SIRT1 activators with good pharmacokinetics and tolerability profiles could provide important new therapeutic paradigms and be developed as candidates to treat inflammatory diseases (134, 135). There is still no Sirtuin activator or inhibitor approved for clinical use; further structural modifications, pharmacological evaluations, and clinical trials are needed.