Adipose tissue regulatory T cells: differentiation and function

Abstract

Rising obesity levels, worldwide, are resulting in substantial increases in cardiovascular disease, diabetes, kidney disease, musculoskeletal disorders, and certain cancers, and obesity-associated illnesses are estimated to cause ~4 million deaths worldwide per year. A common theme in this disease epidemic is the chronic systemic inflammation that accompanies obesity. CD4⁺ Foxp3⁺ regulatory T cells residing in visceral adipose tissues (VAT Tregs) are a unique immune cell population that play essential functions in restricting obesity-associated systemic inflammation through regulation of adipose tissue homeostasis. The distinct transcriptional program that defines VAT Tregs has been described, but directly linking VAT Treg differentiation and function to improving insulin sensitivity has proven more complex. Here we review new findings which have clarified how VAT Tregs differentiate, and how distinct VAT Treg subsets regulate VAT homeostasis, energy expenditure, and insulin sensitivity.

Graphical Abstract

Background

CD4⁺ regulatory T cells (Tregs) have remained at the forefront of peripheral tolerance, required for the suppression of autoreactive T cells but also required at tissue-specific sites to maintain tissue homeostasis [1]. The signature transcription factor associated with Tregs, Foxp3, is essential for Treg development with Foxp3-deficiency resulting in severe autoimmunity in mice and humans. In recent years, a function for Tregs beyond simply regulating peripheral tolerance has emerged. Tissue-resident Tregs residing in different tissues including the skin, muscle, brain, gut, and visceral adipose tissue (VAT) have all been described, with each of these subpopulations performing essential functions to promote tissue homeostasis at their site of residence [2,3].

In 2009, a landmark paper described a substantial population of CD4⁺ Foxp3⁺ Tregs that reside in the VAT in both mice and humans [4]. These visceral adipose tissue-resident Tregs (VAT Tregs) express the transcription factors GATA3, PPARγ, Id2, and Blimp-1, the surface markers KLRG1, CCR2, and ST2, and the cytokine IL-10[4–10]. While Tregs make up ~30% of total CD4⁺ T cells in the VAT, obese mice and humans have reduced frequency of VAT Tregs and decreased VAT Tregs correlates with increased insulin resistance and glucose intolerance [4]. Short-term loss of Tregs using Foxp3-DTR mice also resulted in increased insulin resistance and increased inflammatory mediators in the VAT including IL-6, TNFα, and RANTES [4].

Tregs are not the only immune cell population present in the VAT and several studies have helped to clarify the role of different immune cells in the adipose tissue depots. Anti-inflammatory immune cells are more abundant in the VAT isolated from lean mice and humans. Along with VAT Tregs, anti-inflammatory macrophages [5–8], type 2 innate lymphoid cells (ILC2s) [9,10], invariant Natural Killer T cells (iNKT) [11,12], eosinophils [13], regulatory B cells [14], and B1 cells [15,16] all contribute to the maintenance of the VAT homeostasis in lean mice and humans. In contrast, obese VAT houses a more inflammatory immune cell cohort with pro-inflammatory macrophages [17,18], Th1 CD4+ T cells [19,20], cytotoxic CD8+ T cells [21], and natural killer (NK) cells [22] all present at greater frequencies in the adipose tissue depots of obese mice and humans.

VAT Tregs develop in the thymus, traffic to the lymphoid organs, and seed the VAT at approximately 10 weeks of age in mice [4, 23]. Before VAT-residence, a subset of Tregs in the lymphoid organs upregulate expression of Treg-specific markers including PPARγ and Id2, while downregulating markers associated with conventional Tregs, such as Id3 [24–26]. Upon trafficking to the VAT, VAT Tregs undergo substantial clonal expansion, as evidenced by their restricted T cell receptor (TCR) repertoire [4, 23]. Detailed NextGen RNA Sequencing analysis has succeeded in describing a VAT Treg-specific gene expression pattern, as well as a transcriptional profile that defines tissue-resident Tregs more generally [25]. However, until recently, some critical questions remained unanswered: (1) what are the factors required to induce VAT Treg differentiation?, (2) how do VAT Tregs function in the adipose tissues?, and (3) why are VAT Tregs depleted under high-fat diet feeding conditions? Here we will consider recent data that begin to answer these questions.

