Immune Cells Gate White Adipose Tissue Expansion

Abstract

The immune system plays a critical role in white adipose tissue (WAT) energy homeostasis and, by extension, whole-body metabolism. Substantial evidence from mouse and human studies firmly establishes that insulin sensitivity deteriorates as a result of subclinical inflammation in the adipose tissue of individuals with diabetes. However, the relationship between adipose tissue expandability and immune cell infiltration remains a complex problem important for understanding the pathogenesis of obesity. Notably, a large body of work challenges the idea that all immune responses are deleterious to WAT function. This review highlights recent advances that describe how immune cells and adipocytes coordinately enable WAT expansion and regulation of energy homeostasis.

The prevalence of obesity doubled in the last 25 years (1), increasing the burden of care on medical centers and government (2). The epidemiologic associations of obesity with type 2 diabetes mellitus (T2DM) and cardiovascular disease are unequivocal, but detailed mechanisms accounting for these links are not well understood. Nonetheless, it is clear that obesity promotes chronic alterations in energy storage and utilization resulting in lipid deposition in nonadipose tissues, insulin resistance, and T2DM.

White adipose tissue (WAT) is an endocrine organ that dynamically expands and contracts to meet the metabolic demands of the organism. WAT also secretes peptides, hormones, and metabolites that contribute to insulin sensitivity in other peripheral tissues. WAT mass characterizes obesity and correlates with a strong predisposition for insulin resistance, T2DM, and cardiovascular disease. Adipocytes remain the singular cell type capable of sequestering lipids and protecting the periphery from lipotoxicity.

Subcutaneous (peripheral) and visceral (central) WAT depots broadly constitute the bulk of adipose tissues in adults. In humans, the visceral (omental) fat resembles mouse epididymal fat based on gene expression profiling, inflammation, and expandability, despite anatomical differences; subcutaneous fat depots are anatomically and functionally similar in mice and humans (3–7). Excess calorie intake evokes WAT expansion through both increased adipocyte size (hypertrophy) and number (hyperplasia). Hyperplasia has been linked to increased gene expression of transcriptional regulators essential for adipose tissue formation, such as peroxisome proliferator–activated receptor γ (PPARγ) (8). In chronic states of positive energy balance, such as consuming a Western diet, subcutaneous WAT differentiation becomes impaired (9) and visceral WAT expands (10–13), driving metabolic maladaptation.

Increased hypertrophy is a hallmark of WAT enlargement in obesity and is typically associated with metabolic alterations, proinflammatory response, and increased risk of developing T2DM independent of total fat mass (14–16). Experimental observations indicate that larger, hypertrophic fat cells behave differently than do smaller, hyperplastic adipocytes, namely in responses to lipolytic stimuli, secretory functions, and the anabolic effects of insulin. Consequently, the failure of integral lipid metabolism and insulin sensitivity effectors reflects low subcutaneous adipocyte differentiation and thus limits WAT expandability. Subcutaneous WAT safely sequesters excess energy and promotes insulin sensitivity in the face of diet-induced obesity, though the molecular mechanisms underpinning these observations remain incompletely understood.

Distinct Precursor Cells and Microenvironments Contribute to WAT Mass

WAT depots develop in a spatiotemporal manner, which likely engenders functional identity. Adipocyte progenitor cells express overlapping progenitor and mesenchymal cell surface markers, including platelet-derived growth factor receptor-α (PDGFR-α), CD29, CD34, stem cell antigen-1 (SCA1), and CD24 (17–20). Mural and smooth muscle–related cells that express Acta2, Myh11, or Pdgfrb also contribute to adipocyte formation under certain conditions (21–25). Lineage tracing studies performed in mice argue that subcutaneous and intra-abdominal depots emanate from distinct lineages (26). Visceral adipocytes descend from cells expressing the mesothelial cell marker Wilms tumor 1 (26, 27), whereas subcutaneous adipocytes can be marked with paired related homeobox 1 (PRRX1)–Cre (28, 29) and myxovirus 1 (Mx1)–Cre transgenes (6). Adding to the complexity, anatomically distinct depots may contain a network of adipocytes and other stromal precursor cells that express molecular and secretory programs to restrict or enable responses to dietary challenges (7, 30–35). Such functional cell–cell interactions likely exert stop-and-go signals for WAT development and metabolic plasticity.