Factors that induce VAT treg differentiation

T cell receptor (TCR) stimulation

By sequencing the TCR α and β complementarity determining region 3 s (CDR3s) in VAT Tregs and comparing these sequences to splenic Tregs, it was established that VAT Tregs undergo clonal expansion and have a distinct and limited TCR repertoire relative to conventional Tregs found in the spleen [4, 23, 25]. Given this distinctive and restricted TCR repertoire, as well as the Treg reliance on self-ligand stimulation for survival, it now appears likely that stimulation through the TCR by MHC II-expressing cells in the VAT are responsible, at least in part, for VAT Treg clonal expansion and survival [23]. Analysis of MHC II expression revealed that CD11b⁻ CD11c⁺ dendritic cells and CD11b⁺ CD11c⁺ macrophages within the VAT expressed the highest levels of MHC II and were in closest proximity to Tregs as assessed by flow cytometry and confocal microscopy [23]. Although adipocytes were previously reported to express MHC II for antigen presentation to conventional T cells [27], more recent data could not detect MHC II on adipocytes [23].

Examination of splenic Tregs in vitro also helped to reveal certain Treg characteristics associated with TCR signaling. TCR activation by stimulation of splenic Tregs with anti-CD3 and anti-CD28, and with the addition of IL-2 and IL-33, resulted in upregulation of VAT Treg-associated genes such as PPARγ and ST2 [28]. The addition of IL-4, IL-6, or IL-12, in contrast, did not result in ST2 upregulation on splenic Tregs [28]. Isolation of VAT Tregs ex vivo also revealed that these cells were constitutively high for Nur77, typically upregulated during TCR stimulation, and constitutively low for TCF7, typically downregulated in TCR-stimulated T cells [28]. Together, these data provide several important parameters for the expansion and survival of VAT Tregs: these cells are stimulated through their TCR and this TCR signaling is likely constitutive, driven by VAT-specific ligands presented by MHC II molecules on dendritic cells and macrophages in the VAT.

Recently, a TCR transgenic mouse model was generated from a known VAT-resident Treg clonotype (vTreg53) [29]. In this model, 98% of clonotype-positive VAT-resident Tregs expressed PPARγ, indicating that this clonally expanded population represented an enlarged VAT Treg population [29]. This model expanded the number of VAT Tregs available, eliminating one of the primary barriers to the investigation of VAT Treg function. Comparing adoptively transferred VAT-specific TCR transgenic Tregs to non-transgenic Tregs, it was observed that only the vTreg53 transgenic Tregs experienced significant clonal expansion in the VAT, indicating that TCR signaling is critical to VAT Treg expansion [29].

The question of which specific self-ligands VAT Tregs respond to has also recently been elucidated. Using a peptide library, the protein Fat 1562 was identified as initiating robust expansion of vTreg53 TCR transgenic Tregs in both in vitro and in vivo contexts [30]. Immunization of high-fat diet (HFD)-fed vTreg53 mice with Fat 1562 improved glucose tolerance and increased insulin sensitivity [30], indicative that TCR stimulation with cognate antigen may be useful for improving Vat Treg expansion or function.

Insulin

Through single-cell ATAC, paired single-cell RNA and TCR sequencing, two unique subpopulations of VAT Tregs were newly identified: cells that are CD73hi ST2lo and conversely CD73lo ST2hi [31]. CD73hi ST2lo Tregs expressed lymphoid tissue associated proteins, such as Tcf7, Foxo1, and Lrig1, and showed minimal clonal expansion [31]. In contrast, CD73lo ST2hi Tregs expressed non-lymphoid-associated genes such as St2, Klrg1, and Pparγ and demonstrated clonal expansion [31]. Intriguingly, the adipose tissue CD73hi ST2lo Treg subset expressed the insulin receptor, suggesting a role for insulin in VAT Treg differentiation [31]. Upon exposure to insulin, CD73hi ST2lo altered their transcriptional profile through a HIF1α-Med23 axis and upregulated PPARγ, differentiating into CD73lo ST2hi Tregs [31]. These data suggest that insulin is a key factor for VAT Treg differentiation and results in the transition from CD73hi ST2lo to CD73lo ST2hi VAT-resident Tregs.