Of note, murine and human adipose tissue depots perform physiological and endocrine behaviors based on distinct anatomical locations. Modern molecular and single-cell approaches have partly revealed the specific origins of subcutaneous and visceral adipocytes in mice and humans (30, 34, 36, 37). A few studies indicate that specific anatomical niches provide adipose precursors in various human tissues (38–40). Recent work discovered that DPP4⁺ interstitial progenitors give rise to committed populations of preadipocytes poised to undergo adipocyte differentiation (36). These cells reside in the fluid-filled collagen network of collagen and elastin fibers that surround adipose tissues and many other organs. Other efforts identified distinct populations of adipocyte progenitor cells in human WAT that vary in endocrine function and anatomical distribution by differential CD34 expression (37). These types of studies ultimately elaborate the specification of anatomical fat depots and how fat cells interpret microenvironment cues.

Obesity as an Inflammatory Disease

Obesity-induced chronic inflammation in adipose tissues contributes to the manifestation of insulin resistance and T2DM. However, the precise triggers of obesity-associated inflammation remain poorly characterized. Numerous mechanisms have been investigated in rodent models of dietary and genetic obesity. It is likely that the trigger of inflammation in adipose tissue originates from the anabolic pressure of positive energy balance. The catabolic inflammatory response alleviates anabolic pressure and supports the expansion of adipose tissues to meet the need for increased lipid storage. However, over time, the persistent stress of obesity permanently skews the reparative immune response and new thresholds for adipose tissue expansion cannot be met. This concept suggests that the insults that ultimately constrain fat cell expandability must be buffered appropriately to counter the energetic demands of dietary stress.

Many early observations in humans corroborate links between inflammation and T2DM. The initial observations noted that patients with meningitis also exhibited transient hyperglycemia (41). A large volume of studies in humans continue to underscore the importance of immune cells in T2DM [reviewed in (42)]. Although most of the direct evidence linking chronic inflammation to the comorbidities of obesity stems from rodent studies, primary human cells ex vivo support the notion that inflammation disrupts the metabolic flexibility in adipocytes (43–50).

The causal role for inflammation in the etiology of obesity and T2DM remains unclear in humans. Genome-wide association studies mostly identify loci that predict T2DM risk near genes critical for insulin secretion and processing (51, 52). The lack of T2DM risk loci that enrich in immune pathways may weaken the genetic links between obesity and inflammation. However, human genetic variability likely only partly contributes to the heritability of complex traits. Many questions remain unanswered, including how to restore the endocrine and anti-lipotoxic functions of WAT during excess nutrient intake.

WAT expandability and nutrient storage are closely linked to hypoxia. As reviewed elsewhere, abundant evidence suggests that adipocytes experience hypoxic stress that triggers local inflammatory signatures (53). Hypoxia develops as adipose tissue grows in size due to lack of tissue perfusion, mechanical stress, or increased oxygen consumption (54). The ensuing hypoxia activates hypoxia-inducible factor 1 (HIF1), which stimulates transcription of numerous inflammatory genes and chemokine release. HIF1 knockout (KO) prevents obesity-induced inflammation and insulin resistance (55), which firmly demonstrates that HIF1 mediates maladaptive responses to hypoxia in adipose tissues. Likewise, immunostaining of WAT from obese rodents and humans with obesity reveals that regions of hypoxia correlate with macrophage infiltration and fat cell necrosis (56, 57). Despite the evidence that links hypoxia and inflammation in WAT, it remains uncertain whether hypoxia is a consequence of adipose tissue expansion or a direct causative contributor to obesity-associated metabolic disease.

Hypoxia and inflammation likely foster fibrosis to restrict WAT expandability. During early WAT expansion, hypoxia induces HIF1-regulated profibrotic genes associated with increased inflammation, insulin resistance, and adipocyte cell hypertrophy (58). Notably, subcutaneous WAT fibrosis is negatively correlated with the effectiveness of weight loss following gastric bypass surgery (53). Clinical studies show a strong correlation between increased subcutaneous WAT fibrosis with insulin resistance and obesity (59–61). Conversely, inhibition of fibrosis through collagen VI or HIF1 KO reduces inflammation, glucose intolerance, and adipocyte cell size (62, 63). Hasegawa et al. (64) similarly demonstrated that repression of WAT fibrosis improves glucose homeostasis and insulin sensitivity. Fibrosis is often viewed as a mere consequence of obesity, but these studies suggest that fibrosis may be a pathogenic factor that restricts WAT expansion in obesity and degrades insulin sensitivity.