Interleukin 2 (IL-2)

IL-2 and expression of CD25 (the high-affinity IL-2 receptor) by Tregs have long been described as critical for Treg homeostasis and survival in the lymphoid organs and in peripheral tissues [32–36]. In the lymphoid organs, other T cell populations likely provide IL-2 to support Treg homeostasis [35]. However, in the VAT, the source and availability of IL-2, and how IL-2 pertains to VAT Treg survival and homeostasis, are more nuanced and are still being clarified. Certainly, VAT Tregs require IL-2. Injection of mice with IL-2 complexes increased the frequency of VAT Tregs and decreased insulin resistance [4]. One cell type that has been clearly shown to produce IL-2 for VAT Treg homeostasis is the iNKT cell population (Figure 1). iNKT cells make up ~30% of lymphocytes in the VAT under normal conditions and like VAT Tregs, become depleted during obesity [12]. Lynch and colleagues have shown that iNKT cells produce high levels of IL-2 and can increase IL-2 expression upon stimulation in vitro and in vivo [11]. Stimulation of iNKT cells with their cognate antigen, α-galactosylceramide, resulted in increased IL-2 production by these cells and a corresponding increase in VAT Treg frequency [11]. Moreover, iNKT cell deficiency using CD1d-deficient or Jα18-deficient mice similarly resulted in reduced VAT Treg frequency, and confocal microscopy revealed iNKT cells and Tregs in proximity after treatment with α-galactosylceramide [11]. Thus, at least one source of IL-2 for VAT Treg survival and homeostasis has been described [11]. Interestingly, conditional loss of IL-2 signaling in Tregs resulted in short-term survival of Tregs in the lymphoid organs, in part due to persistent signaling through the IL-7 receptor (CD127) [37]. CD127 is certainly more highly expressed on VAT Tregs relative to splenic Tregs at the mRNA level [4], but whether IL-7 signaling plays a role in VAT Treg survival or homeostasis remains to be determined.

Figure 1.

Cytokine maintenance and production by VAT Tregs. Graphic showing positive regulation of VAT Treg homeostasis by IL-33, TCR stimulation, and IL-2. Cytokine production by VAT Treg showing Csf1, Csf2, IL-10, and TNFα production by VAT Tregs and potential conversion of ATP/ADP to AMP and adenosine by CD39 and CD73 on the surface of the VAT Tregs. VAT Treg-derived IL-10 signaling represses Ucp1 and other thermogenic genes in adipocytes while adenosine helps to promote thermogenesis. VAT Treg-derived Csf1 and Csf2 may help recruit M2 macrophages to the VAT. TNFα can act as a proinflammatory cytokine on adipocytes and M1 macrophages. Figure created with BioRender.com.

Interleukin 33 (IL-33)

Interleukin-33 (IL-33) is a member of the IL-1 family of cytokines and acts as a local tissue alarmin [38]. IL-33 is the most clearly described cytokine in terms of its effect on VAT Tregs. The receptor for IL-33, ST2 (Il1rl1) is highly expressed on VAT Tregs at the mRNA [39] and protein level [28, 40]. Initial descriptions of IL-33 and ST2 deficient mice determined that the frequency of VAT Tregs in these animals was reduced on standard chow and that glucose intolerance and the macrophage frequency in the VAT increased [23, 28]. These data indicated a VAT Treg dependency on IL-33 for survival and proliferation. ST2 expression on VAT Tregs also correlated with increased IL-10, GATA3, CCR4, and Nrp1 expression [40]. Additionally, when high-fat diet (HFD)-fed mice were injected with exogenous IL-33, the VAT Treg population expanded and glucose intolerance could be reduced [23, 28, 40], indicating that increased IL-33 is beneficial when considering obesity treatment.

Which cells then, are required to produce IL-33 to support VAT Tregs? Three publications have helped to shed light on this question. Identification of an IL-33-producing stromal VAT cell population showed these cells expressed Pdpn (podoplanin, gp38), CD26, and Cadherin 11, and intermediate or low levels of platelet-derived growth factor receptor-alpha (PDGFRα) [41] (Figure 1). More recently, a population of Pdpn⁺ PDGFRα⁺ Sca-1⁺ VAT mesenchymal stromal cells were identified as dominant IL-33 producers and conditional deletion of IL-33 using a PDGFRα-driven Cre line resulted in significantly reduced IL-33 in the VAT and corresponding decreases in the frequency of VAT Tregs [42], a phenotype that was similar to loss of ST2 expression in germline deficient mice [28, 42]. IL-33 expression by PDGFRα⁺ Sca-1⁺ VAT stromal cells was confirmed and the IL-33 produced by stromal VAT cells was shown to be critical for the activity of ILC2s and eosinophil maintenance in the VAT [43]. However, while IL-33 is required for VAT Treg homeostasis, it does not appear to be required for VAT Treg differentiation. Li et al. established that while insulin promotes differentiation of CD73hi ST2lo cells to CD73lo ST2hi cells, IL-33 is primarily required for the maintenance of the ST2hi population once they have differentiated [31].

Given the critical role of IL-33 in the maintenance of VAT Tregs, it was therefore somewhat surprising that IL-33 levels increase in mice on HFD [42]. Examination of the stromal VAT cells indicated that the IL-33 producing cells were slightly decreased in frequency after 4 weeks on HFD. However, by 16 weeks on HFD stromal VAT cells had returned to normal frequency, and some IL-33-producing stromal VAT cell subsets had increased in number and frequency [42]. Furthermore, injection of mice with IL-33 resulted in increased frequency of VAT Tregs as expected and reduced frequency of certain stromal VAT cell subsets responsible for IL-33 production. These discoveries have resulted in an updated model for the maintenance of VAT Tregs by IL-33 produced by stromal VAT cells. In this proposed model, under lean conditions, a negative feedback loop exists whereby stromal VAT cells produce IL-33 which helps to support VAT Tregs survival and proliferation and eventually results in repression through an as-yet-unknown mechanism of the stromal VAT cells by the VAT Tregs [42] (Figure 2). Under HFD and obese conditions, this negative feedback loop becomes perturbed, VAT Treg frequencies are reduced which results in decreased repression of stromal VAT cells by the VAT Tregs and thus increasing VAT stromal cells and IL-33 levels [42] (Figure 2). While this model helps to explain some of the phenomena observed, there remains a paradox regarding this hypothesis: if obesity causes an increase in IL-33 production from stromal cells, why does this increased IL-33 availability not support VAT Treg proliferation and homeostasis? Newer data may help to explain this paradox: an alternative soluble isoform of ST2 (sST2) has been discovered that retains the ability to bind IL-33 but lacks the transmembrane and intra-cellular domains required for membrane tethering [44]. In HFD fed mice, sST2 levels were elevated within the VAT and artificial overexpression of sST2 led to the reduction of the VAT Treg population and elevated glucose intolerance [44]. These data indicate that sST2 binds available IL-33 in the VAT, reducing its availability to maintain VAT Treg homeostasis. Mechanistically, it was shown that TNFα facilitates sST2 expression through activation of the NF-κB signaling pathway [44]. More recently, Li and colleagues showed that ST2+ VAT Tregs express high levels of sST2, a finding that indicates that VAT Tregs may limit their homeostasis [31]. Overall, these data suggest a new hypothesis to explain why VAT Tregs become depleted despite elevated IL-33 expression in HFD fed mice: increased decoy sST2 receptors in HFD-fed mice reduce the bioavailability of IL-33 to VAT Tregs, resulting in diminished VAT Treg numbers during obesity [44].

Figure 2.

IL-33 stromal VAT cells and VAT Treg feedback. Graphic showing negative regulation of stromal VAT cell accumulation by VAT Tregs and IL-33 positive regulation of st2⁺ VAT Tregs by stromal VAT cells under standard fat diet (SFD) and high-fat diet (HFD). Figure and data based on Spallanzani et al. [42]. Figure created with BioRender.com.