Nutrient excess and activation of the immune response coincides with the pathogenesis of obesity and its comorbidities, including fatty liver, insulin resistance, T2DM, and cardiovascular disease. Moreover, obesity-induced inflammation involves multiple metabolic and endocrine organs, including WAT, pancreas, skeletal muscle, heart, and brain. Although obesity-induced inflammation occurs in many tissues, WAT remains a primary site for complex skewing of the immune system. The low level of persistent WAT inflammation in obesity should not be confused with acute infection responses. Successful immune responses require short-lived reactions necessary for survival of the organism. In contrast, the nature of obesity-induced inflammation involves sustained low-level activation of the immune response. Macronutrient overload ultimately leads to compromised WAT function accompanied by infiltration of proinflammatory cells, including fibrogenic cells (30), neutrophils (65), M1 macrophages (66), CD8⁺ lymphocytes (67, 68), helper T cells (69), and adipogenesis regulatory cells (34). Supporting these observations, proinflammatory mediators secreted by immune cells and other fibroinflammatory progenitors, such as interferon γ (IFNγ) and TNFα, correlate with insulin resistance and central WAT accumulation (45, 46, 70, 71).

Macrophages comprise up to 40% of all stromal vascular cells (72) and supply inflammatory cytokines that disrupt homeostatic WAT function. Recruitment of polarized M1 macrophages defines the adipose tissue inflammation that accompanies obesity. Although the spectrum of polarization varies across adipose tissue volume (73–75), the term M1 depicts the proinflammatory state of recruited adipose tissue macrophages that express CD11c. In obesity, free fatty acids (FFAs) promote polarization of adipose tissue M1 macrophages, which secrete factors and perform functions that can block insulin action. Indeed, numerous studies demonstrate that genetic alterations that deplete the function of M1 macrophages protect against obesity-induced insulin resistance and glucose intolerance (76). Many other immune cell types, including dendritic cells, mast cells, eosinophils, and lymphocytes, also contribute to the impact of inflammation in WAT. Despite the numerous discoveries linking these cells to the metabolic profile of obesity, the tissue-dependent and pleiotropic functions of the immune system foster inherent challenges that slow development of therapies to minimize the dysfunction caused by diet-induced inflammation.

Immune Signals Mediate Adipose Tissue Expansion

WAT expansion in response to excess nutrients depends greatly on a microenvironment composed of immune cells, blood vessels, and stromal cells. In particular, crosstalk between adipocytes and immune cells involves an intricate signaling network to modulate inflammation and consequently whole-body energy homeostasis. Broadly, insulin sensitivity is preserved when WAT expansion maintains an anti-inflammatory state, primarily through the actions of M2 macrophages (77), innate lymphoid type 2 cells, and regulatory T cells (Tregs) (78). Tregs play a prominent role in WAT to promote an anti-inflammatory environment (79). In contrast, several mouse models of obesity [ob/ob, high-fat diet (HFD), A^y/a, New Zealand obese mice] exhibit dramatically reduced numbers of Tregs in epididymal WAT (eWAT) (79–81). Treg depletion in obesity correlates with infiltration of WAT by Th1 CD4⁺ T cells, cytotoxic CD8⁺ T cells, and classically activated macrophages (M1), leading to a proinflammatory milieu that is highly associated with obesity and insulin resistance (79, 81, 82). These observations highlight the complex relationship between Treg suppressive functions, cytokines, and insulin sensitivity in adipocytes.

The type 1 immune phenotype in obesity derives from M1 macrophages, cytotoxic CD8 T cells, CD4 type 1 helper T cells, natural killer cells, and B cells (Fig. 1). Type 2 immunity in WAT reflects the behavior of M2-like macrophages, eosinophils, group 2 innate lymphoid cells (ILC2s), CD4 type 1 helper T cells, and Tregs. Type 1 inflammation in WAT can be counteracted by the immunosuppressive cytokine IL-10, secreted from Tregs and other type 2 immune cells. eWAT Tregs produce higher levels of IL-10 compared with Tregs from lymphoid tissues and exhibit enhanced IL-10 signaling (79). IL-10 suppresses M1 macrophage polarization and prevents macrophage recruitment by decreasing the adipocyte-derived chemokine MCP-1 (83, 84). Notably, systemic overexpression of IL-10 in mice reduces weight gain and improves insulin sensitivity with fewer eWAT macrophages (85). IL-10 also protects against TNFα-induced expression of inflammatory genes (MCP-1, IL-6, MMP-3, SAA-3, RANTES) in cultured mouse and human adipocytes (72, 79).