Adipokines: leptin and adiponectin

Aside from soluble mediators such as insulin, IL-2, and IL-33, the VAT microenvironment is replete with adipokines, hormones secreted by adipocytes, pre-adipocytes, and stromal cells that can modulate metabolism, satiety, and immune cells [45]. Of the hundreds of described adipokines, leptin and adiponectin are the best characterized [45]. Leptin regulates several different functions in the body, importantly acting on the hypothalamus to reduce appetite. During the development of obesity, leptin levels increase dramatically. Concerning VAT Tregs, obesity caused by a mutation in the leptin gene, the ob/ob mouse strain, resulted in decreased frequency of VAT Tregs although it has not been yet ascertained if this loss of VAT Tregs was due to loss of leptin expression or to increasing obesity and fat mass in these animals [4]. While the expression of the leptin receptor (LepR) has not been tested in VAT Tregs, it is reportedly high in human Tregs isolated from the PBL, relative to naïve T cells [46]. Additionally, antibody neutralization of leptin resulted in increased proliferation of human Tregs in vitro, resulting in the hypothesis that leptin is a negative regulator of Treg proliferation [46]. While this hypothesis helps to explain the loss of VAT Tregs observed during the development of obesity, it would not account for the loss of VAT Tregs observed in the ob/ob mice. Of course, it is possible that other cytokines and adipokines are also affected in the ob/ob mice, thus resulting in depletion of the VAT Treg population. Future work will hopefully elucidate these possibilities. More recent data showed that leptin could modulate the metabolism of effector T cells by upregulating glycolysis in these cells but in this model system, the Tregs were unaffected by leptin [47]. Again, the use of splenic Tregs precluded insight into the effect of leptin on VAT Tregs.

In contrast to leptin, adiponectin levels decrease in individuals with increasing obesity. Adiponectin signals through two receptors: adiponectin receptor 1 and 2, where adiponectin receptor 1 is expressed ubiquitously while adiponectin receptor 2 is specifically expressed in the liver. In terms of VAT Tregs, Adiponectin receptor 1 is highly expressed in VAT Tregs relative to splenic Tregs and its expression is negatively correlated with increasing VAT in mice [48]. Although the function of adiponectin on VAT Tregs is still being elucidated, it has been reported that loss of adiponectin in mice resulted in a decrease in the number and function of Tregs [49]. In the EAE model, adiponectin-deficient animals had worsened disease and decreased frequency of Tregs located in the spleen and CNS [49]. Additionally, expression of Foxp3 and IL-10 was reduced in these animals, and injection of exogenous adiponectin helped to restore Tregs and was protective from EAE [49]. Adiponectin may also indirectly regulate Treg frequency and number. In dendritic cells, the addition of adiponectin in vitro resulted in decreased expression of MHC II, CD80, CD86, and IL-12p40 on the DCs and a corresponding increase in Treg frequency in in vitro co-culture, indicating that adiponectin may act to modulate the immune system [50]. Additionally, adipose tissue-resident macrophages were shown, through adoptive transfer and ablation experiments, to be essential for the differentiation of PPARγ+ VAT Tregs, an effect that was negated in adiponectin deficient animals [51]. How adiponectin directly or indirectly regulates VAT Tregs in normal and obese conditions remains to be determined.

Function of VAT Tregs

IL-10 secretion

In addition to the expression of PPARγ and ST2, initial descriptions of VAT Tregs described these cells as substantial IL-10 producers [4, 28]. Regarding IL-10 and suppression of effector T cell proliferation, conventional Treg-derived IL-10 is a critical component of Treg-mediated suppression. One well-described positive regulator of IL-10 expression by Tregs is the transcription factor, B lymphocyte-induced maturation protein-1 (Blimp-1) [52–54]. However, Blimp-1-deficient VAT Tregs, which express less IL-10, are functional in terms of effector T cell suppression in vitro [23, 28, 54]. Indeed, the function of IL-10 signaling in adipose tissue and obesity has remained controversial. Two studies initially showed that germline loss of IL-10 did not affect insulin sensitivity, body weight, plasma triglycerides, or blood glucose [55,56]. However, IL-10 deficiency did result in increased triglycerides in the liver and increased plasma free fatty acids in HFD-fed mice [55,56]. These data, therefore, suggested that IL-10 was beneficial in preventing obesity-associated inflammation, in keeping with a known function for IL-10 in suppressing inflammation.

In contrast, more recent studies have shown that IL-10 deficiency improved insulin sensitivity and reduced the body weight of HFD-fed mice [57,58]. These data showed that IL-10 signaling in adipocytes resulted in reduced thermogenesis and energy expenditure (Figure 1), while loss of IL-10 in HFD-fed mice was protective from insulin resistance. It was shown that IL-10 signaling in adipocytes suppressed the expression of genes required for energy expenditure and thermogenesis in white adipocytes [58]. Although the specific source of IL-10 in these studies was not identified, the data did show that IL-10 producing cells originated in the bone marrow [58]. Likely reasons for the conflicting IL-10 data are the specific background strain(s) of the mice as well as subtle differences in the microbiome of each cohort. Indeed, the development of type I diabetes in non-obese diabetic (NOD) mice was shown to be critically dependent on the microbiota and IL-10 expression in the gut [59], although there are currently no data about how the microbiome might affect VAT Tregs in the VAT.

Through conditional Blimp-1 ablation in Tregs, it was confirmed that IL-10 expression in VAT Tregs is significantly reduced [52]. Moreover, Treg-specific deletion of Blimp-1 or IL-10 resulted in increased thermogenic gene expression in the adipose tissue depots in mice [52], supporting the data regarding the germline IL-10 deficiency phenotype outlined above. Physiologically, deletion of Blimp-1 or IL-10 in VAT Tregs resulted in reduced weight gain and improved insulin sensitivity in male mice fed a high fat diet [52]. In a separate earlier study, it was observed that the VAT Tregs from female mice express reduced IL-10 compared to their male counterparts [53]. Furthermore, female mice appeared to have improved insulin sensitivity relative to male mice but whether this is due to reduced IL-10 secretion by VAT Tregs was not directly investigated [53]. Curiously, in the same study Treg specific Blimp-1 deficient male mice were not protected from weight gain and insulin resistance, in contrast to the finding by Beppu and colleagues [52,53]. However, like the germline IL-10 studies discussed above, the conflicting data are likely due to differences in how and under what conditions the glucose tolerance and insulin tolerance tests were performed, as well as possible differences in the microbiota and background strains of the animals used in the different studies [52,53]. Without a doubt, however, IL-10 remains a critical component in the VAT Treg arsenal.

Adenosine

As mentioned above, a recent publication identified two distinct VAT Treg populations: immature CD73hi ST2lo Tregs and CD73lo ST2hi cells that are fully differentiated [31]. An intriguing aspect of the CD73hi ST2lo VAT Tregs is their ability to produce adenosine, which was shown to induce adipocyte thermogenesis, thus improving insulin sensitivity in mice [31, 60,61]. Ablation of Nt5e, the gene encoding CD73, in Tregs reduced the expression of thermogenic genes in adipocytes. Physiologically, Treg-specific deletion of CD73 resulted in increased insulin resistance and reduced body temperature upon cold challenge [31] (Figure 1). Taken together, adenosine expression by VAT Tregs represents a new, previously unappreciated, facet of VAT Treg biology.

Tnfα

Tumor necrosis factor α (TNFα) is a pro-inflammatory cytokine produced primarily by macrophages and monocytes. Previous data have indicated high TNFα mRNA expression in VAT Tregs relative to Tregs in the lymphoid organs [2, 39, 62]. In conventional Tregs, under certain inflammatory conditions, it has been reported that Tregs can produce TNFα For example, in humans with acute hepatitis A, Tregs isolated from these patients were better TNFα producers than Tregs from healthy donors, and acute hepatitis A patient-derived Tregs were less immunosuppressive [63]. Similarly, in patients with rheumatoid arthritis (RA), isolated Tregs could be induced to produce TNFα in vitro upon co-culture with inflammatory monocyte subsets [64]. A detailed analysis of TNFα by VAT Tregs has not yet been explored but our unpublished observations indicate that VAT Tregs can produce TNFα after stimulation with PMA and ionomycin in vitro. In terms of the potential effects of TNFα on cells of the VAT, TNFα has previously been described as a potent pro-inflammatory cytokine for adipocytes and can negatively modulate lipogenesis, thermogenesis and carbohydrate metabolism [65]. Additionally, TNFα can help to recruit and activate inflammatory (M1) macrophages to the adipose tissue, particularly under obese conditions [65]. It is intriguing to speculate on the function of TNFα secretion by VAT Tregs, particularly in the context of obese adipose tissue.