Figure 1.

White adipose immune cell composition in obesity. (A) WAT from lean individuals is primarily composed of Tregs, alternatively activated M2 macrophages, ILC2s, and eosinophils that suppress proinflammatory responses and support insulin-sensitive adipocytes. In contrast, adipocytes from individuals with obesity are hypertrophic and insulin resistant due, in part, to infiltrating B cells and various T cells (CD4⁺ Th1, CD8⁺, and natural killer), as well as differentiation and polarization of M1 macrophages with reductions in Tregs and ILC2s. (B) The inset table summarizes skewing of immune cells in obesity. This proinflammatory environment impinges on adipocyte insulin signaling, adipocyte hyperplasia, and function.

IL-10 and forkhead box protein p3 (Foxp3) show reduced expression in both rodent and human eWAT during obesity (79, 80, 86, 87). Decreased Foxp3 expression and Treg function can be attributed, at least in part, to proinflammatory cytokine secretion by adipocytes and infiltrating CD4⁺ Th1 T cells, causing an altered Treg phenotype marked by upregulation of IFNγ (79, 80, 86, 87). This shift in Treg phenotype might promote Treg fragility (88) and may underlie some loss of eWAT Tregs in obesity. IL-2 treatment of HFD-fed mice increased eWAT Tregs and IL-10 expression, although adoptive transfers of eWAT Tregs into obese mice have been limited by isolation of sufficient Tregs, among other technical challenges (79). Induction of Tregs by anti-CD3 and β-glucosylceramide administration reduced WAT inflammation and restored insulin sensitivity in ob/ob mice (89). Although these studies suggest that expansion of IL-10–producing Foxp3⁺ Tregs in eWAT might be an effective therapeutic strategy, further studies are needed to understand how Tregs and adipocytes communicate and mediate metabolic effects.

Eosinophils and ILC2s support Treg suppression of type 1 immunity. ILC2 secretion of IL-5 stimulates the maturation and infiltration of eosinophils, and both cell types promote M2 macrophage polarization through type 2 cytokines IL-4 and IL-13 (90–92). However, HFD-induced obesity reduces ILC2s and eosinophils in mice, whereas ILC2s are decreased in WAT of humans with obesity (90, 92–94). Activation of the ILC2–eosinophil axis induces metabolic effects related with subcutaneous WAT expandability and “browning,” which may involve immune cell secretion of peptides and other hormones (93–96).

IL-33 secretion from adipocytes and stromal cells activates ILC2s (90, 94, 95, 97, 98). Interestingly, IL-33 also acts directly on eWAT Tregs through the ST2 receptor (Il1rl1) (81). IL-33–deficient mice show reduced numbers of Tregs specifically in eWAT, indicating the significance of IL-33 to maintain eWAT Tregs. IL-33 administration in HFD and New Zealand obese mice restored eWAT Tregs with increased Foxp3 and PPARγ expression, resulting in improved glucose tolerance. One study suggested that ILC2s are not required for IL-33–induced Treg effects (99); in contrast, Brestoff et al. (93) demonstrate IL-33–elicited ILC2 induction of uncoupling protein 1 (UCP1)⁺ beige adipocytes in subcutaneous WAT from mice lacking adaptive immune cells. Although Tregs have not been shown to promote recruitment of UCP1⁺ adipocytes in subcutaneous WAT, Tregs and ILC2s interact through ICOS and ICOSL costimulatory molecules, which are inhibited by IFNγ (100). Perhaps the ILC2–eosinophil axis and Tregs act as integrated or partly redundant systems to suppress type 1 inflammation and enable WAT expandability and insulin sensitivity.