TGFβ, IL-35, Csf1 and Csf2

Although TGFβ and IL-35 are cytokines associated with conventional Treg suppressive activity, their expression specifically in VAT Tregs is low relative to splenic Tregs [2, 39, 62]. Surprisingly, Colony Stimulating Factors 1 and 2 (Csf1 and Csf2) are highly expressed at the mRNA level in VAT Tregs [2]. Csf1 and Csf2 are growth factors, typically required for their pleiotropic effects on macrophages [66], and Csf1 produced by local stroma and epithelial cells can support tissue macrophage homeostasis and differentiation [66,67]. Although the function of Csf1 and Csf2 produced by VAT Tregs is not yet known, one possibility is that VAT Tregs produce these growth factors to recruit IL-10-producing M2 macrophages to the adipose tissue. Future studies will no doubt help to elucidate the specific function of VAT Treg-derived Csf1 and Csf2.

Loss of VAT Tregs during obesity

One aspect of VAT Treg biology that has long been perplexing is the relative paucity of VAT Tregs after high-fat diet feeding in mice and humans [4]. New data by Li and colleagues have helped to clarify how this process occurs [68]. Investigating transcriptional changes that occur in the VAT after long-term high fat diet feeding, it was shown that Interferon-α (IFNα) is substantially increased and that the addition of exogenous IFNα reduces the frequency of VAT Tregs by increasing VAT Treg cell death [68]. Deletion of the IFNα receptor on Tregs protected VAT Tregs from cell death during high fat diet feeding and resulted in improved insulin sensitivity in mice [68]. Further investigation revealed that plasmacytoid dendritic cells (pDCs) become enriched in the VAT during high fat diet feeding and are the primary producers of IFNα [68]. Together these studies help to explain why VAT Tregs become during the process of weight gain and obesity.

Conclusion

Here we have reviewed the soluble proteins that are critical for VAT Treg differentiation and function, and the known cytokines produced by VAT Tregs that likely influence the VAT microenvironment. The function of VAT Tregs in regulating local inflammation and VAT-specific cells is still an active area of research. Similarly, how VAT Tregs are maintained in the VAT, is of crucial interest, particularly as VAT Treg regulation can be insightful in considering treatments for obesity and its associated morbidities. Since their discovery, VAT Tregs have been considered as positive mediators in the VAT, responsible for protection from inflammation and dampening effector T cell responses under homeostatic conditions. However, VAT Tregs may not always be beneficial. A seminal study showed that VAT Tregs could be detrimental in older mice [69] and that depletion of VAT Tregs using an ST2-depleting antibody was beneficial in preventing insulin resistance [69]. Thus, VAT Tregs may provide tissue homeostasis benefits in the adipose tissue under steady-state conditions but upon perturbation due to age, obesity, or even infection, the positive effects of VAT Tregs are less clearly defined. Future studies into VAT Tregs in different disease states and conditions will be critical to fully understand the function of these cells and to provide insight into how therapeutics can modulate VAT Tregs for protection from obesity and metabolic disease.

Abbreviations:

Tregs

Regulatory T cells

VAT Tregs

visceral adipose resident regulatory T cells

VAT

visceral adipose tissue

trTregs

tissue-resident regulatory T cells

SLOs

secondary lymphoid organs

TCR

T cell receptor

MHC II

major histocompatibility complex class II

CDR3s

TCR α and β complementarity determining region 3s

Availability of data and material

All data generated or analyzed during this study are referenced in this article.