Treg-Derived Signals Promote Adipocyte Insulin Sensitivity and Metabolism

In addition to Treg-mediated suppression of proinflammatory immune cells, Tregs might also directly affect adipocyte function and metabolism. Depletion of Tregs in Foxp3–diphtheria toxin receptor mice worsens the inflammatory profile and degrades insulin action in eWAT (79). Most studies suggest that Tregs likely promote insulin signaling in adipocytes via IL-10–directed repression of inflammatory cytokine synthesis (72, 79). Seminal work by Hotamisligil and many others demonstrated that inflammatory cytokines (TNFα, IL1β, and IFNγ) interfere with insulin receptor signaling in adipocytes (43, 44, 72, 101, 102). Accordingly, TNFα and IFNγ lower Glut4 expression in adipocytes, contributing to reduced insulin-stimulated glucose uptake (72, 79). However, treatment of adipocytes with IL-10 enhances insulin-stimulated glucose uptake and antagonizes TNFα blockade of insulin receptor signaling and glucose uptake (72). These studies suggest an important role of Tregs to modulate adipocyte function through IL-10 signaling.

Several studies, however, established divergent effects of IL-10 on metabolism and insulin resistance in IL-10 KO mice (103–106). A recent study demonstrated that injection of IL-10 receptor (IL-10Ra)–targeted antisense oligonucleotides decreased body weight and fat mass in chow-fed mice and increased UCP1 expression in subcutaneous WAT (107). Accordingly, IL-10 treatment of subcutaneous WAT ex vivo suppressed UCP1 and other thermogenic genes and oxygen consumption. As a direct target of PPARγ, IL-10Ra expression likely coincides with adipogenesis and other critical insulin sensitivity genes in adipocytes (107). It is also unclear whether other type 2 cytokines (such as IL-4, IL-13) and associated signaling pathways might be altered in the absence of IL-10. Treg depletion studies in vivo and IL-10–stimulated adipocytes in vitro present (72, 79, 108) a strong argument for their role in adipocyte metabolism and insulin sensitivity; however, the adipose microenvironment is a diverse network of cells and signals that requires further study given recent findings.

Inter-Cell Crosstalk Performs Metabolic Functions in WAT

While cytokines comprise a substantial portion of intercellular communication within the WAT microenvironment, adipokines and extracellular vesicles also play a role. In particular, Tregs express receptors for leptin, adiponectin, and FFAs (109–113). The suppressive actions of leptin on Treg function and proliferation have been well described (110, 111, 114). De Rosa et al. (110) demonstrated that in vitro stimulation of activated human Tregs reduced proliferation and the ability to suppress CD4⁺ T effector cells, which was partially restored with addition of a leptin monoclonal antibody. In vivo, Treg proliferation and Foxp3 expression increased with leptin monoclonal antibody administration in wild-type mice, although transfer of wild-type Tregs into ob/ob leptin-deficient mice also exhibited a higher degree of proliferation. The nutrient-sensing mTOR signaling cascade mediates leptin effects in Tregs while leptin receptor (db/db)–deficient Tregs exhibited reduced mTOR signaling and higher proliferative capacity (111). In contrast, leptin stimulates proliferation of CD4⁺ Th1 T cells coupled with increased IFNγ expression (110, 114, 115). Thus, elevated leptin production during obesity might contribute to Treg depletion and expansion of CD4⁺ Th1 T cells, thereby altering the inflammatory environment in WAT.

Adiponectin production by adipocytes exerts anti-inflammatory, antidiabetic, and cardioprotective effects (116–118). Adiponectin KO mice have fewer Tregs in eWAT (119) and, although not directly tested to date, might induce IL-10 production by Tregs, as observed in macrophages (79, 120). Ramos-Ramírez et al. (112) showed that eWAT-resident Tregs expressed higher levels of adiponectin receptor 1 than did Tregs in the spleen, and its expression on adipose tissue Helios⁺ Tregs correlated negatively with eWAT mass. Thus, it is possible that the inverse relationship between adiponectin and leptin reflects the dynamic regulation of Tregs by adipocytes during changes in the nutritional status of lean individuals and those with obesity.

FFAs derived from WAT may also contribute to peripheral insulin sensitivity (121). For example, palmitoleate treatment of eWAT-derived adipocytes reduces proinflammatory cytokine expression (MCP-1, TNFα). Interestingly, anti-inflammatory IL-4 may also induce lipolysis, in addition to dampening of WAT inflammation and mitigating effects of diet-induced obesity (122, 123). FFAs are the preferred substrate for functional Tregs (114, 124, 125) and therefore might secrete lipolytic cytokines to increase the local concentration of FFAs for use as fuel (122). Kahn and colleagues (126) identified a new class of lipids called fatty acid esters of hydroxy fatty acids, which include palmitic acid–hydroxystearic acids (PAHSAs) abundantly found in mouse serum, WAT, brown adipose tissue, and, to a lesser extent, liver. PAHSAs correlate with insulin sensitivity in serum and subcutaneous WAT in humans. PAHSAs also regulate insulin-stimulated glucose uptake through GLUT4 translocation in adipocytes in vitro and decrease inflammatory cytokine production by macrophages in HFD-fed mice. Future studies should help illustrate how specific classes of lipids act on Treg metabolism and function, and how they might impact Tregs during obesity.

Aside from adipokines, local insulin levels are abundant in WAT. Whereas insulin acts on adipocytes to suppress lipolysis and promote energy storage, insulin stimulation of Tregs decreases IL-10 production and reduces the ability to suppress CD4⁺ Th1 T cells in vitro (86). In obese, hyperinsulinemic mice, elevated serum insulin was associated with decreased Foxp3⁺ Tregs in the eWAT, expressing lower levels of IL-10 with a switch toward increased expression of the type 1 cytokine, IFNγ (86). A recent report showed a moderate reduction in insulin production (Ins1^−/−;Ins2^fl/+;Pdx1-Cre^ER mice) induced weight loss on an HFD, mainly affecting eWAT mass, without altering glucose tolerance (127). It will be interesting to consider how eWAT Tregs and other immune cell populations might vary in the setting of reduced insulin. However, anecdotal evidence suggests that Treg-specific insulin receptor deletion might not impact glucose tolerance (128). Thus, although reducing insulin appears important for adipocyte metabolism, it might not be a key signal during the obesity-related shift to a proinflammatory environment and the suppression of Treg function.

Another potential signaling mechanism between immune cells and adipocytes may occur through extracellular vesicles, such as exosomes. Local and circulating exosomes carry cargo-containing proteins, RNA, and particularly miRNAs that can affect function of acceptor cells [reviewed in detail in (129)]. Multiple studies have shown that adipocyte-derived circulating exosomes regulate gene expression and insulin sensitivity of peripheral tissues (130–133). Mainly, miRNAs within circulating exosomes from obese mice confer glucose intolerance and insulin resistance when transferred to lean mice (130, 134). Obese adipose tissue secretes twice as many exosomes as do lean mice, which induce differentiation of bone marrow–derived cells into macrophages (135). Similar findings were observed with human adipocyte exosomes, which induced differentiation of monocytes into macrophages that, in turn, could reduce adipocyte insulin signaling (136). Adipose-derived exosomes from ob/ob mice also activate circulating macrophages marked by increased serum IL-6 and TNFα levels (131). In contrast, adipose-derived stem cell exosomes from lean mice induced M2 macrophage polarization in obese mice associated with improved insulin resistance and hepatic steatosis (137). Additionally, macrophages within adipose tissue also secrete exosomes that influence glucose tolerance and insulin sensitivity. Exosomes derived from macrophages in lean mice improve insulin sensitivity, whereas macrophages collected from obese mice secrete exosomes that promote insulin resistance (134). Of note, miR-155 within macrophages targets PPARγ, which might lead to restricted WAT expansion (134).

We know less about how exosomes affect Treg function in obesity and metabolic diseases. In a mouse model of type 1 diabetes, treatment with exosomes derived from adipose tissue mesenchymal stem cells increased the percentage of splenic Tregs, along with increased type 2 cytokines IL-4, IL-10, and TGF-β (138). In support of these observations, human mesenchymal stem cell exosomes induced differentiation of Tregs with enhanced suppressive capacity (139, 140). Studies have also shown in various contexts that Tregs generate exosomes (141–144). Okoye et al. (142) demonstrated that Tregs produce miRNA-containing exosomes dependent on Dicer (required for biogenesis of most miRNAs) expression and canonical exosomal extrusion pathways. The authors subsequently showed that Treg exosomes suppress Th1 T cell proliferation and IFNγ production through let-7d. Treg-derived exosomes also skew dendritic cells toward a type 2 phenotype, marked by increased IL-10 and decreased IL-6 production, which involved transfer of miR-150-5p and miR-142-3p (144). Whether resident adipose Tregs exhibit similar exosomal suppressive activity upon type 1 immune cells during obesity will be an important question in the emerging area of the adipose tissue exosomes. Additionally, Treg exosomes might also influence adipocyte-specific inflammatory pathways, such as IFNγ and TNFα, to modulate insulin sensitivity. Meanwhile, adipocyte-derived exosomes carry lipids (135) that could influence Treg metabolism and function.

In summary, the adipose microenvironment involves complex and delicate signaling between adipocytes and immune cells. In insulin-sensitive WAT, Tregs produce IL-10 to suppress CD4⁺ Th1 T cell function, polarize M2 macrophages, and repress adipocyte-derived inflammatory cytokines (Fig. 2). IL-33 derived from adipocytes and stromal cells activates Tregs and ILC2s. ILC2s recruit and activate esosinophils, and in combination, they produce type 2 cytokines that polarize M2 macrophages. Adipocytes secrete adiponectin and FFAs, which also suppress macrophage cytokine production and might influence Treg function, whereas the relative absence of leptin allows Tregs to proliferate and maintain suppressive activity. Both adipocytes and Tregs produce exosomes, with adipocyte-derived exosomes influencing macrophage differentiation and function, although Treg exosomes suppress T cells in some contexts. Lastly, adipocytes secrete fatty acid esters of hydroxy fatty acids, such as PAHSAs, that regulate macrophages and peripheral insulin sensitivity. In contrast, obese insulin-resistant adipocytes secrete inflammatory cytokines (IFNγ, IL-6, RANTES, SAA) and recruit M1 macrophages (via MCP-1). Tregs, ILC2s, and eosinophils are reduced, with Tregs shifting toward IFNγ production. High levels of leptin and insulin also reduce Treg proliferation, function, and fatty acid metabolism. Accumulation of CD4⁺ Th1 T cells and M1 macrophages contributes to the increasing inflammatory milieu that restricts adipocyte insulin signaling, function, and expansion. Obesity-derived adipocyte or macrophage exosomes also reduce insulin sensitivity. Notably, these observations largely focused on immune cell–adipocyte signaling in eWAT given its propensity for greater immune cell infiltration in obese mice; however, some studies are beginning to uncover roles of immune cells in subcutaneous WAT and the signals that might preserve insulin sensitivity and WAT expansion (43, 145). Finally, the complexity of the adipose tissue microenvironment suggests that therapeutic intervention will likely require targeting of multiple signals and cell types to restore adipocyte expansion and insulin sensitivity in individuals with obesity.

Figure 2.

Inter-cell crosstalk regulates metabolic and inflammatory functions in WAT. In lean WAT, insulin-sensitive adipocytes expand appropriately to anabolic pressure and secrete adipokines (adiponectin), FFAs and PAHSAs, and exosomes. Tregs produce IL-10 to suppress CD4⁺ Th1 T cell function, polarize M2 macrophages, and repress adipocyte-derived inflammatory cytokines. IL-33 derived from adipocytes and stromal cells activates Tregs and ILC2s. ILC2s recruit and activate esosinophils and, in combination, they produce type 2 cytokines that polarize M2 macrophages. Adipocytes secrete adiponectin and FFAs, which also suppress macrophage cytokine production and might influence Treg function, whereas the relative absence of leptin allows Tregs to proliferate and maintain suppressive activity. Both adipocytes and Tregs produce exosomes, with adipocyte-derived exosomes influencing macrophage differentiation and function, whereas Treg exosomes suppress T cells in some contexts. Lastly, lean adipocytes secrete fatty acid esters of hydroxy fatty acids, such as PAHSAs, that regulate macrophages and peripheral insulin sensitivity. In summary, type 2 immune cells promote adipocyte insulin sensitivity, differentiation, and function through cytokines IL-4, IL-10, and M2 macrophage-derived exosomes. In obese adipose tissue, insulin resistance develops, and adipocyte differentiation is restricted, which contributes to ectopic lipid deposition. Adipocytes secrete inflammatory cytokines (IFNγ, TNFα, IL-6, RANTES, SAA) and recruit M1 macrophages (via MCP-1). Tregs, ILC2s, and eosinophils are reduced, with Tregs shifting toward IFNγ production. High levels of leptin and insulin also reduce Treg proliferation, function, and fatty acid metabolism. Obese adipocyte- or macrophage-derived exosomes also reduce insulin sensitivity. Collectively, accumulation of CD4⁺ Th1 T cells and M1 macrophages contributes to the increasing inflammatory milieu (IFNγ, TNFα, IL-6, and exosomes) that restricts adipocyte insulin sensitivity, expansion, and function.

Future Perspectives

More than 20 years of detailed mechanistic and physiological studies firmly establish that chronic inflammation in WAT depots leads to systemic glucose and lipid dysregulation. Although WAT inflammation is one conserved pathway that links obesity to insulin resistance and T2DM, numerous questions remain unanswered, including how to sustain healthy noninflamed subcutaneous WAT expansion during excess nutrient intake. For people with T2DM, we are hopeful that the immune system will be leveraged as a tool to maintain insulin sensitivity. However, targeting the immune component of obesity has proved elusive. For example, antibodies that neutralize TNF improve insulin sensitivity in obese mice, but humans showed no metabolic improvements (146). In clinical trials, anti-inflammatory salicylates improved glycemia and adipose inflammation profiles in individuals with obesity and T2D, but insulin sensitivity was unaffected (147–149). New strategies that target higher baseline inflammation in adipose tissues appear promising (150) in small clinical trials. Note that several well-established antidiabetic drugs influence anti-inflammatory endpoints. Metformin inhibits reactive oxygen species production and the production of numerous inflammatory cytokines (151, 152). Thiazolidinediones also exert anti-inflammatory effects in macrophages and adipocytes (153). In both cases, these drugs block nuclear factor κB (151, 154) and restore the M2 phenotype of macrophages (134, 155). Further exploration into how biologics targeting inflammation in rheumatoid arthritis and Crohn disease (156) may reveal therapeutic vulnerabilities for obesity. A thoughtful survey of how manipulating the immune response impacts physiology is warranted. Importantly, note that general disruption of inflammatory pathways may compromise immune responses, but also result in tissue damage, dysbiosis, and amplification of other autoimmune conditions (157).

The anatomical and functional differences of WAT depots portend distinct immune cell niches that contribute to adipocyte function and expansion during obesity. The subcutaneous WAT has a tremendous capacity to expand and store excess nutrients, as demonstrated by MitoNEET transgenic mice (158, 159). In addition to varied WAT depots, a recent study suggested that beige fat may also exist in multiple distinct forms, conventional and glycolytic beige fat (160). The immune cell composition also greatly varies, with the eWAT containing a greater abundance of immune cells, even during metabolic homeostasis, in contrast to the less infiltrated brown adipose tissue. The tissue- and depot-specific roles for immune cells is unknown, but there is clearly an extensive communication network between adipocytes and immunity through multiple signaling mechanisms (e.g., hormones, cytokines, exosomes, polyunsaturated fatty acids) within anatomical niches. The characterization of immune cells within WAT marches forward, as highlighted by the identification of multiple macrophage subtypes in WAT beyond the traditional M1/M2 subtypes (161). Ultimately, treatment of obesity and T2DM will require a multipronged approach to address adipocyte insulin sensitivity and nutrient storage while maintaining an anti-inflammatory environment. Advances in immunotherapy and gene editing tools might provide unique opportunities to skew macrophage and T cell populations toward an anti-inflammatory state. Collaborative efforts to couple cell biology with immunology and genetics will be pivotal to address the current knowledge gaps and identify new therapeutic strategies to improve insulin sensitivity and nutrient storage for patients with obesity and T2DM.

Glossary

Abbreviations:

eWAT

epididymal WAT

FFA

free fatty acid

Foxp3

forkhead box protein p3

HFD

high-fat diet

HIF1

hypoxia-inducible factor 1

ILC2

group 2 innate lymphoid cell

IFNγ

interferon γ

KO

knockout

PAHSA

palmitic acid–hydroxystearic acid

PPARγ

peroxisome proliferator–activated receptor γ

T2DM

type 2 diabetes mellitus

Treg

regulatory T cell

UCP1

uncoupling protein 1

WAT

white adipose tissue